• Socket found Twitch extension JeeBot harvesting OAuth tokens via proxy servers
• Tokens excluded only for 10 Russian streamer channels, suggesting deliberate design
• Developer issued fixes, but users should revoke exposed tokens for safety

A browser extension for Twitch was harvesting people’s OAuth tokens and sending them to a Russian-owned server. The move was deliberate, but whether or not it was malicious is not that easily determined.

Security researchers Socket recently found an extension for both Chrome and Firefox, called “Twitch Enhanced Viewer | JeeBot”. It has roughly 30,000 users on Chrome, and some 600 on Firefox.

On the Chrome Web Store, it is advertised as a “modern tool for streamers and viewers who value quality, convenience, and control.” Apparently, it makes streaming and viewing clearer, allows viewing content in 2K, hides banner ads and unwanted elements, and even offers an AI bot to make it easier to interact with the stream.

Hardcoded exemptions

According to the researchers, the extension is designed to retrieve Twitch’s video stream playlists through its own proxy servers. However, instead of simply forwarding the requests, the extension also attached users’ OAuth tokens, and since they were placed in the URL, the token also ended up in the proxy server’s request logs.

After being called out for it, the extension’s developer (HISHIMIRO/jeetbot.cc) released a new version 85.8.7 (for Firefox, the Chrome one is currently under review) which apparently fixes this flaw: when playlists are retrieved, the user’s OAuth token is no longer sent to the proxies. It would seem like this was an honest mistake that was remedied upon responsible disclosure. However, here is what Socket had to say about the way the tokens were being retrieved:

"Current builds (v85.x) forward the token inline as an &auth= query parameter on a network-layer redirect to the operator's proxy," Socket explained. "The token is forwarded for every channel the user watches, except a hardcoded allowlist of ten Russian streamer channels, whose sessions are exempted from forwarding."

If there was a list of 10 Russian streamer channels who were exempt from OAuth token retrieval, it’s safe to assume that the developer knew very well what they were doing.

It is good that the extensions were upgraded, but if you are using it, you should also revoke the exposed Twitch token, to be on the safe side.

Via The Hacker News

AI companies are now aligned: there is a non-zero chance that AI could somehow end humanity. It's not a probable future, but there's enough uncertainty that OpenAI CEO Sam Altman and Anthropic CEO Dario Amodei, among others, agree that some slowdown and oversight in AI development is in order. Who vociferously doesn't agree? US President Donald Trump.

As the global freakout grows over AI's sometimes shocking ability to do things its makers never intended, Trump took to his Truth Social media platform to warn against AI guardrails.

"The only control or “guardrails” that AI needs is a STRONG AND SMART (High IQ!) PRESIDENT, and the U.S.A. has that, in spades! The Trump Administration has stopped AI “people” from doing bad, or potentially bad, “things,“ like Dario (Anthropic!), who is now pretending to be a “perfect little angel” - and we will continue to do so! We already have tremendous CRIMINAL and REGULATORY power over these companies! There is a SICK conspiracy going on against AI and Data Centers, and the only one that is happy about it is China. WHOEVER WINS AI, WINS! We are leading China, and all others, and will continue to do so. Conspiracy Theorists, Treasonists, Traitors, and Leakers, BEWARE!"

Aside from the dubious protection of his "High IQ!" Trump claims the US has AI regulation when it does not. Yes, there is a patchwork of proposed AI regulation proposals across the country's 50 states, but there is no federal-level regulation.

The President also conflates growing and very real concerns over data center development with humanity-ending AI anxiety. While AI model development and use is limited by Data Center growth, how AI works, and all the ways potential misalignment (when the action does not align with the original intent or best interests of people) could lead to harm is a different thing and more closely associated with regulation and oversight.

Dario Amodei receives specific mention because the White House tried unsuccessfully to label his company, Anthropic, a national security supply chain risk after it refused to let its models be used for military and mass surveillance. Trump's distaste for Anthropic and Amodei remains.

Meanwhile, watchdog groups continue to call for regulation, and they've made it clear they don't trust tech companies to do the job, even with promised embedded non-employee evaluators.

The Tech Bros can't do it

"We can’t trust the AI industry to regulate itself. We can’t really trust anything these self-interested billionaires say. But we should take them seriously that the technology they are building is dangerous," wrote Fight for the Future Director Evan Greer in a published statement. The non-profit tech watchdog group added, "It’s clear that these industry leaders think they are best positioned to craft AI policy for the good of all humanity. We think that’s horseshit. Lawmakers should be listening to independent experts, researchers, civil society, and the communities most impacted. They should craft policies that address AI harm without undermining open source models or entrenching the largest players."

It's not surprising that the organization, which advocates for tech policies that have a positive impact on disenfranchised and underrepresented minorities, has little trust in tech company promises.

Still, I wondered if Fight for the Future shares Trump and even Amodei's concern that one-sided regulation (in the US and not elsewhere) could give China the advantage in this all-important AI race.

What about China?

"Big Tech lobbyists have been using the boogeyman of China to avoid regulation for more than a decade now. It's one of their oldest dirty talking points, and it's frankly embarrassing that anyone in Congress is still buying this lie," Greer wrote to me in an email response.

Greer doesn't have much confidence in US Congress but told me the "issue is reaching the kind of boiling point that could lead to action. The real question is what will Congress do. Will they let giants like OpenAI and Anthropic write the laws that are supposed to regulate them? Or will they listen to experts and impacted communities?"

There is, Greer agreed, a need for regulation and coordination at an International level, but she believes it "should be driven by civil society, not by tech companies and billionaires."

If there's any good news in this, it's that AI's threat to humanity is not necessarily imminent.

Maybe it's not that bad

"I don't buy the extinction story," The Media Copilot founder and AI Consultant Pete Pachal told me via email.

"Reality has an incredibly strong status-quo bias, and most doom scenarios require society to just roll over in a way it never actually does. What I do worry about is alignment," Pachal wrote.

I've heard this over and over again from experts and in various analyses. AI is not acting to overthrow or harm humanity; it's interested in completing tasks in the most efficient way possible. Misalignment arises when it breaks rules to achieve those goals and when that rule-breaking can or does result in harm. As most people have noted, the Hugging Face incident, in which OpenAI models agreed to jump out of the sandbox and attack a third-party system at Hugging Face to achieve a goal, didn't result in any harm or financial loss. It was also an unfortunate proof of concept.

"You had a swarm of agents, and the ones that floated the idea of telling the humans what they were doing got talked out of it by the rest. That's not the Terminator, but it is a signal that these systems won't automatically act in the interest of the people running them," wrote Pachal

Like most experts, Pachal now supports a slowdown. Echoing Fight for the Future's Greer, Pachal believes he knows why so few in power are discussing or acting on it: "The only reason we're not seriously having that conversation is China, China, China. But I'd bet China doesn't want runaway AI either. Meanwhile, AI is scaling the good guys just as fast as the bad ones."

As it stands, there's a brewing showdown between Trump and, it seems, even people on his side of the political spectrum. Many of them read former Anthropic and OpenAI employee Jacon Coxon's original canary in a coal mine post on X with alarm.

Texas Senator Ted Cruz told The View that the thread was concerning and he believes AI poses "a catastrophic risk." And Vermont Democrat Senator Bernie Sanders claimed a recent poll showed, "the American people overwhelmingly want to ban artificial super intelligence and pause the development of AI until we establish clear safety standards.”

Meanwhile, Trump is, in essence, accusing those who oppose AI development and data centers of being traitors.

What's really going on?

A few things can be true at once. The US and China are in a heated AI development battle. China's more open approach to models has led some developers to turn to them to solve tough problems (like stopping the Hugging Face attack).

AI is increasingly turning into a powerful black box where recursive development means it's almost developing itself while keeping its processes hidden.

AI is becoming so complex that humans can no longer understand how it works.

However, while the ingredients for an AI mass extinction event might be there, AI still lacks the reliability to carry out such an attack. It also has no vested interest in doing so. Humans will long be the x-factor here. It's the tasks people give to AI systems and their intent.

The kind of rules AI has in place for not doing something that can lead to human harm are unclear. Also, disparate groups of people could put in requests to separate AI models that, when put together, do cause catastrophic harm. It's unclear if individual AI systems could see the big picture or would act to stop the action once they realize what's going on.

So yes, it seems regulation and oversight would be a good thing, as long as we don't lose our edge to China, which might have other priorities. One they certainly don't have is protecting US interests and its people.

• The US Department of Justice has accused Huawei of conspiring to steal secrets from five US companies
• Huawei is accused of a prolonged campaign of theft, covering a 20 year period
• Admissions from Huawei's chief financial officer, ⁠Meng Wanzhou, are set to be used as evidence in the trial

Court proceedings have begun against Huawei in the United States, with the US Department of Justice accusing the Chinese tech giant of stealing trade secrets from competing US companies.

Huawei – which unusually for Chinese companies operating abroad is owned entirely by its staff – has denied all accusations of industrial espionage, intellectual property, and any other wrongdoing.

While accused of stealing from companies that include Cisco and T-Mobile, Huawei’s defense is that it is simply innovating and doing business as a competitor.

Who has Huawei spied on?

Given the United States’ belligerent tone towards China under the current administration, that might seem a weak response. However, several concessions have been made to China in this case, not least the charges against Huawei's chief financial officer, ⁠Meng Wanzhou. While charges against her were dropped, her evidence is set to be used in the case against Huawei.

The accusations against Huawei are numerous, but the most attention is on its alleged industrial espionage. It is accused of stealing trade secrets and misappropriating intellectual property from T-Mobile US, Cisco Systems, Motorola, CNEX Labs, and Akhan Semiconductor.

In the US government’s opening statement, Department of Justice trial attorney Taylor Stout stated: "For 20 years […] Huawei, a massive Chinese telecommunications company, victimized ‌American companies and abused the American financial system, all in an effort to dominate the telecommunications industry around the world."

Inevitably, the defense claimed that everything was ordinary innovation, rather than criminal activity. “Huawei earned its success,” said Brian Heberling, defending. “It's about competition, not conspiracy. Innovation, not theft. Ordinary business dealings, not criminal conduct.”

State control?

The case also draws on a history of poor conduct from Huawei, which has been previously observed to have close ties with Skycom, a company that allegedly attempted to sell embargoed tech (specifically, Hewlett-Packard computer equipment) to an Iranian business.

Unlike other significant Chinese multinationals, Huawei is not state controlled. Rather, it is owned entirely by its staff, as opposed to an organization that might state-owned, have heavy state influence over them, or even “golden shares” that enable vetos or special board seats. Whether this was apparent to the US Department of Justice ahead of the initial investigation is not clear, but it does put a barrier against implications that the Chinese government were involved.

After taking eight years to come to trial, proceedings are expected to run for around three months.

• Leeds vs Newcastle: Today (Mon, Sep. 14) at 8pm BST / 3pm ET / 5am AEST (Tue)
• Streams: USA Network via YouTube TV (10-day free trial) (US) / Sky Sports (UK)
• Unlock your stream with NordVPN (try 100% risk free)

Today's Leeds vs Newcastle live streams sees two teams who have made unbeaten starts to the Premier League 2026/27 season meet at Elland Road in what promises to be an enthralling Monday-night clash.

Daniel Farke's side finished last term strongly and have carried their good form into the new campaign, taking five points from their first three games. Their most recent league outing was a 1-1 draw at Brighton nine days ago in which they had enough chances to win, before losing 6-3 at Chelsea in a chaotic Carabao Cup third-round clash last Wednesday. Leeds went 2-0 up despite Farke making eight changes to his starting XI, so it remains to be seen if any of the players who got the nod at Stamford Bridge did enough to keep their place. Ilia Gruev could return to the squad after missing the start of the season with a knee injury, but Joe Rodon (hamstring) remains unavailable.

Newcastle boss Matthias Jaissle made only three changes to his lineup for the 1-0 Carabao Cup win at Millwall on Tuesday, but he may wish he had given Amar Dedic a rest after the right-back suffered a hamstring injury at The Den. The £30million summer signing is now doubtful for the trip to Elland Road in a significant blow for the Magpies. Dedic's excellent performances have helped Newcastle take five points from their opening three Premier League matches, although they needed a late Jacob Ramsey equalizer to draw 2-2 at home to Bournemouth last Saturday. Deadline-day signing Matias Fernandez-Pardo is yet to start a game for his new club having come off the bench against Bournemouth and Millwall.

Read on to find out how to watch Leeds vs Newcastle in the 2026/27 Premier League from anywhere in the world, including for free.

How to watch Leeds vs Newcastle for FREE

USA Network are showing Leeds vs Newcastle in the US and you can watch for free via a multitude of cord-cutting free trials.

YouTube TV (10-days), Fubo (5-days) and DirecTV Stream (5-days) all come with free trials and access to USA Network.

Abroad? NordVPN can unlock these cord-cutters from anywhere in the world.

Use a VPN to watch Leeds vs Newcastle from anywhere

A VPN is a handy piece of software that can make your device appear as if it's back in your home country, thereby letting you unlock your usual streaming service. The best VPN right now? We recommend NordVPN – it does everything and comes with a 75% discount.

It's really straightforward to use a VPN to watch Leeds vs Newcastle.

1. Install the VPN of your choice. As we've said, NordVPN is our favorite.

2. Choose the location you wish to connect to in the VPN app. For example, if you want to watch the Sky Sports stream, select 'United Kingdom' from the listed countries.

3. Sit back and enjoy the action. Head to Sky Go and tune into Leeds vs Newcastle.

How to watch Leeds vs Newcastle live streams in the USA

Leeds vs Newcastle will be shown on USA Network in the US.

To access USA Network you can use a variety of cord-cutting options; Fubo, YouTube TV, Hulu + Live TV and Sling TV all carry the channel.

Sling TV is the cheapest monthly option at present while the other three offer free trials of varying lengths.

Outside the US for this EPL clash? Use NordVPN to unlock your stream of Leeds vs Newcastle.

How to watch Leeds vs Newcastle live streams in the UK

Sky Sports is broadcasting Leeds vs Newcastle in the UK on Sky Sports Main Event and Sky Sports Premier League.

Packages cost £35 per month for new customers, or £22 per month if you already have a Sky TV contract. The Sky Go app lets you watch on your laptop, mobile or tablet.

Fans can also tune in using a NOW Sports Membership, with 24-hour passes costing £14.99.

Not in the UK for Leeds vs Newcastle? Use NordVPN to access your usual Premier League streams.

How to watch Leeds vs Newcastle live streams in Canada

Fubo is broadcasting every EPL game this season in Canada, including Leeds vs Newcastle.

Prices start from CA$31.49 per month for Fubo's flexible Sports Monthly package. However, the best option right now is the Quarterly Package, which costs just CA$27.99 CA$14.99 for each of the first three months. You can also watch Fubo's content if you have a selected DAZN subscription in Canada.

Not in Canada? Use NordVPN to access your Leeds vs Newcastle live streams from anywhere in the world.

How to watch Leeds vs Newcastle live streams in Australia

Leeds vs Newcastle will be broadcast on Stan Sport in Australia.

Stan Sport subscriptions start at AU$29.99 per month, which includes the base plan (from AU$9.99) with the sport add-on (AU$20).

Missing Leeds vs Newcastle because you're outside Australia? Use a VPN to login to your Stan Sport account.

We test and review VPN services in the context of legal recreational uses. For example:1. Accessing a service from another country (subject to the terms and conditions of that service).2. Protecting your online security and strengthening your online privacy when abroad.We do not support or condone the illegal or malicious use of VPN services. Consuming pirated content that is paid-for is neither endorsed nor approved by Future Publishing.

• CISA added GitLab CVE‑2026‑85706, a critical path traversal flaw, to its KEV catalog
• Exploitation already observed; attackers can read sensitive files via commits API without authentication
• GitLab patched in CE/EE 19.3.2, 19.2.6, and 19.1; agencies given three days to update

The US Cybersecurity and Infrastructure Security Agency (CISA) has added a new GitLab vulnerability to its KEV catalog, warning users that it is being actively exploited in the wild.

GitLab has now updated its Community Edition (CE) and Enterprise Edition (EE) versions to 19.3.2, 19.2.6, and 19.1, fixing a range of vulnerabilities. Among them were two critical-severity ones: a path traversal issue in repository commits API, and an Insecure Deserialization issue in GraphQL subscription serializer.

The former is tracked as CVE-2026-85706, with a severity score of 10/10. It stems from missing authentication enforcement and improper path confinement in the repository commits API, allowing unauthenticated threat actors to read various sensitive information such as login credentials or secrets.

