SoylentNews

JoeMerchant writes:

https://www.bleepingcomputer.com/news/microsoft/microsoft-august-security-updates-break-windows-recovery-reset/

Microsoft has confirmed that the August 2025 Windows security updates are breaking reset and recovery operations on systems running Windows 10 and older versions of Windows 11.

"After installing the August 2025 Windows security update [..] on any of the client versions mentioned below in the 'Affected platforms' section, attempts to reset or recover the device might fail," the company said in a new Windows release health update.

Installing this month's security updates will cause issues for users who want to reinstall their system while keeping their files using the Reset my PC feature, or reinstall it and keep their files, apps, and settings using the Fix problems using Windows Update tool.

The known issue may also impact users who want to remotely reset devices using the RemoteWipe configuration service provider (RemoteWipe CSP).

According to Redmond, the bug only impacts client platforms after installing the following updates, including:

• Windows 11 23H2 and Windows 11 22H2 (KB5063875),
• Windows 10 22H2, Windows 10 Enterprise LTSC 2021, Windows 10 IoT Enterprise LTSC 2021 (KB5063709),
• Windows 10 Enterprise LTSC 2019, Windows 10 IoT Enterprise LTSC 2019 (KB5063877).

The company is currently working on a fix for this known issue, which will be delivered via out-of-band updates for all impacted platforms over the coming days.

Original Submission

looorg writes:

SlowTV. When they need to move a 600-tonne church to the other side of town.

As interesting as it was I still prefer 'The Great Moose Migration' (x) they have on TV instead. It's more relaxing.

This is just unnerving.

https://www.bbc.com/news/articles/cde3xp4xlw9o
https://www.forbes.com/sites/davidnikel/2025/08/19/swedens-kiruna-church-is-on-the-move-after-more-than-100-years/
https://www.nytimes.com/2025/08/19/world/europe/sweden-church-kiruna-move.html
https://www.latimes.com/world-nation/story/2025-08-19/this-swedish-church-is-moving-3-miles-down-the-road-its-relocating-before-a-mine-swallows-the-town

https://lkab.com/en/events/the-moving-of-kiruna-church/
https://www.svtplay.se/video/KnDgQ9G/den-stora-kyrkflytten/forsta-dagen?position=4104 [in Swedish]
https://player.vimeo.com/video/1108407438?h=11c3a1f23e&dnt=1&app_id=122963
https://www.byggnadsarbetaren.se/600-ton-och-k-markt-sa-ska-kiruna-kyrka-flyttas/ [in Swedish]

https://en.wikipedia.org/wiki/Kiruna_Church

(x) https://apnews.com/article/sweden-great-moose-migration-slow-tv-7c13e3b13a6b5bec01dfc3bb7465506d

Original Submission

upstart writes:

How Chefs and Scientists Are Using Kombucha and Kimchi to Study Microbiology:

Scientists and chefs have collaborated on a new study that demonstrates how fermented foods can be used to drive participatory science projects that both engage the public and advance our understanding of microbial ecology. The study focused on working with food experts and the public to examine the microbial communities associated with kombucha, kimchi and chow chow.

"One of the things we demonstrated here is that this approach works, it's relatively inexpensive, and it is easy to scale," says Erin McKenney, co-lead author of a paper on the work and an assistant professor of applied ecology at North Carolina State University.

"This proof-of-concept study focused on questions that have been answered using conventional approaches, allowing us to determine that the findings from our approach are consistent with established findings," McKenney says. "But now that we have that proof of concept, we can begin using this technique to address additional questions."

For the study, the researchers hosted three participatory science workshops at the North Carolina Museum of Natural Sciences in which scientists and chefs instructed K-12 teachers and members of the public on how to make fermented foods. Each workshop focused on a specific fermented food: kimchi, chow chow and kombucha.

While workshop participation varied, the researchers ended up with 18-23 samples of each fermented product.

Liquid samples were taken from each of the fermented foods at different points, to see how the microbial communities in each sample changed as the fermentation progressed. Samples were taken from chow chow and kimchi on days 3 and 10; kombucha samples were taken on days 4 and 8.

The researchers conducted DNA sequencing of each sample to get a snapshot of both the diversity and overall abundance of microbes in the sample.

An Anonymous Coward writes:

Sni5Gect research crew targets sweet spot during device / network handshake pause

https://www.theregister.com/2025/08/18/sni5gect/
https://archive.ph/buKXp

Security [scientists] have released an open source tool for poking holes in 5G mobile networks, claiming it can do up- and downlink sniffing and a novel connection downgrade attack - plus "other serious exploits" they're keeping under wraps, for now.

"Sni5Gect [is] a framework that sniffs messages from pre-authentication 5G communication in real-time," the researchers from the Singapore University of Technology and Design explained of their work, presented this week at the 34th USENIX security bash, "and injects targeted attack payload in downlink communication towards the UE [User Equipment, i.e. a phone]."

Designed to take advantage of the period just after a device connects to a 5G network and is still in the process of handshaking and authentication - which, the team points out, can occur when entering or leaving a lift, disembarking a plane and turning aeroplane mode off, or even passing through a tunnel or parking garage - Sni5Gect takes advantage of unencrypted messaging between the base station and a target handset.

"Since messages exchanged between the gNB [Next-Generation Node B, the base station] and the UE are not encrypted before the security context is established (pre-authentication state)," the researchers wrote, "an attacker does not require knowledge of the UE's credentials to sniff uplink/downlink [traffic] nor to inject messages without integrity protection throughout the UE connection procedure."

That's a flaw, and one the framework is designed to exploit. The team's testing showed it capable of sniffing both uplink and downlink traffic with more than 80 percent accuracy, at ranges of up to 20 meters between an off-the-shelf software-defined radio and the target mobile. For packet injection, the success rate varied between 70-90 percent - and delivered, among other things, proof of a novel downgrade attack by which a ne'er-do-well equipped with Sni5Gect could downgrade a connection from 5G to 4G