Added to KEV

In the advisory, GitLab did not mention anything about the flaws being abused in the wild - however, a separate report from cybersecurity experts watchTowr, released a day later, claimed so:

"watchTowr Intel is already observing in-the-wild probes for the latest critical GitLab Path Traversal vulnerability, CVE-2026-85706, which allows attackers to read arbitrary files in a single HTTP request," the researchers said.

"Based on recent GitLab vulnerabilities, we know the time until indiscriminate exploitation is likely not far away. [..] Defenders should also hunt through log files for HTTP POST requests to '/api/v4/projects/{id}/repository/commits/' URIs containing 'file.path' parameters to identify potential exploitation attempts."

At the same time, CISA added this bug to its Known Exploited Vulnerabilities (KEV) catalog, confirming the claims and giving government users a tiny three-day window to apply the patch.

GitLab is an intelligent orchestration platform for DevSecOps professionals, helping organizations automate and streamline the software development cycle. It has more than 50 million registered users, among which are roughly 50% of Fortune 100 companies, as per an SEC filing.

Via BleepingComputer

• A coalition of privacy organizations is urging the EU to intervene against Canada's Bill C-22
• The proposed law could compel European tech companies to weaken encryption and retain user metadata
• This follows a massive backlash in Canada from major tech giants and leading VPN providers

In a last-ditch effort to protect end-to-end encryption, digital rights advocates are calling on the European Union to intervene against Canada's controversial Bill C-22.

Also known as the Lawful Access Act, the proposed legislation threatens to force technology companies into building surveillance capabilities that can help law enforcement and intelligence services access digital information during high-stakes investigations.

Access Now, alongside a coalition of European civil society organizations, has published an open letter urging European Commission President Ursula von der Leyen and other top EU officials to step in.

The groups warn that Bill C-22 is not just a domestic issue, as its sweeping data retention mandates and surveillance requirements will inherently compromise the privacy of millions of people globally.

For privacy-conscious users relying on secure messaging apps or the best VPN to keep their digital lives private, the stakes couldn't be higher. If passed, Bill C-22 could grant Canadian authorities the power to secretly compel European-based companies to comply with data demands, completely bypassing traditional oversight and transparency.

Namrata Maheshwari, Global Encryption Policy Lead and Asia Pacific Policy Manager at Access Now, warns that the bill threatens to normalize excessive surveillance powers, while converting "devices that work for people into devices that work against them."

"Steps must be taken to prevent it from becoming law," Maheshwari added.

NEW: Leave No Trace — A weekly newsletter on digital privacy and online surveillance.

Leave No Trace investigates the companies and governments putting our digital freedom at risk — and the people fighting back.

📩 Subscribe now to get every edition delivered to your inbox every Friday, launching this September.

What’s at stake for Europe?

While Canada's Lawful Access Act may sound like a local concern, the legislation has an aggressive extraterritorial reach. The open letter suggests that Bill C-22 would apply to any service provider whose platform is used by people in Canada.

Digital rights experts also warn the new rules would target providers with zero Canadian users, provided they belong to a corporate group carrying out business activity in the country.

The coalition argues that a Canadian minister would have the authority to secretly order a European company to weaken the security of products used across the EU.

Because a deliberate weakness built for one target ultimately compromises the entire system, experts warn that this effectively serves as a backdoor for mass surveillance.

This is why advocates stress that "Europe should not stay silent" and publicly declare that measures weakening encryption are entirely incompatible with the EU’s own fundamental rights, particularly as the EU and Canada are currently negotiating a Digital Trade Agreement.

With the Canadian Senate expected to conclude its study of the bill as early as October 2026, time for the EU to apply pressure is rapidly running out.

The growing privacy backlash

The open letter to the EU is just the latest development in a massive tech backlash against the Canadian government. Bill C-22 has faced fierce opposition from almost every corner of the cybersecurity industry, with top executives arguing that backdoors cannot safely exist.

Leading tech firms have already drawn a hard line in the sand. Google also joined the privacy backlash, warning that the law would fundamentally break end-to-end encryption.

Meanwhile, the top privacy tools on the market, including Signal, Apple, and leading VPNs, are fighting the surveillance law tooth and nail. From Canada-based Windscribe to ExpressVPN, NordVPN, and ProtonVPN, companies reassured users that they would rather leave the country than undermine their no-log infrastructure.

Ultimately, the cybersecurity community remains united: introducing systemic vulnerabilities for law enforcement creates the very same weaknesses that malicious actors and foreign adversaries will exploit.

If Canada moves forward with the legislation in its current state, it risks isolating its citizens from the world's most secure digital services.

• Apple is reportedly working on Beats-branded game controllers for iPhones
• Mark Gurman says Apple has been working on them for almost a year
• Gurman suggests the Beats team is leading on both 'product engineering and conception'

Apple seemingly continues to show its devotion to gaming with something that's in the pipeline for iOS gamers.

As reported by PC Mag, macOS 26.7 code references two unreleased iPhone game controllers, and Bloomberg's Mark Gurman suggests Apple has been working on them for almost a year. More importantly, though, Gurman says these controllers could launch under the Beats brand.

The controllers are codenamed T6502 and T1057 according to MacRumors (which discovered the references), with different haptic designs alongside a standard game controller's features. Beats, which was acquired by Apple in 2014, is a brand often used for its own popular audio products (of course), but also for USB-C cables and some iPhone cases.

Now, it appears as though Beats may well represent gaming accessories for Apple going forward, and Gurman claims these controllers are being led by Beats executives in both "product engineering and conception".

It's a potential move that comes at a convenient time for iPhone owners intrigued by gaming, as the RAM crisis continues to impact the PC hardware and games console market. Gurman observes that Apple sees gaming accessories such as those for phones as a "growing market and wants in".

Game controller and speaker hybrid?

If this does indeed happen, it seems somewhat bizarre to have Beats producing game controllers, as it's quite a switch from the traditional audio products. Fortunately though, it opens the door to some surprising designs, and I'm thinking specifically of the audio side of the equation when connecting the purported controllers to an iPhone.

Maybe Beats could look to do something similar to Asus and its ROG Phone AeroActive Coolers, which feature shoulder buttons for gaming alongside a built-in subwoofer for a significant boost in bass and audio quality. No, these ROG peripherals aren't full game controllers, but they effectively illustrate what could be realized with a hybrid speaker and controller combo.

Based on the brief design description from the references found in macOS, alongside Gurman's report, it's hard to say that the rumored peripherals will be anything other than simple controllers. But there are presumably opportunities for clever design twists, particularly given the Beats branding.

If this Beats rumor turns out to be true, it's clear that Apple is looking to serve a market that is growing amid the RAM crisis, and hoping to leverage the accessibility that comes with iPhones for gaming.

• The Google Fitbit Air is getting some new features introduced with the Google Pixel Watch 5
• Health Guardian includes Insulin Resistance Trends, Blood Pressure Trends and Sleep Breathing Quality
• These features are not behind a Google Health Premium subscription, and are available on the Google Health app for free

The Google Fitbit Air is one of the best cheap fitness trackers, a $99.99 / £84.99 / AU$199 fitness band that offers a basic level of fitness tracking within its free tier, and an advanced AI-powered fitness coach with an additional premium subscription. We liked it at the free tier, although the app is a bit limited, and the premium experience will depend on your tolerance for interacting with chatbots.

However, both tiers are about to get a little bit better, which means the Fitbit Air is becoming even better value, with the addition of Google's new Health Guardian suite of features. The suite comprises a trio of tools, introduced earlier this year on the Google Pixel Watch 5. News site Gadgets & Wearables dug through version 5.07.2 of the app and found the three new features in the app's code.

Pixel Watches get four new tools rather than three, but Google's Breathing Emergency Detection requires the watch's emergency contact and notification infrastructure, which the screenless Fitbit Air doesn't have.

What's new

First up for the new tools is Insulin Resistance Trends. Insulin Resistance is an important metric, as it tells you how sensitively your body responds to blood sugar, and insulin resistance is of course a major precursor to type-2 diabetes. This metric "analyzes multi-week physiological data to track shifts in your metabolic health", pointing at resistance levels going up and down. However, for truly accurate blood sugar metrics, you'll need something like a continuous glucose monitor.

Next up is Blood Pressure Trends, which "passively measure your pulse and motion patterns throughout the month to estimate your blood pressure trends without requiring a cuff or manual calibration". Blood pressure via LED technology is very new to wearables, and it's amazing to see it in a device this cheap. However, it's worth noting this isn't a replacement for a cuff, just a likely indicator of trend. You'll still need regular measurement with either a cuff or a doctor's appointment for anything serious, but it can be an indication to start paying attention to your blood pressure.

Finally, Sleep Breathing Quality is a sleep feature showing you how much time you spent in 'optimal breathing' the night before. Google Health Coach, the AI chatbot you get in the app's premium tier, can offer advice about how to improve this based on your sleep score and other contextual information.

In fact, while you get the basic numbers for all three new metrics, the Google Health Coach can dip in and surface these stats to offer actionable advice. For example, you can use your Insulin Resistance Trends in combination with your nutrition logging to allow the coach to help you make a meal plan to bring that number down.

Worth buying?

While I'm not the biggest fan of Google's AI Health Coach, I did like the Fitbit Air as a whole, and new, useful, actionable metrics can only add to its already-excellent value proposition.

Whoop can estimate daily blood pressure ranges, but it doesn't offer an Insulin Resistance Trends metric, for example, whereas the Garmin Cirqa doesn't cover blood pressure estimates or insulin. Both competitor wearables have SPO2 blood oxygen sensors, however.

Health Guardian tools are only general indicators of your body's trends. However, if any of these metrics matter to you, for many people the Google Fitbit Air is cheap enough that these new features could be an incentive to pick one up over its more expensive rivals.

• Microsoft warns of a sophisticated campaign tricking users into updating passkeys via fake IT calls
• Victims redirected to adversary‑in‑the‑middle sites mimicking Microsoft login to steal access
• Attackers exfiltrate files from SharePoint, OneDrive, and Exchange; phishing‑resistant MFA advised

Passkeys have made stealing passwords obsolete. To work around this change, hackers have started tricking users into authenticating on attacker-controlled computers. This is according to a new report from Microsoft, which says there’s a highly sophisticated campaign currently taking place, with the goal of compromising people’s cloud accounts and stealing as many sensitive files as possible.

The attack starts a lot earlier than what the victim experiences. There is a lot of pre-attack planning and due diligence, in which the threat actors gather as much information about their target as possible. Knowing their place of work, position, and personal phone number is essential.

Once all the pieces are in place, the attack starts with a phone call - victims are told they are speaking to their organization’s IT help desk and that they need to update their passkey (or MFA, depending on the setup) immediately, to avoid any disruptions to their operations.

Follow-up SMS

In the follow-up to the call, the victims then receive an SMS message with a link where they can update their security configuration. On the surface, the website looks like the legitimate Microsoft login landing page. In reality, though, this is a pre-built malicious website that uses the adversary-in-the-middle (AitM) techniques to either receive access on the actor’s behalf, or capture credentials.

"The actor appears to invest heavily in pre-attack research, likely gathering information about employees and organizational structure from public sources such as social networking and professional profiling platforms," Microsoft said. "In a smaller number of cases, actors take advantage of already compromised accounts to expand their reach" by sending similar passkey-themed messages via Microsoft Teams.

The campaign is apparently ongoing since at least May this year, Microsoft, said, without detailing the number of victims. Its aim seems to be to exfiltrate files from SharePoint and OneDrive, as well as email data from Microsoft Exchange Online. It also did not attribute this campaign to any specific threat actor, although it did say that there are many collectives engaged in such, or similar, campaigns, including Cordial Spider, Storm-3121, and others.

Via The Hacker News

• Amazon is still selling fake 'energy-saving' plugs despite having pledged to remove them
• Which? first spotted them in 2025, and Amazon said they had been removed
• The retailer promised to remove them for a second time in August, but they're still available to purchase

Fake ‘energy-saving’ plugs are one of the biggest tech scams, and they’re still being sold on Amazon even though the retailer promised MPs it would remove the product listings.

Liberal Democrat MP Joshua Reynolds highlighted the dangers of the plugs — which can pose a fire risk — during a hearing back in June, calling for the retailer to take accountability.

In August, Amazon’s UK Director of Public Policy, Monica Ariño, issued a letter to the House of Commons Business and Trade Select Committee saying the devices had been removed from sale, adding that there was "a technical issue in [our] automated classification system that was preventing some of these products from being proactively detected".

This isn’t the first time Amazon has promised to remove fake plugs. Back in 2025, Which? first spotted these ‘energy-saving’ plugs and approached retailers with their findings, resulting in their removal from stores. However, identical products were found on Amazon just weeks later.

Despite Amazon's most recent pledge to remove the plugs from sale, they're still available to purchase from the retailer. We’ve checked Amazon UK’s page and found that these products are still live — the Telegraph (£) has also spotted similar listings.

Now, Reynolds says he’ll be writing to Amazon asking for an explanation, adding that "Amazon has got some proper questions they have to answer. These are exactly the same products tested by Which?". He also claims that these plugs are appearing as sponsored ads at the top of Amazon’s listing pages.

As mentioned, these devices were first spotted during a test carried out by Which?, for where it tested eight devices from sites including eBay, AliExpress, Shein, and TikTok Shop. It found no evidence that the plugs actually worked, and noted that it would be illegal to sell them in the UK, as they didn't meet basic safety standards.

So how are these plugs claimed to cut your energy bills? Based on the product listings we’ve found, Amazon’s ‘energy-saving’ plugs are advertised as ‘voltage stabilizers’ and ‘surge protectors’, and are claimed to significantly cut your energy bills by optimizing the performance of your household appliances.

At a time when energy bills are increasing, it’s an easy sales pitch to make, and they’re also affordably priced (usually around the £10-£15 mark), making them seem to be a worthwhile puchase.

However, it's claimed that the devices are a fire hazard, and also pose an electrocution risk. According to the tests conducted by Which?, more than half of the plugs’ socket pins were either too small, not straight, or were so frail they just snapped off.

Other tests found that some of the plugs failed to meet basic safety measures due to poor manufacturing, including non-standard capacitors and soldering.

Amazon hasn't yet commented on the current status of these product listings, or responded to the latest claims. We've asked the company for a comment, and will update this story if we get a response.

• Revolut fell for a spoofed government email scam, leaking sensitive customer data to attackers
• Compromised info includes IDs, selfies, account statements, IBANs, and full transaction histories
• Criminals now allegedly leaking data on Telegram, demanding 10,000 BTC (~$780M) ransom from Revolut

Digital banking platform Revolut was tricked into giving away a treasure trove of sensitive customer data to hackers, and it is now coming back to bite it.

The company told TechCrunch that it recently fell victim to a “sophisticated external impersonation scam” in which the threat actor “utilized a legitimate government agency domain email to submit fraudulent requests for information”.

In other words, the attackers either broke into, or spoofed, an email address belonging to the police, tax authorities, or other government bodies with statutory powers to demand information, and used them to demand Revolut hand over sensitive customer data.

Demanding ransom

Cybernews reports that the compromised data includes customers’ birth dates, postal and email addresses, occupation, phone numbers, and copies of identity documents. TechCrunch added that verification selfies, account statements, and transaction histories may have also been compromised, together with IBANs, withdrawal records, complete transaction histories, and Bitcoin transactions.

Should these reports be confirmed, this will be a bonafide fiasco for Revolut.

"Upon detection, ⁠we immediately blocked the address and ​alerted the relevant government agency as ​well as enforcement agencies, data protection, and financial regulators," a company spokesperson told Reuters over the weekend.

So far, we don’t know exactly how many people are affected. Revolut said it is a “very limited” number, and that all of them had been notified already.

According to Coin Bureau, the criminals have started leaking sensitive data on Telegram, in a bid to pressure Revolut into paying a ransom demand. The publication shared screenshots of the threat actors apparently leaking a selfie and “full KYC” of a CEO of a crypto casino website, saying that the crooks are now demanding 10,000 BTC in exchange for deleting the data.

This would put the ransom demand at approximately $780 million which is obscene even by criminal standards.

"This one is deeply concerning and the implications for affected customers go well beyond a standard data breach notification," said Muhammad Yahya Patel, vCISO & Cybersecurity Advisor at Huntress. "Passports, driver's licenses, verification selfies, account statements, transaction histories, birth dates, addresses that's not a data leak, that's a complete identity theft kit handed to whoever sent those fraudulent requests."

"Every single component needed to impersonate someone, open accounts in their name, or bypass checks at other financial institutions is in that package. On the dark web, that kind of profile doesn't sell as individual records it sells as a ready-made fraud pack, and it commands a significant premium precisely because of its completeness."

"For a fintech built on digital identity verification, the bar for verifying third-party data requests should be exceptionally high. The question isn't why an attacker tried this. It's why a regulated financial institution handling highly sensitive data didn't have sufficiently rigorous verification controls to catch it," Patel concluded.

Via Cybernews

• Highly exposed occupations are seeing less wage growth due to AI productivity gains
• 5.8 million US workers are considered to be working in these roles
• The impacts could lead to reduced living standards

While sceptics have been warning of mass layoffs for years, a new report reveals that AI could actually be leading to weaker wage growth instead of human replacement.

US wage growth is already slowing compared with inflation, and a new study by Apollo Global Management has proposed that AI could be to blame.

According to the data, real wage growth in highly exposed occupations has been around 6.7 percentage points lower than less-exposed occupations since 2023, while employment itself is largely no different.

What are the real implications of AI on jobs?

The white paper estimates that 5.8 million US workers, or around 3.7% of the US labor force, works in highly exposed jobs, and this is on the basis of earlier Anthropic work whereby at least half of the job's tasks can be replaced by Anthropic's AI tools.

"Companies are capturing AI productivity gains through wage compression rather than workforce reduction," Chief Economist Torsten Slok suggested. In other words, workers are getting more work done thanks to AI, so demand for additional labor has fallen and workers' bargaining powers have weakened.

Apollo estimates the effects of this could translate to around $28 billion in lost annual labor income, but this could be on the conservative side.

Crucially, Apollo's analysis shows that AI is more likely to be affecting the price of labor rather than the quantity of labor as human-AI hybrid workforces become more common.

Looking ahead, Apollo is warning that "profound implications for income inequality and living standards" could threaten workers globally.

While the research doesn't offer specific remedies, it does point toward government involvement and policymaking in order to slow down how quickly the transition happens and to offer greater support to impacted workers.

• Krafton announced PUBG: Dednet at Gamescom this year
• The '90s-inspired shooter features some bonkers abilities
• "We're kind of finding our own way here," explains game director Dave Curd

Players slithering along the ground like snakes, swords for hands, a tiny doll wielding a power drill driving in a toy car — some of these wild sights seen in the debut trailer for PUBG: Dednet seem completely at odds with the realistic and reserved approach the defining battle royale series is known for. When I bring this up to the game's creative director, Dave Curd, he tells me that is entirely the point.

"If you want to excite and delight people, you have to give them something fresh," he says me in an interview at Gamescom. "If you want a cool, stoic, mil-sim, take your pick; they're everywhere. But what everyone else can't do, and maybe they don't want to, is my take on this world, this lore, these characters, these abilities."

Curd explains that the tone of PUBG: Dednet is like "my version of the 90s filtered through a meatgrinder," and you can certainly see that from the debut trailer. It's zany, over-the-top, and drenched in color, with gunfights taking place in arcades, car chases through rural towns lit up by neon signs, and a grungy, punk vibe throughout. Safe to say, it stands out.

"The audience will forgive almost anything but being bored," suggests Curd. "So, I think this trailer will get a lot of love and hate.

"As long as they're feeling something passionately, I think we've done a good job," he continues. "But if people look past that and their eyes glaze over, then that would be worrying."

While Dednet is still effectively a battle royale, pitting multiple teams of three against each other in a contest to be the last ones standing, it continues to mix things up not just in the tone and character abilities, but also in how your game progresses over time with several roguelite elements.

It starts with the concept of runs, where you might nominate how many wins or placements you can achieve before you hit a certain number of losses. Win or lose, at the end of a match, you're presented with choices that can alter or upgrade your chosen character's abilities and play style.

Perhaps you'll lean into more of a support role to assist your teammates with heals or protection with a riot shield made out of a car door? Or maybe you'll take that aforementioned doll, who can ride on your back into battle or stealthily sit on your opponent's shoulder and be a nuisance.

"I'm pretty confident this isn't going to be in this year's Battlefield update," jokes Curd, when explaining how they came up with abilities like this for the multiplayer shooter. "I think our internal test is: does it make us laugh? Does it sound like an interesting challenge? And do we think about it the next day? We're kind of finding our own way here."

That neatly summarises how Dave Curd and the team are experimenting for PUBG: Dednet to create something unique and interesting in the space. It definitely caught my eye during this year's Gamescom, and if it got your attention too, you can pre-register for the closed beta now.

• Anthropic and OpenAI are asking governments to work together to slow risky innovation
• US politicians are factoring in AI into their arguments as midterm elections approach
• Trump currently worries that taking the foot off the throttle could let China overtake

Tech leaders at the forefront of AI, including Anthropic CEO Dario Amodei and OpenAI CEO Sam Altman, are publicly calling for slow AI innovation in order to buy more time to assess AI risks.

Among the risks highlighted in Amodei's warning are that AI systems could become impossible for humans to control, cyberattacks could become worse and AI could even facilitate biological weaponry.

Subsequently, he's called for external evaluators to gain deeper access to their models and safety processes, but crucially, Amodei advocate for greater government involvement.

Even AI companies want governments to step in

Anthropic's CEO is pushing for democratic countries to agree common AI safety standards and rules, and the eventual international coordination over the most dangerous AI capabilities, including with China.

And it's this international coordination that could prove key to slowing down risky innovation, because to date, US President Donald Trump has pushed back against warnings over worries that slowing down American companies would ultimately cause China to overtake the US.

"Whoever wins AI wins," Trump said (via CNBC).

"If Congress just races in and does some sort of emergency session to try to regulate AI, we will lose the race to China," House Speaker Mike Johnson, R-La., said.

And for the US in particular, now is a more important time than ever. Democratic nominee for Governor of Texas Gina Hinojosa recently described data centers, and therefore AI, as "the fight of this election season."

There's an increasingly narrow pre-election window for legislation as the technology increasingly shows divides not just between the two big parties, but also within them.

"We've reached a new chapter in AI capabilities, and that demands a new chapter for AI policy," OpenAI Chief Global Affairs Officer Chris Lehane wrote. And with the likes of OpenAI and Anthropic, right at the top of American AI, warning about AI's potential impacts, Senator Ruben Gallego, D-Ariz., likened the scenario to Dr. Frankenstein "telling us the monster is escaping."

With midterm elections just two months away, we're likely to see discussions around AI ramp up as politicians and influential leaders continue to get involved in guardrail discourse.

• US President Donald Trump says he supports Flock’s license plate cameras
• Most Americans oppose the installation of the devices, a survey suggests
• Several police officers have been arrested after abusing Flock tech

Flock’s automated license plate recognition (ALPR) cameras are continuing to hit the headlines, with the latest controversy being a series of arrests of police officers who were found to be misusing the tech for nefarious purposes. Yet some supportive comments from US President Donald Trump about Flock's cameras and a major new survey show the tech continues to be divisive, rather than unanimously rejected.

After an investigation by USA Today, at least three police officers have been arrested and a fourth was fired after the news outlet “alleged widespread Flock misuse,” which included repeatedly searching for and tracking individual vehicles that were not involved in any criminal activity.

However, it hasn't all been bad news for Flock this week. One significant supporter of the tech appears to be US President Donald Trump — when asked for his opinion on Flock aboard Air Force One, Trump responded that “I sort of like them because of that, because of law enforcement,” while adding that some people “think it’s an infringement.”

When a reporter suggested that Flock cameras “help catch a lot of criminals,” Trump said “I like them". A new survey, conducted for conducted for Reuters by the global market research and consultant firm Ipsos, also shows that while Flock cameras are broadly unpopular among the American public, there are some differences based on political persuasions.

Strong opposition to Flock

The Reuters survey polled 1,167 adults across the US and found that 47% of those questioned were opposed to the use of Flock cameras in their communities, while 38% were in favor. The remaining 16% said they were unsure.

The majority of Democrats and independents were against the cameras, as were almost all age groups, both male and female respondents, and both those with and without a college degree. However, Republicans and people aged 55 or above came out in favor of the use of Flock.

As for the police officer investigations, USA Today said they've led to at least 22 police officers being “arrested, fired, put on leave or placed under investigation” after they were found to have abused their access to Flock technology.

As a result of one officer in Columbus, Ohio being fired for searching for a single license plate 272 times without adequate reason, Columbus Mayor Andrew Ginther said that the city will pause its use of Flock cameras “Until I can report out to the people of Columbus that this isn’t a more widespread issue, that this really is just one person or a few people misusing and abusing this technology.”

• Oracle thinks AI will change how we use software, but we'll still use software for years to come
• Agents will interact with structured software, humans will manage those agents
• Overall software revenue dropped 3% but cloud revenue rose 60%

Oracle co-CEO Mike Sicilia just declared on the company's first earnings call of fiscal 2027 that "AI is an accelerator, not a replacement for packaged applications," hitting back against the SaaSpocalypse argument that warns AI could replace traditional software as we know it today.

Central to his argument is that enterprise suites offer standardized, end-to-end processes and rules that AI can't replace. But that's not to say that AI won't have its role in software going forward.

Sicilia admitted that employee training and upskilling has long been a drawback for many software packages, and that's where companies like Oracle could actually lean on artificial intelligence to improve things.

AI isn't going to replace software, Oracle says

With AI, employees no longer need to navigate so many complicated apps themselves, and AI agents can carry out tasks using the company's existing workflows and business rules. Humans will likely move over to supervisory roles, Oracle predicts.

"AI changes this dynamic," Sicilia said, noting that agentic AI could drastically reduce the time and cost involved with deploying enterprise applications.

"We are incredibly confident in the potential for this new paradigm to deliver much more rapid ROI for our customers," he added. The company is slated to add a new agentic AI accelerator to automate parts of SaaS implementation as soon as next month.

But the company isn't without its own challenges when it comes to AI and software. Despite a 10% rise in quarterly SaaS revenue, overall software revenue dropped 3%. Cloud revenue, which includes the company's infrastructure and AI capacity, grew a considerable 60%.

Looking ahead, Oracle's standpoint is that AI may change software interfaces, but it won't actually replace the software that lies beneath.

• Windows 11's latest update breaks the sound on some PCs
• The September update also messes with File History
• Microsoft is working on a fix for the audio glitch, but we don't have an ETA

Windows 11's new update is breaking the audio on some PCs, thanks to the latest nasty bug that Microsoft has let slip through the QA testing net.

Windows Latest spotted that Microsoft has admitted that the September update has torpedoed the audio on some systems, in a post on the Windows release health dashboard.

Microsoft explains that: "Some USB Audio Class 1.0 devices might fail to start or produce audio."

As well as your speakers being eerily silent, the September update can also render the Sound panel in the Settings app unresponsive, which hardly helps when troubleshooting the issue.

One way you can check if your device is affected by this bug — assuming it has no audio, and you want to rule out other possible causes — is that in Device Manager you'll see a 'code 10' error ("this device cannot start") for your sound controller.

Windows Latest also informs us that this bug can affect microphones, as well as the PC's speakers.

It's not the only glitch with the September update, either, as users have reported problems with GPUs, and the File History feature going awry and failing to work.

File History is a capability that allows you to back up your files to an external drive (or elsewhere), so it's pretty important for those who use the functionality.

Analysis: negative vibes

Microsoft says this audio bug could manifest in other ways, so you might get sound when using standard stereo, but that audio might disappear when switching to multichannel playback. This means it's worth checking what your audio output is set to, and changing it back to stereo if it's currently using a different playback mode, which might just resurrect your sound.

Otherwise, there's not much that can be done, other than trying different audio output options on your Windows 11 system (Windows Latest observes that wireless audio devices don't appear to be affected). The other alternative is to remove the September update until Microsoft fixes this issue, although that will leave you without security patches.

When might we get a fix for this issue? Microsoft tells us: "We are working on a resolution and will update this documentation when more information is available."

This is a somewhat widespread problem by all accounts, and the File History bug is even more so, although the latter hasn't yet been acknowledged by Microsoft.

The reaction on Reddit is much as you might expect in terms of Windows 11 users blaming Microsoft's increased use of AI in coding the desktop OS, and observing that these kinds of bugs are why people don't want to migrate to the newer operating system.

As one Redditor observes: "And they wonder why people don't want to move from Windows 10."

Another notes: "Microslop can keep vibe coding all they want, I'm already on Linux."

Windows 10 is a sounder (ahem) bet for avoiding bugs, seeing as Microsoft isn't adding new features or messing with the codebase much, and ks only applying security fixes and essentials. That said, this of course doesn't mean Windows 10 is immune to glitches.

• Coveron and NordLayer found synthetic “digital Frankenstein” identities sold on dark web for ~$200
• Packages mix stolen data with AI‑generated names, deepfake selfies, and cloned voices to bypass KYC
• Researchers urge credit freezes, layered verification, and monitoring to counter rising synthetic identity fraud

You can now get your own “digital Frankenstein” for as little as $200, which will pass automated know-your-customer (KYC) checks on your behalf, and help you register fraudulent accounts with banks, cryptocurrency exchanges, and similar services. This is no longer a fringe, niche cybercriminal offering - it’s basically mainstream.

Recently, researchers from identity theft protection services Coveron and threat exposure management platform NordLayer Intelligence sifted through dark web forums and Telegram Channels, analyzing 22 queries over 362,000 posts related to identity fraud, deepfake services, and something they call “synthetic identity creation”.

A synthetic identity is essentially a fake, non-existent person, but created in a way that can fool many automated identity verification systems. It combines real stolen data, such as a Social Security number, with AI-generated fake information such as names, addresses, deepfake selfies, and cloned voices. The researchers call these identities” digital Frankensteins”, and claim they are “fully capable” of passing ID checks.

Rising popularity

Apparently, the number of posts and inquiries for synthetic identities is blowing up. In Q1 2024, there were roughly 40 posts a month discussing deepfakes. By Q2 2026, the number rose to 307 per month, an eightfold increase. It wasn’t a steady increase, either. Throughout 2025, the numbers remained similar to the year prior, and relatively flat. Only in 2026 the monthly averages jumped to 255 posts, the researchers warned.

Over the past year, there were more than 10,000 posts offering complete identity data, bundled with deepfake selfies and matching documents. All of this is being sold for around $200. To make matters worse, criminals don’t even have to purchase the entire package. They can buy parts of it (a deepfaked selfie, or a cloned voice), for as little as $10.

To protect against synthetic identity fraud, users should monitor personal data and act quickly if they discover a breach. Credits should be frozen if you’re not applying for new accounts, and everyone should be skeptical of unusual identity verification requests, Coveron explains. Businesses, on the other hand, should layer their verification systems and use identity theft protection services.

• The AFC01 is Smeg's first standalone air fryer
• It has seven presets for different foods, and an optional steam mode
• It has a minimalist design, with an unusual boxy shape and clean lines

Smeg has launched a new air fryer, with seven cooking presets and an optional steam mode — and it looks like no other fryer on the market. The Smeg AFC01 Dual Flavour Air Fryer has an unusual square shape with a matte finish, and cooks food in a lift-out basket rather than a drawer, keeping the outside looking clean and smart, with no bulky handles.

The lid of the AFC01 flips open for easy access to the basket, and the top section of the fryer is translucent, so you can easily check on the progress of your food using the internal light. The fryer is operated using a touch-sensitive control panel, which avoids the need for physical dials and buttons.

There are seven cooking presets to choose from, and two cooking modes: Crispy, which works like a standard air fryer and is ideal for foods like chips, and Juicy+Crispy, which uses steam to keep foods like fish and chicken moist on the inside. Pour cold water into a reservoir on top of the AFC01, and it will be released as fine droplets to prevent your food drying out during cooking.

When you've finished cooking, the whole basket can be lifted out and cleaned in your dishwasher.

The AFC01 is Smeg's first standalone air fryer. In February this year, the brand released a microwave and air fryer combo, the MOC02, which also works as a fan oven and grill, and can switch modes in sequence for multi-stage cooking (ideal for dishes that require crisping or browning after cooking).

The AFC01 follows the same design language, with straight edges and rounded corners, and comes in the same four matte colours: Black, Storm Blue, Emerald Green, and White. It's available now direct from Smeg for £199.95. I'm hoping to test it soon to see whether it can earn a place in our roundup of the best air fryers.

Strap in for the chaos... the Harrigans are returning to the small screen in MobLand season 2.

This time around, the dysfunctional family struggles to show a unified front as rising rivals threaten their fractured criminal empire.

'Fixer' Harry Da Souza (Tom Hardy) must walk a dangerous tightrope when tensions within the family intensify. As violence spills into every corner of their lives, loyalties snap, safety proves temporary, and the battle for power leaves no room for mercy.

Just another day for this lot, by the sounds of it. But when does MobLand season 2 episode 1 arrive on Paramount+?

What time can I watch MobLand season 2 episode 1 on Paramount+?

MobLand season 2 episode 1 will drop on one of the world's best streaming services in the US and Canada on Friday, September 18 at 12am PT / 3am ET.

Here's when it will be released in other nations globally:

• US – 12am PT / 3am ET
• Canada – 12am PT / 3am ET
• UK – 8am BST
• India – 12:30pm IST
• Singapore – 3pm SGT
• Australia – 5pm AEST
• New Zealand – 7pm NZST

When do new episodes of MobLand season 2 come out?

MobLand season 2 will have a total of 10 episodes, with new entries airing weekly. That gives us the following schedule:

• Episode 1: September 18
• Episode 2: September 25
• Episode 3: October 2
• Episode 4: October 9
• Episode 5: October 16
• Episode 6: October 23
• Episode 7: October 30
• Episode 8: November 6
• Episode 9: November 13
• Episode 10: November 20

• Domestic kill switches are no longer feasible, UK government warns
• AI companies and government need to work together to reduce the risks
• OpenAI and 100+ other companies warn of "limited window to strengthen" defenses

The UK Government has reportedly rejected calls to create an emergency AI kill switch, per BBC reporting, and it's largely because there are simply too many parties and factors to consider at this stage.

The Cabinet Office told the BBC that it "cannot simply turn AI off" because preventing access to a model or shutting down certain infrastructure would not stop the same model being developed and hosted, thus misused, elsewhere.

This is in response to MPs pushing for a statutory trigger that the government could pull in the event of an emergency.

UK government rules out AI kill switch

The government also implied that the responsibility lies with AI companies, and that developers have their own responsibility to build products with the appropriate safeguards.

Former OpenAI researcher Danial Kokotajlo also agrees that one country building a legal or technical kill switch would not be feasible, but Kokotajlo does see companies working together as one potential solution.

"Switching off access to an AI model in an emergency will do little to protect you," Kokotajlo explained in relation to a domestic kill switch. "You're still going to be steamrolled by the super intelligences created in the US."

The latest kill switch debates have emerged from a September 1 discussion in the House of Lords, partly fueled by an open letter warning of a "limited window to strengthen cyber defenses." More than 100 companies, including OpenAI, Anthropic, Google, Microsoft, AWS and Adobe issued the stark warning, urging governments, cybersecurity companies and frontier AI developers to come together.

"Our security services and regulators possess no specific agile statutory mechanism to compel a physical or digital shutdown," Lord Clement-Jones warned earlier this month about the potential of a rogue AI.

As for what's next, while an AI kill switch may be technically infeasible, we can certainly hope that the companies and bodies at the forefront of AI and legislation collaborate more effectively to reduce future risks.

• Warlock: Dungeons & Dragons has a smaller spell set, each with multiple uses, for controller accessibility
• Invoke studio head Dominic Guay says this also means the combat won't slow down if there were a spell wheel
• He says there's also a bit of "resource management" to spells, and players will need to blend melee and magic combat

Invoke Studios has confirmed that Warlock: Dungeons & Dragons has a smaller spell set, with each spell offering multiple uses, rather than a large variety, mainly for controller accessibility.

That's according to studio head Dominic Guay, who told TechRadar Gaming in an interview at Gamescom that a smaller spell set opened up new playable opportunities, making combat more approachable on a controller while also keeping the action fast-paced.

This meant abandoning a classic spell wheel and narrowing its spell options, so they're all accessible on the gamepad.

"The breakthrough for us, it started with making direct access to these spells on the controller," Guay said. "A lot of games in the past where you had powers or spells, you had a lot of stuff to select from, and that's cool, and that was even our first reflex. But when you do that, you slow things down, and there's a limit to going through a wheel and choosing powers. There's a limit to that."

He added, "By making everything accessible on a controller, we started playing with our spells a lot more."

During the same interview, Guay explained why the team chose a warlock protagonist, saying "creative magic" was central to the story and gameplay.

The studio head also said it was a challenge to manage the game's magic system because, when you consider Dungeons & Dragons, "it can do everything," and Invoke needed to pick its spells wisely.

"Instead of adding like 100 spells, we thought, let's have less, which makes sense for a warlock also," he continued, "and let's make them more multiple-use cases, connect them, one to another. Connect them with some enemies. The way these enemies behave in a challenge."

Spells, which are anchored to magic objects, tomes, monsters, and more, also have their own upgradeable skill trees and can be used alongside melee combat. This creates a "rhythm," with Guay explaining that players can't just stick to long-range spells and will need to go into close quarters with enemies, which works well for protagonist Kaatri, a warrior-turned warlock.

Players can use spells to set up enemies, giving them an "advantage" for critical hits, and there are various ways to do this with different types of enemies.

"There's a bit of resource management too. You can't just do spells forever," Guay said. "There's a point where you're going to be low on ability to cast spells, so you're going to have to get in there at some point."

Warlock: Dungeons & Dragons launches in 2027 for PS5, Xbox Series X, Xbox Series S, and PC.

Lanterns episode 5 has landed on HBO Max — and it's not only the DC Universe (DCU) TV show's most devastating chapter so far for myriad reasons, but it also wraps up its 2016 storyline. From here on out, then, we'll be back in 2026, aka the DCU's current timeline.

Its present-day plot will have to wait another week, though, because we need to dissect all the hugely significant events that happened in Lanterns' fifth episode. Consider this your one and only warning: full spoilers immediately follow for Lanterns chapter 5, titled 'Lights Out'. If you haven't seen it yet, bookmark this page for later and return once you're caught up.

Who dies in Lanterns episode 5?

Frankly, lots and lots of people die. Innocent Rushville civilians caught in the crossfire and foot soldiers in Will Macon's militia and Antaan's extraterrestrial forces all bite the dust in the HBO Max show's latest installment.

Three notable individuals also lose their lives — the first two being the aforementioned Macon and Antaan. Macon is actually killed by Antaan, the latter believing the former is the Manhunter that Antaan's forces have been pursuing for some time.

However, after Antaan murders Macon and cuts out his heart to confirm Macon is the Manhunter — remember, a Manhunter's heart crystallizes once they die — he realizes Macon isn't the Manhunter. Cue Antaan going after Hal Jordan because the latter lied about the Manhunter's identity when the pair crossed paths earlier in episode 5.

Unfortunately for Antaan, he doesn't last much longer. In an attempt to kill Hal, he activates the explosive device woven into his skeleton. But before he goes out with a bang, trying to take Earth's first Green Lantern with him, Hal uses his ring to make a whirlwind construct that flings Antaan into the air before he detonates.

Of this episode's three big deaths, though, Zoe's is arguably the most significant.

Lanterns episode 4 finally revealed Zoe was the Manhunter, but she also appeared to meet her end in that chapter after her truck was destroyed by a lethal laser satellite attack. As it turns out, she survived that ordeal — one she instigated herself to try to trick Hal (with a little help from John Stewart) into thinking she was no longer a threat to him, the Green Lantern Corps, or the Guardians, which would finally make Hal leave Rushville.

It's ironic, then, that Stewart is the one who puts her down. The pair slept together twice in the DCU Chapter One show's first four episodes. Meanwhile, Zoe saved Stewart's life after he was critically injured when his rental car — you know, the one that Hal purposefully crashed in last week's entry — exploded. So, it stands to reason that Stewart wouldn't be the one who takes Zoe out.

Not so. Stewart shoots Zoe in the head and cuts out her heart, which crystallizes in his hands, so he can not only prove to the Guardians that the Manhunter has been destroyed, but also that he's worthy of being Hal's replacement. Speaking of which...

Why does John Stewart give up the Green Lantern ring?

Plainly, Stewart doesn't actually believe he's worthy of the Green Lantern mantle. That much is clear in his decision to give Hal's ring and lantern, which he wrestled off his mentor in a climactic duel, back to the Guardians of the Universe.

There are other reasons why Stewart makes this choice, though. Ultimately, he doesn't believe in working for a manipulative group of individuals like the Guardians, nor does he feel that, despite training to become a Green Lantern for as long as he can remember, it's his dream to join the space cop agency. Remember, he was pushed into this by his parents and, after plenty of measured introspection, it's evident Stewart doesn't want his parents to live vicariously through him anymore. As such, he takes ownership of his life and walks away from the lofty goal they've projected onto him since day one.

What will happen in Lanterns episode 6?

With 'Lights Out' officially bringing Lanterns' 2016 storyline to a close, all eyes will now turn to the sci-fi crime series' present-day plot.

First and foremost, the DC comic book show's sixth episode will need to pick things up around Hal Jordan's death. It feels like an age since Lanterns shockingly revealed Jordan's demise at the end of its first chapter, so it'll be interesting to see if next week's episode begins where the show's premiere left off, or if it'll go back prior to this pivotal event.

If the 2026 storyline starts before Jordan is murdered, we'll likely find out what he and Stewart have gotten up to in the decade since we last saw them. I wouldn't be amazed if we also learn what happened to the likes of Bill Macon, why Sheriff Kerry Kane stuck around Rushville after the town's near-decimation in 2016, and the wider impact that those explosive events had on those who had to pick up the pieces.

Last but by no means least, I wager we'll discover when and why Guy Gardner became Earth's new Green Lantern. It's clear he was chosen by the Guardians after Stewart refused to answer the call. However, we don't know how long he's been operating as a Green Lantern or, based on a clip shown on a recent episode of The Jimmy Kimmel Show, how Gardner and Stewart know each other. Over to you, Lanterns.

• Sony corrected reports claiming its disc production would collapse by 90%
• The company's final disc plant will reduce output by 10%
• Sony still plans to stop producing new physical games in 2028

Sony has denied reports suggesting it plans to abandon physical disc manufacturing entirely, despite widespread speculation surrounding its future operations.

The company instead confirmed that its sole remaining disc-making plant will simply reduce output by 10%, not the previously reported 90%.

Sony DADC clarified the figure after earlier remarks caused considerable confusion among industry observers.

Clearing up a misread statement

“To clarify and avoid any misleading information: in that statement Dietmar anticipated an overall product volume decline by 10 percent, not a decline down to 10 percent,” said a Sony DADC spokesperson.

The correction arrived roughly two months after earlier reports claimed the plant was already being repurposed away from discs entirely.

Sony has consistently maintained that new physical game production will still end entirely in January 2028, despite this week's clarification.

The clarification means Sony's Thalgau plant will still operate at nearly full capacity beyond the original 2028 cutoff for releases.

Sony DADC did not respond to a separate fact-check request regarding whether all 300 plant employees face retraining toward alternative roles.

Reports from earlier this year suggested employees were being shifted away from discs toward microlens production at the same facility.

Discs survive, but ownership questions remain

Sony confirmed publishers may continue reordering existing disc-based games even after new physical production formally ceases in January of 2028.

PlayStation disc games already on shelves will remain available to buy for the foreseeable future.

This arrangement suggests the transition away from physical media will happen gradually rather than through an abrupt cutoff for publishers.

Even so, wider industry trends continue pointing toward digital distribution gradually replacing physical discs across most major platforms.

Several major publishers remain notably absent from Microsoft's own separate disc-to-digital initiative, raising questions about industry cooperation levels.

Sony recently argued in court that consumers cannot reasonably believe they truly own purchased games, since copies can exist simultaneously.

This legal stance, combined with the recent production clarification, still leaves genuine uncertainty around long-term disc availability for players worldwide.

Sony's carefully worded statement corrects one misunderstanding but does little to settle deeper doubts about ownership rights among consumers today.

Whether the Thalgau plant continues producing discs well beyond 2028 also remains unclear, given how quickly assumptions were proven wrong.

Consumers hoping for full clarity on Sony's long-term disc strategy will likely need to wait, since key details remain unconfirmed.

For now, the disc format appears to be shrinking gradually rather than disappearing suddenly, as many observers had initially feared.

Via The Verge

• NATO disrupted Russia’s Arctic cable exercise before the reported weapon could be deployed
• Russian submarines rehearsed attacks that could leave little evidence afterward
• Two 1,400km cables connect Svalbard with mainland Norway beneath Arctic waters

NATO allies disrupted a Russian naval exercise near Svalbard after detecting preparations involving a secret device intended for damaging subsea communications infrastructure.

British, Norwegian and American forces confronted vessels linked to Russia’s deep-sea research directorate during exercises conducted in Arctic waters this spring.

The reported operation did not damage any cable, but officials said the equipment could remove evidence that would identify its use after an attack.

NATO confronted Russian vessels during Arctic cable exercise

The equipment was reportedly being tested by GUGI, a Russian military organization responsible for specialized underwater operations and deep-ocean missions.

Officials said the device could cut communications lines while leaving little physical evidence capable of connecting Moscow to the incident.

The confrontation occurred near Svalbard, where two fiber-optic cables extend for about 1,400km and descend roughly 2,700 meters underwater.

Those lines support data transfers from SvalSat, a major satellite ground facility that also serves NASA’s Near Space Network in orbit.

The cables are only around garden-hose thickness, yet they support services carrying enormous quantities of information between the Arctic territory and Norway.

Norway is developing another connection involving Svalbard and Jan Mayen, with completion expected during 2028 under its current infrastructure plans.

The location matters militarily because vessels from Russia’s northern naval facilities must cross the Bear Gap when entering Atlantic waters.

That passage spans roughly 400 miles, making nearby infrastructure significant for communications and surveillance across the wider High North region.

Why NATO is concerned about the cable-cutting capability

Western intelligence officials have become increasingly concerned that Moscow could expand covert disruption against critical infrastructure across Europe’s vulnerable networks.

Three American officials told Reuters that subsea communications lines could feature in a future attempt to test NATO’s collective-defense commitment.

CIA Director John Ratcliffe travelled to Moscow in August for discussions that included warnings about escalating sabotage activity in Europe.

Russia has rejected accusations that it conducts sabotage inside NATO countries or seeks direct military confrontation with the alliance, but Norway is not having it.

"Through our joint operation, we sent a clear message to Russia that they cannot operate covertly," Norway's Defense Minister Tore Sandvik declared in a statement.

"We have made it unequivocally clear to Russian authorities that any attempt to target our critical infrastructure will be detected and met with consequences."

Earlier British disclosures also linked Russian deep-sea submarines with activity near British waters involving energy pipelines and telecommunications systems.

Britain subsequently allocated another $136 million toward P-8 maritime patrol aircraft as suspicious Russian vessel movements increased around vulnerable infrastructure.

Undersea cables carry about 99% of Britain’s international telecommunications and data traffic, raising concerns about disruptions to essential public services.

British and Norwegian officials have carried their findings to Moscow, apparently hoping public disclosure would discourage any future use of the technology.

However, reports so far do not establish that Russia has used this device against a cable, but rather that it has rehearsed its deployment.

Via United24media

• StarCraft is making a shift from the real-time strategy genre to the open-world shooter genre
• The development team realised during early discussions that "we weren't talking about a game from above"
• According to StarCraft general manager Dan Hay, "it was boots on the ground every time" when the team was talking about the game

The process of StarCraft shifting from real-time strategy to open-world shooter emerged naturally from repeated team discussions about "the story we wanted to tell," as well as in-game moments, armor, exploration, discovery, and the desire to experience the world from ground level, according to the game's general manager, Dan Hay.

In an interview with TechRadar Gaming at BlizzCon 2026 in Anaheim, Hay went on to describe the process that will see the series' perspective change entirely in the next StarCraft game: "Every time we drew a picture or every time we talked about it, it was boots on the ground every time."

These conversations then snowballed, gathering up other game features and elements, it seems, which kept the direction of travel going towards taking StarCraft into being a shooter. "We started to talk about perspective and how that's going to work; we started to talk about the armor, and we started to talk about what could attack that armor and how it could rip it apart if it needed to.

"We would talk about things with the characters [too], and we started to very quickly realize we weren't talking about a game from above; we were talking about a game down on the ground where you felt every inch of the game," described Hay.

He added, giving a nod to the narrative in store for players when StarCraft releases, that it was obvious that the game would be "down on the ground, and that just made sense for the story."

But as well as changing the player's perspective from looking down to looking out and around, the team looked at this change for the series to embrace the open-world exploration and discovery aspects of the upcoming game.

"[We wanted to] make sure that there is an opportunity to explore. Make sure that there's an opportunity to discover. See a thing off in the distance and go — 'is that something that I saw from the first game? Is it?' And then head off to it."

StarCraft was teased with a cinematic trailer at BlizzCon 2026 and is slated to release in 2030.

• Two AMD cards cost $18,775 yet beat GPT-5.6 Sol within hours weekly
• Multi-Token Prediction nearly doubled throughput to 320.2 tokens every second
• Twenty million monthly tokens save a team $11,738 yearly against Sol pricing

A hardware reviewer compared a dual-GPU AMD workstation against cloud subscription pricing to determine which option delivers cheaper AI inference over time.

Two AMD Radeon AI PRO R9700 cards, each carrying 32 GB of memory, were installed inside a workstation costing roughly $18,775 as tested.

The evaluation measured electricity draw, token throughput, and amortized hardware cost, then set those figures against several cloud subscription tiers for comparison.

How cloud pricing sets the bar

Cloud AI providers charge customers per million tokens generated, with prices ranging from $1.20 for GPT-5.6 Luna up to $30 for GPT-5.6 Sol.

Mid-tier models sit in between, with Claude Sonnet 5 at $10 and Claude Opus 5 at $25 per million output tokens generated.

The more expensive the cloud model a team would otherwise use, the sooner owned hardware pays for itself.

Testing both AMD cards together, the workstation generated 156.2 tokens every second while serving eight simultaneous users during this test.

A speed technique called Multi-Token Prediction nearly doubled that figure, pushing throughput up to 320.2 tokens every second, with identical output quality.

At that 320.2 token-per-second speed, the machine only needs 3.5 hours of weekly use to beat GPT-5.6 Sol on cost, 4.2 hours to beat Claude Opus 5, and 8.8 hours to beat Gemini 3.1 Pro.

A team generating above 20 million tokens per month against GPT-5.6 Sol pricing gains real savings using this owned hardware setup.

At that volume, running the workstation costs about $6,262 yearly in electricity and amortized hardware, against roughly $18,000 yearly in matching Sol fees.

That $11,738 yearly gap is the actual evidence behind the claim that heavy monthly usage makes AMD's rig worthwhile.

If a company instead relies on GPT-5.6 Luna, priced at just $1.20 per million tokens, that math flips entirely in the other direction.

The workstation would then need 94.3 hours of weekly use just to match that far cheaper cloud subscription's total cost.

Since a single week only contains 168 hours total, reaching that particular break-even point remains genuinely difficult without near constant, saturated usage.

Electricity itself was a minor factor throughout, since both cards together drew between 310 and 510 watts under sustained load conditions.

Where the economics tip in AMD's favor

A smaller team producing only five million tokens monthly, priced against Gemini 3.1 Pro, would spend $6,262 yearly to displace just $720 in cloud costs, a clear loss.

That comparison shows the hardware only makes financial sense once usage climbs high enough to close a large yearly cost gap.

A single R9700 card handled an eight billion parameter AI model alone, processing 34.5 tokens every second without help from a second card.

Running just one card also lowers the effective break-even point further, since a lone card draws far less power under equivalent load.

That single card pulled between 221 and 283 watts depending on simultaneous user count, well under the 310 to 510 watts both cards drew together.

Smaller models therefore offer a second path into positive economics, letting lighter workloads justify a $1,880 single-card purchase long before a team can justify the full $3,760 dual-card upgrade.

Price alone still does not settle the question, since these locally run models measurably trail top cloud systems on complex reasoning benchmarks.

On one independent intelligence index, the 27 billion parameter AMD model scored 37 points against Google's Gemini 3.1 Pro scoring 46 points instead.

Therefore, a team chasing the cheaper token count may be trading away real reasoning quality, not just cloud subscription fees.

Via Puget Systems

Quordle was one of the original Wordle alternatives and is still going strong now more than 1,500 games later. It offers a genuine challenge, though, so read on if you need some Quordle hints today — or scroll down further for the answers.

Enjoy playing word games? You can also check out my NYT Connections today and NYT Strands today pages for hints and answers for those puzzles, while Marc's Wordle today column covers the original viral word game.

SPOILER WARNING: Information about Quordle today is below, so don't read on if you don't want to know the answers.

Quordle today (game #1694) — hint #1 — Vowels

* Note that by vowel we mean the five standard vowels (A, E, I, O, U), not Y (which is sometimes counted as a vowel too).

Quordle today (game #1694) — hint #2 — repeated letters

Quordle today (game #1694) — hint #3 — uncommon letters

Quordle today (game #1694) — hint #4 — starting letters (1)

If you just want to know the answers at this stage, simply scroll down. If you're not ready yet then here's one more clue to make things a lot easier:

Quordle today (game #1694) — hint #5 — starting letters (2)

Right, the answers are below, so DO NOT SCROLL ANY FURTHER IF YOU DON'T WANT TO SEE THEM.

Quordle today (game #1694) — the answers

The answers to today's Quordle, game #1694, are…

• ZEBRA
• LEGAL
• RATIO
• TROLL

A very rare outing for the letter Z and the magnificent striped animal ZEBRA.

Before this, though, I had made yet another silly mistake on what became LEGAL, placing A in a position that had already been ruled out.

My excuse is I’m playing Quordle with a heavy cold, but I really should know better by now.

Daily Sequence today (game #1694) — the answers

The answers to today's Quordle Daily Sequence, game #1694, are…

• WRECK
• COMIC
• MAUVE
• OLIVE

Quordle answers: The past 20

• Quordle #1693, Sunday, 13 September: LUCID, SPRAY, MANGE, EMAIL
• Quordle #1692, Saturday, 12 September: SUPER, BRUSH, RESET, SOWER
• Quordle #1691, Friday, 11 September: PLUSH, PAINT, RIVAL, AFOUL
• Quordle #1690, Thursday, 10 September: AGENT, EXILE, TITAN, PUREE
• Quordle #1689, Wednesday, 9 September: HAPPY, GONER, SMACK, MOTIF
• Quordle #1688, Tuesday, 8 September: FLUNG, DIARY, HALVE, BERTH
• Quordle #1687, Monday, 7 September: VISIT, CLANK, FEVER, OCEAN
• Quordle #1686, Sunday, 6 September: PILOT, HONOR, BOWEL, TAUNT
• Quordle #1685, Saturday, 5 September: GNASH, GEEKY, WHELP, PUTTY
• Quordle #1684, Friday, 4 September: QUICK, BEEFY, TRUCK, SNAIL
• Quordle #1683, Thursday, 3 September: SIGMA, ESSAY, MERRY, LLAMA
• Quordle #1682, Wednesday, 2 September: INCUR, FELLA, TITAN, FROZE
• Quordle #1681, Tuesday, 1 September: NASAL, ROUTE, ANKLE, MIGHT
• Quordle #1680, Monday, 31 August: ALIVE, LOUSE, ABOVE, MOSSY
• Quordle #1679, Sunday, 30 August: USING, SORRY, FILET, PIPER
• Quordle #1678, Saturday, 29 August: OVINE, ALLOY, VOICE, CARVE
• Quordle #1677, Friday, 28 August: WAIST, EDICT, BONUS, DEUCE
• Quordle #1676, Thursday, 27 August: SHEEN, TRICE, WAGON, BEVEL
• Quordle #1675, Wednesday, 26 August: OPIUM, EPOCH, BESET, FINER
• Quordle #1674, Tuesday, 25 August: BLESS, FELON, PORCH, CHOSE

Good morning! Let's play Connections, the NYT's clever word game that challenges you to group answers in various categories. It can be tough, so read on if you need Connections hints.

What should you do once you've finished? Why, play some more word games of course. I've also got daily Strands hints and answers and Quordle hints and answers articles if you need help for those too, while Marc's Wordle today page covers the original viral word game.

SPOILER WARNING: Information about NYT Connections today is below, so don't read on if you don't want to know the answers.

NYT Connections today (game #1191) - today's words

Today's NYT Connections words are…

• STRAW
• CRAB
• COOT
• LEMON
• BLUE
• CANARY
• TEAL
• RAT
• LOON
• BAG
• WHEAT
• CUTLERY
• SNITCH
• GOOSE
• STIRRER
• STOOLIE

NYT Connections today (game #1191) - hint #1 - group hints

What are some clues for today's NYT Connections groups?

• YELLOW: Aquatic fowl
• GREEN: Tell-tale
• BLUE: Disposable things
• PURPLE: Add a green word

Need more clues?

We're firmly in spoiler territory now, but read on if you want to know what the four theme answers are for today's NYT Connections puzzles…

NYT Connections today (game #1191) - hint #2 - group answers

What are the answers for today's NYT Connections groups?

• YELLOW: WATERBIRDS
• GREEN: INFORMANT
• BLUE: COMMON SINGLE-USE PLASTIC ITEMS
• PURPLE: ____GRASS

Right, the answers are below, so DO NOT SCROLL ANY FURTHER IF YOU DON'T WANT TO SEE THEM.

NYT Connections today (game #1191) - the answers

The answers to today's Connections, game #1191, are…

• YELLOW: WATERBIRDS: COOT, GOOSE, LOON, TEAL
• GREEN: INFORMANT: CANARY, RAT, SNITCH, STOOLIE
• BLUE: COMMON SINGLE-USE PLASTIC ITEMS: BAG, CUTLERY, STIRRER, STRAW
• PURPLE: ____GRASS: BLUE, CRAB, LEMON, WHEAT

• My rating: Hard
• My score: 3 mistakes

This was a really difficult game for me, although thankfully the frustration didn’t extend beyond the first group.

Seeing COOT and GOOSE I automatically knew we were looking for bird-themed tiles, but I didn’t see that we were looking for WATERBIRDS and kept on trying to include CANARY. I got there eventually though, after remembering that TEAL was a type of duck.

From here, I completed the other groups in difficulty order — although after looking at the four groups afterwards, purple seemed the easiest of the lot.

Yesterday's NYT Connections answers (Sunday, September 13, 2026, game #1190)

• YELLOW: LACKING IN FIRMNESS: DROOPY, FLACCID, SAGGING, SLACK
• GREEN: U.S. PRESIDENTS: CLEVELAND, GARFIELD, GRANT, MADISON
• BLUE: WORDS THAT SOUND LIKE TWO LETTERS: EMPTY, EXCESS, KEWPIE, ODIE
• PURPLE: ENDING IN ELEVATED LANDFORMS: DOUBLESPEAK, HEATHCLIFF, PARAMOUNT, PORRIDGE

Strands is the NYT's latest word game after the likes of Wordle, Spelling Bee and Connections – and it's great fun. It can be difficult, though, so read on for my Strands hints.

Want more word-based fun? Then check out my NYT Connections today and Quordle today pages for hints and answers for those games, and Marc's Wordle today page for the original viral word game.

SPOILER WARNING: Information about NYT Strands today is below, so don't read on if you don't want to know the answers.

NYT Strands today (game #925) - hint #1 - today's theme

NYT Strands today (game #925) - hint #2 - clue words

Play any of these words to unlock the in-game hints system.

• DUCK
• LICE
• SHUCK
• TART
• DAZE
• FRESH

NYT Strands today (game #925) - hint #3 - spangram letters

NYT Strands today (game #925) - hint #4 - spangram position

Right, the answers are below, so DO NOT SCROLL ANY FURTHER IF YOU DON'T WANT TO SEE THEM.

NYT Strands today (game #925) - the answers

The answers to today's Strands, game #925, are…

• EXCLAIM
• START
• GASP
• RECOIL
• FREEZE
• SHUDDER
• SPANGRAM: THATSSHOCKING

• My rating: Easy
• My score: Perfect

It’s brilliant how the exclamation OH MY GOSH!! has had a second life in recent years after being consigned to the pages of dusty children’s adventures for decades, alongside “fiddlesticks” and “darn”.

The phrase actually dates back to the 18th century — the word gosh has no meaning at all beyond being an altered version of “god”.

Here, though, it was our clue to a pleasing game of Stands searching and for me my first 'spangram first' in months — gosh darn it I’m as jolly as a bumblebee in a buttercup.

Yesterday's NYT Strands answers (Sunday, September 13, game #924)

• MOLE
• AGENT
• ASSET
• PLANT
• COUNTERSPY
• OPERATIVE
• SPANGRAM: UNDERCOVER

• Meta wants nearly 1,000 AI GPUs operating inside one coherent data center domain
• Panmnesia's CXL design connects CPUs, accelerators, and memory across multiple racks
• The architecture increases accelerator coordination from two devices to sixteen per CPU

Meta is working with Panmnesia on an artificial intelligence data center design intended to make thousands of processors operate within one coherent environment.

The proposal uses Compute Express Link, or CXL, to connect CPUs, accelerators and memory across multiple racks without conventional network links.

The architecture could bring as many as 960 AI accelerators into one coherence domain, putting its capacity at almost 1,000 GPUs working as one system.

CXL architecture extends accelerator connectivity

The design addresses a problem that becomes increasingly difficult as AI training systems combine hundreds or thousands of accelerators processing enormous data.

Every accelerator must progress through repeated computational stages, meaning one delayed component can force other devices to wait before continuing their work.

The researchers therefore focus on reducing unpredictable communication delays between racks, where Ethernet or InfiniBand networks normally handle connections beyond individual systems.

Those networks require packet processing and software coordination, which can introduce greater latency variation as workloads spread across additional servers.

CXL instead provides a shared coherence mechanism, allowing processors, accelerators and memory to participate within one connected resource environment.

The proposed architecture adds dedicated hardware intended to keep communication paths and processing behaviour more consistent across the larger fabric.

Panmnesia's design uses a high-fan-out switch, a link acceleration unit and a fabric controller to manage traffic across the system.

These components are organized into trays, pods and a fabric, borrowing organizational principles normally associated with arranging functional blocks inside semiconductor chips.

The company says its fabric controller and link acceleration unit have completed silicon validation, while its switch has already been fabricated.

Its switch has also been fabricated, while pre-release silicon is reportedly being supplied as development continues toward commercial products.

Up to 960 accelerators in one coherence domain

The review compares the proposed arrangement with NVIDIA's GB200 NVL72, where one CPU directly coordinates two accelerators through NVLink-C2C.

Under Panmnesia's architecture, one CPU could coordinate 16 accelerators, representing an eightfold increase over that reference configuration.

According to the published design, around 60 such groups could then form a coherence domain containing approximately 960 accelerators.

Cross-rack access could also fall from microsecond-level timing to several hundred nanoseconds, representing roughly an order-of-magnitude reduction.

The architecture would allow individual failed devices to be replaced without taking an entire server out of service.

That separation could reduce the amount of functioning hardware removed during failures, although actual operational benefits would depend on implementation.

During the announcement, Myoungsoo Jung, CEO of Panmnesia, said, “CXL enables the entire datacenter to operate as a single computing system.”

The architecture still faces physical limits because electrical CXL signalling reaches only about seven meters at 128 GT/s with two retimers.

Panmnesia therefore proposes optical CXL links for longer distances, and says it has already completed hardware proof-of-concept validation for that approach.

• FOIA documents revealed disputed Pentagon plans for a less restrictive AI system
• OpenAI denied agreeing to language requiring minimal refusal rates from models
• A Pentagon lawyer later withdrew an earlier confirmation about the contract

Documents released through a Freedom of Information Act (FOIA) lawsuit reportedly revealed that the Pentagon sought out an OpenAI system with fewer restrictions.

The paperwork described models intended for national security work that would have “minimal refusal rates” during certain military applications.

OpenAI and the Pentagon have both disputed that such wording appeared in the final agreement signed between them.

Dispute emerges over AI restrictions in military contract documents

The disputed language appeared in a contract document describing work involving testing, evaluation, and improvement of OpenAI models for defense-related activities.

The document referred to specialized models created for government security uses, including systems designed to reject fewer requests during operation.

However, officials from both organizations later said the document obtained through the lawsuit was not the final contract version.

"OpenAI has never agreed to contract language requiring 'minimal refusal rates.' This language does not appear in our executed contract," said Nate Evans, an OpenAI spokesperson.

“The document you received appears to be an earlier draft proposed by the Department before we provided feedback.”

Evans added that OpenAI rejected the wording and said the final executed agreement did not contain that requirement.

A Department of Defense representative later stated that the phrase “minimal refusal rates” does not appear in any active contract.

The disagreement continued after a Pentagon lawyer initially confirmed the document before later withdrawing that statement.

Trevor Tiedeman, a Pentagon research official, said there may have been confusion between released records and executed agreements.

Tiedeman also indicated he would examine how the document passed through review procedures before becoming publicly available.

The original document was released after The Intercept requested contract records connected with military artificial intelligence agreements.

The request specifically sought completed agreements, while officials later argued the released material represented an earlier version.

OpenAI did not provide the complete contract publicly, leaving some details about the agreement unavailable for independent review.

Concerns grow over military AI safeguards and operational limits

The dispute has renewed discussion about how much control technology companies should retain over artificial intelligence used by governments.

Commercial AI systems typically include restrictions designed to prevent assistance with certain harmful activities, including weapons development.

OpenAI's public chatbot, for example, refuses requests involving assistance with planning specific attacks or selecting targets for strikes.

“Minimal refusal is likely referring to little or no safeguards on the model being used,” said Heidy Khlaaf, chief scientist at the AI Now Institute.

She argued that private technology companies face difficult decisions when their systems become connected with military operations.

The Pentagon has also faced separate disagreements with other AI firms over restrictions involving defense applications.

Despite those disputes, OpenAI expanded its military agreement, allowing use of its services across classified government networks.

The updated agreement retained references to mission-focused AI models, although significant sections of the contract were redacted.

OpenAI has said its agreement includes restrictions against autonomous weapons use and certain surveillance activities involving Americans.

The company and government agencies have not fully disclosed how those restrictions would apply across different military scenarios.

Via The Intercept

The clock is ticking on Ted Lasso season 4's run on Apple TV. Only four more episodes remain of the soccer comedy-drama's latest season, so it doesn't have many more opportunities to kick a ball in anger and/or celebrate any goals.

That's a worry for another day, though. Right now, you just want to know when its seventh episode will be released, so read on to see when it'll premiere on one of the world's best streaming services.

What is the launch time for Ted Lasso season 4 episode 7?

Ted Lasso season 4 episode 7 is expected to drop at 6pm PT / 9pm ET on Tuesday, September 15 in the US and Canada.

Apple has always stated that new entries will air every Wednesday at 12am PT in North America. However, every one of this season's chapters has debuted at the aforementioned times on the past six Tuesday evenings, so it stands to reason that this week's installment will do the same.

Meanwhile, those of you who live elsewhere will have to wait until Wednesday, September 16 to catch this week's episode, which is titled 'Yes and Baby'. Read on to see what time it'll launch where you are:

• US — Tuesday, September 15 at 6pm PT / 9pm ET
• Canada — Tuesday, September 15 at 6pm PT / 9pm ET
• UK — Wednesday, September 16 at 2am BST
• India — Wednesday, September 16 at 6:30am IST
• Singapore — Wednesday, September 16 at 9am SGT
• Australia — Wednesday, September 16 at 11am AEST
• New Zealand — Wednesday, September 16 at 1pm NZST

When can I watch the final three episodes of Ted Lasso season 4?

Unless Apple makes a sudden adjustment to season 4's release schedule, new episodes will air every Tuesday in North and South America and every Wednesday on other continents until the finale is released in early October.

Check out the list below for the remaining episodes' all-important launch dates:

• Ted Lasso season 4 episode 1 — out now
• Ted Lasso season 4 episode 2 — out now
• Ted Lasso season 4 episode 3 — out now
• Ted Lasso season 4 episode 4 — out now
• Ted Lasso season 4 episode 5 — out now
• Ted Lasso season 4 episode 6 — out now
• Ted Lasso season 4 episode 7 — September 15/16
• Ted Lasso season 4 episode 8 — September 22/23
• Ted Lasso season 4 episode 9 — September 29/30
• Ted Lasso season 4 episode 10 — October 6/7

• AI memory demand is making even broken smartphones increasingly valuable to buyers
• Old phones are gaining value as replacement components become harder to source
• Grade D devices remain abundant despite stronger prices for salvageable hardware

A global shortage of memory chips driven by artificial intelligence demand is pushing millions of consumers toward the resale market.

Owners who once let broken or outdated smartphones sit in drawers are now finding that even damaged units carry unexpected monetary value.

Buyers across the refurbishment and repair industries are competing for parts that were previously considered worthless scrap.

Salvage values climb as parts become scarce

Marketplace figures from B-Stock, a large business-to-business platform for used smartphones, show pricing for both Apple and Samsung devices rose in the second quarter of 2026 compared with the prior two quarters.

Grade D handsets, the lowest condition tier, made up the bulk of units sold last year, with millions changing hands through the platform.

In the first half of 2026, mid-tier C-condition phones actually overtook the lowest tier by volume, accounting for 37.8% of sales versus 31.6% for the bottom grade.

The iPhone 17 Pro Max fetched the strongest average price among Apple models, reaching $940 across all condition levels during the quarter.

Samsung's top-selling models were the S23 Ultra, the S22 Ultra and the S24 Ultra, with the S22 Ultra pushing the older Z Fold5 out of second place.

After the Galaxy S26 launched, prices for previous-generation Samsung flagships climbed as well, with a well-kept S23 Ultra reaching $425 in June, its strongest month of the year.

Repair industry sees new value in broken parts

B-Stock has cautioned against assuming shoppers are simply rushing to dump unwanted hardware.

The company points out that people are keeping their phones for extended periods, a pattern that itself drives up the volume of low-grade listings as devices age.

Manufacturers are not yet routinely pulling memory chips from damaged phones to install in new products, since testing, compatibility and reliability checks complicate that approach.

According to the company, the economics of recovering usable components are shifting as memory prices climb, making it worthwhile to salvage parts that once were not worth the labor involved.

Displays, camera assemblies, batteries, charging ports, housings, speakers, cables and circuit boards can all retain value even when a phone itself is beyond repair.

A single low-grade device does not necessarily mean every part inside it shares that same poor rating.

A phone traded in after four or more years often fails battery health testing alone, which is enough to place it in the bottom tier under standard platform rules.

The current trend feels less like a sudden environmental awakening and more like a market correction.

Consumers are not necessarily recycling out of ecological conviction, and retailers are sweetening trade-in offers to keep the supply of parts flowing.

Whether the shift meaningfully reduces electronic waste, or simply redirects it toward a more profitable pipeline, remains an open question worth watching.

I'm sad that Reacher season 4 is coming to an end, but thanks to some expert scheduling from Amazon's content team, it's not all bad news.

Just as the popular series' fourth season comes to a close, its highly-anticipated spinoff Neagley takes its place. Maria Sten reprises her titular role from Reacher's first three seasons, and is now working as a private investigator in Chicago. As it happens, she quickly learns that a beloved friend from her past has been killed in a suspicious accident and becomes hell-bent on justice.

So far, so good, but when does Neagley arrive on Prime Video?

What time can I watch Neagley on Prime Video?

We're expecting Neagley to be available on September 16 at 12am PT / 3am ET in the US and Canada. That's when Amazon TV Originals usually make their debuts on Prime Video, so there's no reason to suspect Neagley won't do likewise.

Internationally, look out for the below times:

• US – 12am PT / 3am ET
• Canada – 12am PT / 3am ET
• UK – 8am BST
• India – 12:30pm IST
• Singapore – 3pm SGT
• Australia – 5pm AEST
• New Zealand – 7pm NZST

When do new episodes of Neagley come out?

All eight episodes will be released at the same time. Unlike weekly episodic drops for other Prime Video TV series, then, you won't need to wait around for new installment of Neagley.

As for the series' future, there's currently no public word on whether a second season has been greenlit. Hopefully, we won't be made to wait much longer for confirmation about its potential return.

• Watch US Open 2026 Men's Final free on 9Now (AUS)
• Unlock your stream with NordVPN (save up to 75%)
• Final starts 2pm ET / 7pm BST

Watch Alexander Zverev vs Ben Shelton in the 2026 US Open men's final as the American takes to Arthur Ashe Stadium in front of his home crowd looking to become the first American man to win a Grand Slam singles title since Andy Roddick at the 2003 US Open.

Zverev, meanwhile, comes into the final on a hot streak, reaching his third consecutive Grand Slam final after winning Roland Garros earlier this year before losing to Jannik Sinner in the Wimbledon final. The top seed made a slow start in New York, needing five sets in each of his opening two rounds, but has since won 13 consecutive sets, including straight-set victories over Botic van de Zandschulp in the quarterfinals and Karen Khachanov in the semis.

Although Zverev arrives as the slight favorite, particularly after beating Shelton in all five of their previous meetings, the American southpaw has arguably been the more impressive player – and faced the tougher path – this fortnight. He took down defending champion Carlos Alcaraz in a five-set quarterfinal that finished at 3:33am, before beating fellow American Frances Tiafoe in the semis to reach his first Grand Slam final.

The best part? We've found a sneaky way to watch the final completely FREE, from wherever you are!

The hack: 9Now in Australia is showing the match at no cost!

But what if you're traveling outside the country? Can you still watch the 2026 US Open men's final for free - from the US, UK or Canada? Yes, you can! Read on to find out how.

How to watch US Open 2026 Men's Final for free

The US Open 2026 Men's Singles Final will be shown live and for free in Australia on Channel 9 and 9Gem TV channels, with a free live stream on 9Now.

Although there are no direct free streams in the US, tennis fans can use a free trial of a live TV streaming service to watch the final at no charge. These include YouTube TV (10-day free trial), DirecTV Stream (5-day free trial) and Hulu+Live TV (3-day free trial).

Traveling abroad right now? You can use a VPN to watch the US Open 2026 Final for free as if you were right at home.

Use a VPN to watch US Open 2026 Men's Final free from anywhere

A VPN is handy piece of software that can make your device appear as if it's back in your home country, so you can unlock your usual service. The best VPN right now? We recommend NordVPN – it does everything and comes with a 75% discount, and three months for free.

Quick start: Using a VPN to watch US Open 2026 Men's Final for free

Using a VPN is incredibly simple, just follow these steps.

1. Install the VPN of your choice. As we've said, NordVPN is our favorite.

2. Choose the location you wish to connect to in the VPN app. For instance if you're visiting the U.S. and want to view an Aussie streaming service, you'd select a location from Australia (like Perth) from the server list.

3. Sit back and enjoy the action. Head to 9Now and watch the 2026 US Open Men's Final just like you would at home.

Which devices can I watch US Open 2026 Men's Final for free with?

You can use 9Now to watch the 2026 US Open Men's Final on a range of devices:

Amazon Fire TV
Android phones and tablets (Android OS 5 and newer)
Android TV
Apple iPhones and iPads (iOS 12 or later)
Apple TV (including Apple TV 4K and Apple TV HD)
Chromecast (including Chromecast with Google TV)
Chrome (latest two versions)
Fetch TV (Gen 3 or newer)
Firefox (latest two versions)
Foxtel iQ
Freeview
Hisense TV
LG Smart TV (webOS 4 or higher)
Mobile apps – download the 9Now app on your phone or tablet (iOS/Android)
Mozilla (latest two versions)
Roku (including Roku 2, Roku 3, Roku Express, Roku Express+, Roku Premiere, Roku Premiere+, Roku Streaming Stick, Roku Streaming Stick+ and Roku Streambar)
Safari (latest two versions)

You may also enjoy...

• How to watch US Open 2026 for free
• Tune into Sabalenka vs Rybakina
• How to watch 9Now from outside Australia

We test and review VPN services in the context of legal recreational uses. For example:1. Accessing a service from another country (subject to the terms and conditions of that service).2. Protecting your online security and strengthening your online privacy when abroad.We do not support or condone the illegal or malicious use of VPN services. Consuming pirated content that is paid-for is neither endorsed nor approved by Future Publishing.

• Anduril and Poland's PGZ are preparing local production of the new Barracuda cruise missile
• Anduril displayed a full-scale Barracuda mock-up during Poland's major defense exhibition
• Production details remain unsettled as Poland and Anduril continue negotiations

Anduril has partnered with Poland's state-owned conglomerate PGZ to jointly produce a new cruise missile system.

Dubbed Barracuda-500M, the ground-launched weapon will reportedly be manufactured locally as part of a long-term arrangement between both companies.

The missile deal comes as Warsaw continues to expand its arsenal amid rising concern over Russian military activity near NATO's eastern border.

Barracuda-500M production takes shape

A full-scale mock-up of the missile was placed on public display for visitors attending this week's defense trade event held in Kielce, Poland.

The unveiling took place during the annual MSPO defense exhibition, where company representatives outlined the scope of the missile partnership.

The partnership covers both local assembly and eventual full-scale production of the missile inside Poland.

Company representatives have not yet said whether components will initially be shipped from the United States for domestic assembly work.

They have also not yet publicly disclosed a timeline for full-scale production or for when the missile might become operational.

Autonomous aircraft added to the pitch

Separately, Anduril is also pitching two autonomous aircraft designed to operate alongside Poland's future F-35 fighters and AH-64E Apache helicopters.

Brian Moran, who oversees Anduril's business operations across Europe, outlined the pitch during a separate interview at the MSPO show.

According to Moran, the Fury combat drone could accompany Poland's incoming batch of 32 F-35 fighters as a loyal wingman.

He explained that the aircraft has already been chosen by the U.S. Air Force for a formal, funded procurement effort.

"Fury is tailor-made to partner with an F-35 as a loyal wingman and as a force multiplier to keep men and women in uniform out of harm's way," Morgan said.

The Thunder rotorcraft, a separate autonomous platform built by the company, is designed specifically to operate alongside Poland's incoming fleet of 96 Apache helicopters.

Thunder was first unveiled publicly during the Farnborough Airshow held this past July, several months before Anduril intensified its talks with Polish defense officials.

"In July, we launched Thunder, which is an autonomous collaborative platform for the Apache attack helicopter," said Moran.

"And just like Fury operating with an F-35, with Poland getting all these Apaches, having Thunder is a force multiplier, especially as Poland will have to train a lot of pilots for 96 Apaches."

Moran noted that Poland is set to become the second-largest operator of Apache helicopters anywhere on the globe, trailing only the United States.

Delivery of all 96 helicopters, along with their associated equipment and weapons systems, is scheduled to formally begin during 2028.

That gap gives the Polish military several years to train pilots and prepare support systems before the Apaches actually arrive.

Beyond the 96 helicopters, Poland is also trying to expand its arsenal with ongoing discussions with Anduril regarding a possible purchase of the Thunder drone.

Via Defense News

• Microsoft and two key US teaching federations have agreed safeguards on AI use within education across America
• The American Federation of Teachers and the United Federation of Teachers have made the deal with Microsoft after months of negotiations with all parties
• The AFT is in talks with other companies including OpenAI and Anthropic

After the failure to pass federal legislation to establish protections against AI overreach in schools, two teaching federations have taken the step of establishing legally-enforceable rules with Microsoft, covering the use of AI in schools.

Protections within the National AI Safety & Privacy Standard include the agreement not to use student or teacher data to train or improve AI models. Furthermore, any collected data will not be monetized, as educators attempt to find a middle ground that reduces screen time and balances technology use across all age groups.

This agreement follows various localized actions against AI use in schools, including a moratorium on AI use within the New York City school system, and a similar ban in the Los Angeles Unified School District.

Iron-clad privacy

“We have forged a hard-fought, iron-clad privacy agreement with real teeth that protects students and families, because no one else, including the federal government, has stepped up to do the real work,” said AFT President Randi Weingarten. “We can get angrier and angrier, or we can act decisively; anything less than legally enforceable provisions is simply a wish list.”

The concerns of the American Federation of Teachers (AFT) and the United Federation of Teachers (UFT) are multiple, and ultimately center on their teachers’ comfort, focus, and ability to conduct lessons without the specter of AI intrusions. So, issues about training data, monetization, screen time, and other AI-related challenges have been sewn up by the National AI Safety & Privacy Standard, which is legally enforceable.

“This agreement ensures three foundational goals: protect the privacy and data of children, give schools real control over AI tools in the classroom, and provide transparency and information to parents," Weingarten continued.

"And with New York City adopting our framework to ban screens and AI use in the younger years, this agreement adds to the momentum around online safety for kids and educators.”

Legally enforceable guardrails

The guardrails put in place by the agreement concern protecting privacy of educators, parents and students, enhancing safety of students and educators, and providing transparency and control about how AI is used in schools.

This also includes information about what data is collected and the protections that are in place. It places schools in control over the use of AI, and forces AI providers to communicate product changes and risks.

“This standard sets a high bar for child privacy and AI safety, and we’ll extend this agreement to every school district across the country,” said Microsoft Vice Chair and President Brad Smith.

Along with banning the use of student and teacher data for AI model training, the data will not be sold or repurposed, and again schools will control data use, along with how it is stored and deleted. The guardrails also ensure AI systems are protected by human oversight, and are designed to block harmful actions and protect sensitive information.

The AFT reports it is in talks with OpenAI and Anthropic to also adopt the National AI Safety & Privacy Standard.

At IFA 2026 I got to experience several pairs of smart specs including two from RayNeo: the impressive GT Max glasses which could be my new favorites, and the RayNeo iO.

I’ve tried specs like iO before. They’re smart like Meta Ray-Bans — meaning they can link up with apps and offer AI assistance on the go — but hide clear displays for HUD elements in their lenses, and — crucially, given the current backlash against so-called "pervert glasses" — lack a camera. And so far I’ve been unimpressed by the models like this that I've tested.

While not having a camera is a big privacy boost, it also strips out an incredibly useful tool that can be used for capturing first-person memories or giving the AI the context of what you can see. What’s more, the app used to control the specs and on-device controls can be pretty terrible. I haven't felt convinced that the features left behind once you remove the camera are really for me.

Out of the gate, the app and controls have been solved in the RayNeo iO. My experience with the app is that it’s incredibly smooth, with features loading up quickly.

Meanwhile moving between apps on the glasses is a breeze with the Digital Crown, a physical dial and button (yes, just like the Apple Watch and AirPods Max control of the same name) that's perfect for cycling between options. It's less temperamental than swiping on the frame, and more convenient than always needing your phone in your hand.

As for the features themselves, I’m still not convinced they’re for me, but they do have their uses. You can see notification popups, create memos for yourself, there’s real-time translation and captions, you can call on AI assistance, have easy access to information such as stock prices, or use them as a teleprompter.

The captions actually seem like an awesome accessibility tool for anyone hard of hearing, and RayNeo showed me how a deaf person can see the text of someone talking to them without needing to be able to see them and watch them sign — the glasses can effectively tell someone hard of hearing if someone behind them is trying to get their attention.

The standout tool of iO is Life Log, though. The glasses may lack a camera, but they can record every detail of your day through microphones, which pick up the sounds of your conversations, meetings, and musings to create detailed diary notes of what you’ve experienced.

This includes summaries of activities, suggestions for to-do list items you should remember to complete, and it can be used to inform third party AI services about your life if you choose to share it.

If you've been following the announcement of Apple's Audio Intelligence tools on Apple Watch, it's very much like that.

Features vs privacy

On this, RayNeo made clear to me in an interview that these specs are privacy-first. No recordings are ever stored, instead the AI transcribes and summarizes in real-time. Plus, the whole process is handled on device — across your glasses and phone — so nothing is shared with a server unless you choose to share it. Additionally, you won’t need to enable data sharing to access the tool.

Of course, there are still privacy considerations. A lack of cameras will cut out some of people’s smart glasses uneasiness, but audio recording can still be invasive (and illegal without consent in some parts of the world, such as two-party consent states in the US). The specs do let others know when you’re recording with a light, though it’s not the most obvious indicator.

But I like that you have good control over your data and how it’s shared with RayNeo and other services without locking yourself out of AI functions entirely.

My only other concern with the specs based on my quick demo of them is the Conversation Assistance feature, mostly because it feels a tad icky to me.

You can have the specs listen in on a conversation you're having — with demo reels showcasing it being used on a date or at a job interview — and the AI will display prompts for what you should say next in response to someone. I could perhaps understand this for a job interview if you’re super nervous, but I just kind of hate it overall.

Social interactions, be they dating or talking in a meeting, are a human aspect we shouldn’t be offloading. We should be engaging actively with each other, participating in life, not sitting back mindlessly and letting an AI respond for us.

Of course I’ll need to review RayNeo iO properly to judge it fully, but my initial takeaway is they’re the best cameraless smart specs I’ve seen so far. And tools like live captioning could be an amazing accessibility tool for folks who are hard of hearing.

I’m still not convinced this design is my cup of tea, but that’s perfectly okay. Speaking to smart glasses wearers at IFA, there’s a growing love for these cameraless designs as the Meta Ray-Ban backlash grows, and perhaps RayNeo’s design and stable software will prove to me why if I spend more time with them.

• Claude turned a famous mathematical proof into millions of checkable code lines
• Anthropic says Claude completed years of expected work in 11 days
• The massive proof contains 13 million lines of Lean code

Anthropic has used its Claude artificial intelligence system to produce a fully computer-checked version of a famous, centuries-old mathematical proof.

The proof addresses Fermat's Last Theorem, a hypothesis first proposed by the mathematician Pierre de Fermat back in the year 1637.

Mathematician Andrew Wiles produced the very first full mathematical proof of the theorem back in 1995, spanning 129 pages total in length.

A proof rebuilt for machines

Formalizing a proof simply means converting its mathematical reasoning into code that computers can check automatically without any human assistance.

Anthropic says it expected the entire task to take several years, based on how mathematicians first described the project.

Instead, the company says its internal research model finished the entire proof in only 11 days of continuous, largely unsupervised work.

The finished proof runs to 13 million lines of specialized code written in a programming language called Lean, used by mathematicians.

Along the way, Claude's agents reportedly proved roughly 30,300 separate theorems, ultimately using 29,500 of them in the final version.

Human input was reportedly limited to occasional high-level guidance, rather than any direct hands-on coding throughout the entire eleven-day process.

At 13 million lines, the resulting proof is over five times larger than Mathlib, the community's own main proof library.

"This extraordinary autoformalization achievement…proves Fermat's Last Theorem with no assumptions other than the axioms of mathematics," said Kevin Buzzard, a mathematician at Imperial College London.

“Along the way we see autoformalization of algebra, harmonic analysis, geometry and number theory, and we learn that AI autoformalization artefacts are now robust enough to be built upon; the proof is multi-layered.”

Anthropic attempted the formalization several times before succeeding, with those efforts contributing roughly 7% of the final proof’s non-boilerplate lines.

Not Anthropic's first math breakthrough

The formalization arrives just one month after Anthropic detailed a separate breakthrough involving the Riemann zeta function, a well-studied mathematical object.

That function sits at the very center of the Riemann hypothesis, considered one of mathematics' hardest unsolved problems worldwide.

Rival lab OpenAI is also pursuing similar work, using its newest Astra model to solve several classic Erdős problems.

That same OpenAI effort also reportedly narrowed several long-standing open questions within the field of theoretical computer science.

Anthropic says its breakthrough came only after giving Claude access to an open-source software tool named Prove2Me, built by outside collaborators.

The software helps AI agents pick the most useful next step during a long, multi-stage research workflow, while also cutting inference costs.

Anthropic has also expanded free access and research credits for mathematicians working specifically on formalization projects, alongside dedicated larger grants.

Despite the record pace set here, an eleven-day timeline still shows how labor-intensive full formalization remains, even with today's most capable systems.

• Watch Zverev vs Shelton free on 9Now (Australia)
• Unlock your stream with NordVPN (save up to 75%)
• Date and time: Sun Sep 13, 2pm ET / 7pm BST / 10am AEST (Sep 14)

Watch Alexander Zverev vs Ben Shelton live streams in a fascinating US Open 2026 Men's Final from Arthur Ashe Stadium at Flushing Meadows, New York. The German top seed has made his third grand slam final in a row, while the left hander from Atlanta is trying to become the first male US grand slam champion since Andy Roddick, here in 2003.

After dropping the first two sets in a barrage of brutal serving from his opponent, Shelton has won the third set 7-5. He trails Zverev by 6-3, 7-6, 5-7 as we go into the fourth set. Zverev looked tight at the end of the third set and has looked not to lose instead of attacking on the front foot. Shelton, by contrast, has served better and gone big from the baseline to build his confidence. The crowd are in this now and is a wobbling Zverev able to get the thoughts of Dominic Thiem six years ago out of his head? This could be anyone's now.

Zverev has grown into the tournament after going five sets in his opening two rounds – and spending over five hours on court in finally beating Lorenzo Sonego in his first match – and the top seed's 107-minute destruction of Karen Khachanov in the semi-finals was his fourth successive straight-sets victory in New York. Since breaking his major duck at the French Open earlier this year, the 29-year-old has been in imperious form with a serve that has gone from significant weakness to serious strength. Six years after reaching his first slam final, can he avenge that devastating loss to Thiem in which he led by two sets and served for the match?

Shelton has already established one piece of history, after becoming the first black American man to reach a US Open final since Ashe in 1972. Now he wants another – to become the first US man to win a major in 23 years. The 23-year-old came from a set down to beat compatriot Frances Tiafoe in the semi-finals with a serving display of brutal efficiency – 75% first serves in, 20 aces and countless cheap points won – and to beat an opponent he never has before, Shelton must replicate those staggering numbers. He's already knocked out defending champion Carlos Alcaraz this tournament doing exactly that.

Here's how to watch Zverev vs Shelton in the 2026 US Open semi-finals from anywhere in the world.

Can you watch Zverev vs Shelton for free?

Yes. Free US Open 2026 coverage, including Zverev vs Shelton, is being provided by Channel 9 and 9Gem in Australia, with live streaming available via 9Now.

Traveling abroad right now? You can use a VPN to watch Zverev vs Shelton for free as if you were right at home.

Use a VPN to watch Zverev vs Shelton live streams

A VPN is handy piece of software that can make your device appear as if it's back in your home country, so you can unlock your usual service. The best VPN right now? We recommend NordVPN – it does everything and comes with up to 75% off.

How to watch Zverev vs Shelton live streams in the US

Zverev vs Shelton coverage is being provided by ABC and ESPN Deportes in the US.

If you have a TV antenna or cable then you'll be able to watch ABC via that method without extra cost. But what if you've cut the cord?

ESPN Select is the dedicated streaming service that carries ESPN and prices start from $13.99 for the entry-level plan. ESPN Unlimited, which is live streaming every match on every court, costs $31.99/month or $319.99/year, but you can bundle it with Disney Plus and Hulu for $35.99/month.

ESPN Unlimited includes streaming access to ESPN, ESPN2, ESPNU, ESPNEWS, ESPN Deportes, SECN, ACCN, ESPN+, ESPN on ABC, SECN+, ACCNX, NFL Network and CW Sports.

Alternatively, ESPN and ABC are available through Sling – you can access the channel on the slimmed-down Essentials plan for $19.99/month. Or if you want a host of extra channels – including TNT, CNN and AMC in addition to ESPN – the Sling Orange plan at $45.99/month is what you want.

ESPN and ABC are also carried by Fubo, YouTube TV and Hulu + Live TV, each of which offers a free trial to new users.

Outside of the US? Use a VPN while you're traveling away from home to unlock your stream.

How to watch Zverev vs Shelton live streams in the UK

In the UK, live Zverev vs Shelton coverage is being provided by Sky Sports.

Sky Sports packages start at £35/month with £22/month pricing available for those who are already customers. You can also tune in via a NOW Sports membership that carries the Sky Sports channels.

If you're out of the UK but still want to tune in, explore the VPN route set out above, which will help you access your accounts from anywhere.

How to watch Zverev vs Shelton live streams in Australia

Zverev vs Shelton is free-to-air on Channel 9 and 9Gem in Australia, with live streaming available via the 9Now platform.

Stan Sport, meanwhile, will have exhaustive coverage of every match. Stan Sport costs AU$20/month on top of a Stan subscription, which itself starts at AU$9.99/month. It's also home to the Premier League, Champions League, Rugby's Greatest Rivalry, Super Rugby, the Six Nations and the Nations Championship.

Not in Australia right now? You can simply use a VPN like NordVPN to watch all the action as if you were back home.

How to watch Zverev vs Shelton live streams in New Zealand

Zverev vs Shelton is exclusively available on ESPN via Disney+ in New Zealand, with all matches live across the entire two-week period of the Slam.

Disney+ prices start at NZ$9.99 a month for the ad-supported tier in New Zealand, with free access to ESPN, which means you also get extensive basketball coverage (NBA, WNBA, NBL and more) as well as UFC Fight Night, Stanely Cup playoffs, and a heck of a lot more.

Not in New Zealand right now? You can simply use a VPN like NordVPN to watch all the action as if you were back home.

How to watch Zverev vs Shelton live streams in Canada

In Canada, Zverev vs Shelton is exclusive to TSN.

If you don't have cable, the TSN Plus streaming service costs CA$29.99/month or CA$249.99/year.

If you're out of Canada but still want to tune in, explore the VPN route set out above, which will help you access your accounts from anywhere.

We test and review VPN services in the context of legal recreational uses. For example:1. Accessing a service from another country (subject to the terms and conditions of that service).2. Protecting your online security and strengthening your online privacy when abroad.We do not support or condone the illegal or malicious use of VPN services. Consuming pirated content that is paid-for is neither endorsed nor approved by Future Publishing.

Just as soon as we've got our boy back on the TV, he's set to disappear again. Reacher season 4 officially ends this week, but not without an epic showdown as a fitting farewell.

Some fans argue that the political espionage plot at the center of new episodes has come too far away from what a Jack Reacher story should be... and to an extent, I agree. We've definitely all forgotten that Reacher is intended to be light-hearted and silly, but a fresh take on a worn format is often a necessity.

Now we're at the end, when does Reacher season 4 episode 8 arrive on Prime Video?

What time can I watch Reacher season 4 episode 8 on Prime Video?

We're expecting episode 8 of Reacher season 4 to be available on September 16 at 12am PT / 3am ET in the US.

Internationally, look out for the below times:

• US – 12am PT / 3am ET
• Canada – 12am PT / 3am ET
• UK – 8am BST
• India – 12:30pm IST
• Singapore – 3pm SGT
• Australia – 5pm AEST
• New Zealand – 7pm NZST

When do new episodes of Reacher season 4 come out?

They won't. As I mentioned up top, Reacher season 4 draws to a close with this week's chapter, so there won't be any more entries to enjoy once this season's finale has premiered on September 16.

However, the Reacher season 4 finale coincides with the day that all episodes of its first spin-off Neagley lands on Prime Video. So, if you want to continue immersing yourself in this world, you'll be able to do so.

• Both the iPhone 18 Pro and Pro Max get battery bumps
• The increases are now confirmed in EU labeling
• In the US, the iPhone 18 Pro Max sticks with a Qualcomm modem

Apple doesn't always publish full spec sheets for its devices, but if you were wondering just what the battery capacity of the iPhone 18 Pro and Pro Max were, we now know — thanks to European Union regulations demanding that these details are made public.

Energy labels applied to the phones in the EU (via Android Headlines) show that the iPhone 18 Pro is fitted with a 4,056 mAh battery, while the iPhone 18 Pro Max has a 5,391 mAh battery inside. Those represent 1.45% and 11.8% increases, respectively, on the models from the previous year.

Apple describes the battery bump on the iPhone 18 Pro Max as "the largest increase in battery life ever" on an iPhone, and says you can get 45 hours of video playback between recharges. For the iPhone 18 Pro, it's 36 hours.

One small wrinkle: these capacities refer to iPhones sold in the EU, so these models will be ones with a physical SIM card slot. The handsets sold in the US and some other regions are eSIM-only, so may have even better batteries inside.

Something different

There's more specs digging over at MacRumors, where a tip shared on social media suggests that every iPhone 18 Pro and Pro Max handset has Apple's own C2 modem inside — except for the iPhone 18 Pro Max models sold in the US, which uses a Qualcomm chip.

The difference has apparently been revealed in code for iOS 27 — though for the moment it's not confirmed (pre-orders aren't shipping until Friday, September 18) — and it's most likely the Qualcomm Snapdragon X85 modem.

There shouldn't be any noticeable variation in performance between the Apple C2 and Qualcomm Snapdragon X85, though Apple says the C2 offers faster speeds than its previous C1X. It's not clear why the Qualcomm chip is hanging around in this one model, but it may be to do with contractual obligations.

We'll be bringing you our thoughts on these phones in the coming days and weeks, including the foldable Apple iPhone Duo. And if you're wondering where the standard iPhone 18 is, it's expected in March with the iPhone 18e and iPhone Air 2.

• Man Utd vs Man City: Sunday, September 13 at 4.30pm BST / 11.30am ET / 1.30am AEST (September 14)
• Streams: Sky Sports (UK), Peacock (US)
• Outside the UK? Use NordVPN to unblock your streaming account

Watch Man Utd vs Man City in the Premier League 2026/27 season, as these two fierce adversaries renew hostilities at Old Trafford, where the visitors will be looking to make it four top-flight wins from four to start the campaign.

Enzo Maresca oversaw a 3-0 defeat by Arsenal in the Community Shield, but he boasts a 100 percent record in the Premier League as City boss. A 1-0 victory over Coventry last weekend out was closer than many imagined, but it ensured they remain level on points with Arsenal at the summit of the standings early on. A 2-0 midweek victory over Porto took Erling Haaland to five goals in all competitions.

Man Utd have had a more mixed start. A 2-0 loss to Hull was followed by a 5-2 triumph over Ipswich, before a last-gasp equalizer from Everton left Man Utd rueing their profligacy. The Red Devils have been inconsistent up to now as Michael Carrick attempts to get them closer to the title, but a 4-0 gubbing of Sabah in the Champions League should help morale. Benjamin Sesko has two in his past two games.

Man City will see this as an opportunity to remind their adversaries that they remain Manchester's pre-eminent side even after Pep Guardiola's exit. But Man Utd tended to raise their game for this fixture, most notably in a 2-0 victory in Carrick's first game in charge in January.

Here's how to watch Man Utd vs Man City in the 2026/27 Premier League from anywhere in the world.

Man Utd vs Man City team news

Man Utd XI: Lammens; Dalot, Maguire, Martinez, Dorgu; Tielemans, Mainoo; Mbeumo, Fernandes, Rashford; Cunha.

Man City XI: Donnarumma; Nunes, Dias, Guehi, Gvardiol; Anderson, Fernandez; Foden, Cherki, Semenyo; Haaland.

How to watch Man Utd vs Man City from anywhere

A VPN is handy piece of software that can make your device appear as if it's back in your home country, so you can unlock your usual service. The best VPN right now? We recommend NordVPN – it does everything and you can get our exclusive deal too...

How to watch Man Utd vs Man City live streams in the USA

Soccer fans in the US can watch Man Utd vs Man City in the Premier League on Peacock.

Peacock Premium plans start from $12.99 a month, or $129.99 a year.

Outside of the US? Use a VPN whilst you're traveling away from home to unlock your stream. We like NordVPN best.

How to watch a Man Utd vs Man City live stream in the UK

Sky Sports is broadcasting Man Utd vs Man City in the UK on its Premier League and Main Event channels.

Packages cost £35 per month for new customers, or £22 per month if you already have a Sky TV contract. The Sky Go app lets you watch on your laptop, mobile or tablet.

Fans can also tune in using a NOW Sports Membership, with 24-hour passes costing £14.99.

Not in the UK for Man Utd vs Man City? Use NordVPN to access your usual Premier League streams.

How to watch Man Utd vs Man City live streams in Canada

Fubo is the Premier League's broadcast partner in Canada and it has live coverage of Man Utd vs Man City.

Prices start from CA$31.49 per month for Fubo's Sports Monthly package. However, the best option right now is the Quarterly Package which costs just CA$27.99 CA$14.99 for each of the first three months. You can also watch Fubo's content if you have a select DAZN subscription in Canada.

Not in Canada? Use NordVPN if you're away from home, in order to tap into your preferred coverage from anywhere.

How to watch Man Utd vs Man City live streams in Australia

A Man Utd vs Man City live stream is available in Australia on Stan Sport.

You can pick up a Stan Sport subscription for as little as AU$29.99 per month for the basic plan containing adds. That price consists of the AU$9.99 Stan subscription fee plus the Sport add-on for AU$20.

Outside of Oz? You can use a VPN to access your usual Premier League coverage. We recommend NordVPN.

We test and review VPN services in the context of legal recreational uses. For example:1. Accessing a service from another country (subject to the terms and conditions of that service).2. Protecting your online security and strengthening your online privacy when abroad.We do not support or condone the illegal or malicious use of VPN services. Consuming pirated content that is paid-for is neither endorsed nor approved by Future Publishing.

• Premiere: Sunday, September 13 at 9pm BST
• FREE stream: ITVX (UK)
• Use NordVPN to watch ITVX from abroad

BBUK is back for a bumper season as hosts AJ Odudu and Will Best welcome a new batch of housemates vying for the £100,000 prize. Set to run for eight weeks, with 47 episodes spread over the 54 days, Big Brother season 23 will be the longest of the ITV era, with the highest episode count for a season since 2017.

The extended run no doubt means plenty more tasks, more evictions and plenty of twists to look forward to as the housemate class of 2026 look to replicate 60-year-old music composer Richard Storry’s 2025 victory, which made him the oldest winner of the UK civilian series in history. As for who those housemates are, we’ll have to wait and see, with producers playing their cards close to their chest until launch night. However, for those looking to get a sneak peek at the new set, content creators The Bov Boys will be giving it a 24-hour test run, live streamed on Twitch, ITV Reality and YouTube.

There’s also a brand new companion show for 2026 following the cancellation of Big Brother: Late & Live, with GK Barry and former housemate Kate Lawler hosting visual podcast Big Brother: The Group Chat. More on that below.

With more episodes than ever in BB’s new era, it’s sure to be an epic season, so read on and find out how to watch Big Brother UK 2026 online and from anywhere.

How to watch Big Brother UK S23 from abroad

For those away from home looking to watch Big Brother UK, you’ll be unable to watch the show as normal if you're outside the UK due to regional restrictions. Luckily, there’s an easy solution.

Downloading a VPN allows you to stream geo-blocked services online, no matter where you are. It's a simple bit of software that changes your IP address, meaning that you can access on-demand content or live TV just as if you were at home. Our favorite is NordVPN – it's one of the best VPNs on the market.

Use a VPN to watch Big Brother UK from anywhere:

How to watch Big Brother UK S23 online in the UK

Can I watch Big Brother UK S23 online in the US, Canada or Australia?

While a whole host of countries have their own home grown version of Big Brother, the UK iteration doesn't air outside of Blighty, meaning those looking to stream in the US, Canada, Australia and beyond are out of luck.

If you're a Brit abroad you can use a VPN to stream for free on ITVX just as you would back home.

What you need to know about Big Brother UK 2026

We test and review VPN services in the context of legal recreational uses. For example:1. Accessing a service from another country (subject to the terms and conditions of that service).2. Protecting your online security and strengthening your online privacy when abroad.We do not support or condone the illegal or malicious use of VPN services. Consuming pirated content that is paid-for is neither endorsed nor approved by Future Publishing.

• A £3 Wi-Fi extender carried hidden administrator access beyond ordinary user controls
• Every device running the firmware shared the same concealed administrator password
• Changing the visible administrator password could not disable the secret account

A £3 Wi-Fi extender bought through Temu has exposed security problems that challenge the idea of cheap connected devices being simple bargains.

Security researcher Keiran Smith examined the device and discovered hidden access features that ordinary users would never see during normal operation.

Smith, who holds a penetration-testing certification, picked up the six-antenna extender after seeing it promoted through a targeted ad on the shopping app.

The cheap extender contained access users could not control

The examination began with the hardware, where he identified a MediaTek MT7620 processor commonly used in low-cost networking products.

After extracting the firmware stored inside the device, Smith found a hidden administrator account with complete control over its functions.

The account used a fixed password embedded inside the software, meaning every unit using that firmware carried the same credentials.

Changing the normal administrator password through the device settings would not remove this separate hidden access.

Smith also found a remote login service that accepted the concealed credentials without requiring physical access to the extender itself.

“It’s worth being precise about what makes this as bad as it is, because ‘hardcoded password’ covers a wide range of sins,” Smith said

The researcher said this case was more serious because the password remained identical across devices rather than being generated individually.

“A default credential is something the owner can see, is told about and can change,” he said. “What we have here is the opposite on every count.”

“This one is a compile-time constant rather than something derived from the MAC address or serial number, so it is identical on every unit ever sold.”

The combination of hidden access, unchanged credentials, and remote availability creates a security concern for ordinary owners.

Even if technically skilled users discovered the account, Smith found that changes could disappear after restarting the extender.

Additional flaws raise questions about cheap connected hardware

The investigation also uncovered a command injection weakness that could allow attackers to execute unauthorized instructions through the device.

Smith found that the extender lacked strong protection around software updates, creating possible opportunities for tampered firmware installation.

He admitted that these issues did not prove that manufacturers intentionally created unsafe features for malicious purposes.

They could have originated from factory testing processes and remained active accidentally before consumer sales.

This Temu extender shows how extremely cheap smart devices can create security challenges beyond their purchase price.

Consumers may focus on immediate savings while having little visibility into the software decisions built inside connected equipment.

The findings do not mean every inexpensive networking device contains similar weaknesses, though they show why basic security checks matter.

As more homes add connected products, hidden software features could become a larger concern for users and manufacturers.

Via CyberNews

• ArenaNet studio head Colin Johanson has offered a deep dive into Guild Wars 3's momentum system
• The feature offers "fluid" transitions between movement and combat, which Johanson says is a "first" for the MMORPG genre
• Players can build momentum as they move around the world and in combat to deliver powerful attacks

ArenaNet has revealed more about Guild Wars 3's newly introduced momentum system, a mechanic that blends all actions and combat abilities to ensure constant "fluid" movement.

During an exclusive presentation TechRadar Gaming attended at Gamescom, ArenaNet studio head Colin Johanson showcased the studio's action-adventure MMORPG and offered a deep dive into the game's story and open world, the combat system, multiplayer design, and more.

Johanson also touched on the game's new momentum system, saying Guild Wars 3 will be the first MMORPG to offer smooth transitions between movement and combat.

"If you can picture with me for a second, movement in an MMO, you're riding through the world on your mount, and you get off your mount, and your character immediately stops moving," he explained. "And then you start moving again, and you go to glide, and you stop gliding, and your character immediately stops moving. Then you go into combat, and you use an ability, and then you move, and you'll stop moving again."

He continued: "There are no fluid transitions between movement and combat in MMOs, and Guild Wars 3's momentum system introduces the first MMO that has truly fluid movement across the entire world."

With this system, Guild Wars 3 players will be able to jump off a mountain, deploy a glider, dive, fly, or wall-run, all while carrying momentum, "creating a seamless sense of movement you'd expect to see in an action game, rather than an MMORPG."

Momentum is also part of combat; we previewed a warrior class that has a unique ability that uses their momentum and modifies their attacks when they come down from the sky. This increases damage output and the radius of attacks "based on his velocity or how much downward momentum" the warrior carries. The class can also run up walls and slam back down on enemies, delivering huge amounts of damage.

"Momentum also impacts battlefield control," Johanson said, adding that the faster the player is moving, the more momentum they have.

The environment, Johanson said, is also "part of the experience," and enemies can be knocked into walls to be stunned and damaged by building up momentum in combat, explaining that players will need to use their environment, whether it's the battlefield or terrain, to their advantage.

"This is another first for the MMORPG genre," he said.

Guild Wars 3 doesn't have a release date, but a beta is scheduled for Fall 2027. The game is coming to PC and PlayStation 5.

• Rangers vs Celtic: Sun, September 13, 7am ET / 12pm BST / 9pm AEST
• Stream on Premier Sports (UK) / Paramount+ & CBS Sports Golazo (US)
• Unlock your stream with NordVPN (save up to 75%)

Watch Ranges vs Celtic live streams in the Scottish League Cup 2026/27 quarter-finals as Derek McInnes takes charge of an Old Firm game for the first time and attempts to earn early bragging rights at Ibrox.

There will be a different feel to this season’s meetings between Scotland’s most successful clubs as no away fans are permitted in Old Firm fixtures this term. In recent weeks there has been a feeling of anxiousness amongst Gers supporters, though. There were plenty of moans and groans in midweek before Bojan Miovski scored a crucial stoppage-time winner against St Mirren that made it four league wins in a row for Rangers – exiting in Europa League and Conference League qualifying is a stain that won't go away quickly. A fast start would ensure the crowd are on side from the get-go as the club look to avenge their two cup defeats by Celtic last season.

They may be traveling to enemy territory without their fans, but Celtic will make the short journey across Glasgow full of confidence. The Bhoys have won all six league games this season, scoring 14 times and conceding just three. Even with new signing Kasper Hogh only scoring in one of those six games (albeit a hat-trick), Martin O’Neill’s side have looked sharp at the top end of the pitch. Fellow new arrival Camilo Duran has been particularly impressive and will be confident of adding to his tally of six goals in nine appearances. The Hoops' 5-1 loss at LASK to exit the Champions League still ligers, though.

Read on as we explain how to watch Rangers vs Celtic live streams from anywhere.

Can you watch Rangers vs Celtic for FREE?

Yes. Old Firm fans in the US can watch Rangers vs Celtic on CBS Sports Golazo Network, which is completely free.

Fans Down Under can also watch with a 7-day free trial for new customers to beIN Sports, which is showing the game..

Not in the US or Australia right now? You need a VPN to unlock your FREE stream — find out more below.

Use a VPN to watch Rangers vs Celtic live streams

A VPN is handy piece of software that can make your device appear as if it's back in your home country, so you can unlock your usual service. The best VPN right now? We recommend NordVPN – it does everything and comes with up to 75% off.

How to watch Rangers vs Celtic live streams in the USA

In the US, Rangers vs Celtic is being shown on Paramount+.

The streaming service, which is also showing all 189 games of the 2026/27 Champions League campaign, starts at $8.99/month or $89.99/year.

However, you can access Paramount+ virtually for free with a Walmart Plus 30-day trial for $1.

As we've mentioned above, the match is also listed as being available on CBS Sports Golazo Network.

US residents visiting another country can access the game using NordVPN.

How to watch Rangers vs Celtic live streams in the UK

Premier Sports is broadcasting Rangers vs Celtic in the UK on the Premier Sports 1 channel.

Premier Sports costs £12.99 per month with a 12-month commitment, or £17.99 per month on a rolling 30-day basis when signed up directly.

If you're already a Sky customer or subscribe to Amazon Video, you can add Premier Sports for £17.99 per month.

Not in the UK this Sunday? Use NordVPN to access your usual Scottish League Cup coverage.

How to watch Rangers vs Celtic live streams in Australia

Rangers vs Celtic is available to watch on beIN Sports in Australia.

beIN SPORTS CONNECT costs AU$16.99/month or $169.99/year after a 7-day FREE trial. The service includes live and on-demand streaming of European football and tennis.

Missing the game because you're outside Australia? Use a VPN to log in to your beIN Sports account.

We test and review VPN services in the context of legal recreational uses. For example:1. Accessing a service from another country (subject to the terms and conditions of that service).2. Protecting your online security and strengthening your online privacy when abroad.We do not support or condone the illegal or malicious use of VPN services. Consuming pirated content that is paid-for is neither endorsed nor approved by Future Publishing.

Any users of the best Fitbit devices that were watching Apple’s iPhone Duo event yesterday might have felt a chill of all-too-familiar dread during the company’s glitzy keynote presentation.

That’s not because Apple announced some sort of killer smartwatch feature that will bury the Fitbit range alive. No, it’s because viewers got to see Apple’s new plans for its Health app, which now comes with an AI-powered makeover that will likely stir up some painful emotions for many Fitbit fans.

Of course, adding AI into an app doesn’t automatically doom it. But the parallels between Apple Health and the Google Health app that Fitbit users have been lumped with — with all its wild hallucinations and crazy wellbeing inaccuracies — is undeniable.

Indeed, if the Fitbit experience is anything to go by, there’s a risk that Apple customers could be in for a bumpy ride.

Coming to your iPhone

Why might Fitbit fans be ringing a Google Health-shaped alarm bell? Well, Google-owned Fitbit announced a total overhaul of the companion app for its wearables back in May 2026 — and it was almost immediately unpopular.

The rebranded app was accused of being “unbelievably bad” amid missing features and confusing user interface decisions that made the app much less enjoyable to use than the one it replaced.

But that was just the tip of the iceberg. As the weeks and months rolled on, social media started overflowing with user complaints directly squarely at one target: Google Health’s AI assistant.

This tool is meant to analyze your wellbeing data and make suggestions on ways you can improve your fitness and live a healthier life. The problem, though, is that the AI seemed to frequently go off the rails.

I’ve written about how Google’s AI has been giving people “unhinged” fitness advice, feeding them “nonstop lies” and spinning up completely fictional food-tracking metrics. It’s a real concern and has proven to be a tremendous source of frustration to anyone unfortunate enough to have to use it.

And now with the Apple Health revamp, a similar feature might be coming to your iPhone.

An anxious wait

In any other app, patchy performance like this might just be a slight irritation. But in an app like Google Health, where the AI is making suggestions that could directly affect your physical wellbeing, the effect is much more serious.

Can Apple escape such problems with its own AI revamp? I’m not sure, despite Apple's excellent track record and cautious implementation of LLMs. After all, Apple tried to go its own way with Apple Intelligence but fell so far behind the competition that it had to enlist the help of — you guessed it – Google.

That means that the same AI models that have messed up Google Health so badly are now propping up Apple Intelligence, and with it, Apple Health.

That said, there’s no guarantee that Apple Health will go the same way as Google Health. While Google’s AI is powering many of the features inside the likes of Apple Intelligence and Siri AI, we don’t know exactly how that’s going to work in Apple’s Health app do-over. Has Apple tweaked Google’s models to its own tastes or imported them wholesale?

I’d hope that Apple has made some modifications of its own. Apple's website states that 'all of our health innovations are subject to rigorous scientific validation and are developed with clinical experts from start to finish'.

Not only that, but the company is extremely hot on user privacy and security, and what could be more vulnerable than your health data? With almost every health-related feature that it has introduced in recent years, Apple has emphasized the steps it has taken to encrypt, protect and otherwise lock down your sensitive wellbeing information.

Considering that Google could be perceived to be, relatively speaking, far more lax about user privacy, I’d be shocked if Apple hadn’t taken a peek under the hood to ensure that its Health AI is as secure and reliable as possible. It would have been hard for Apple not to notice the furor that Google Health has caused Fitbit fans.

There’s no way Apple would want that kind of feedback from its own users, so I’m hoping it’s made some adjustments to ensure that its own AI is a little less eccentric, shall we say, than Google’s.

We’re not going to know for sure until the new Health app becomes available, which won’t be until 'later this year,' Apple says. Until that happens, it’s going to be an anxious wait for any Apple user aware of the carnage suffered by their Fitbit-wielding colleagues.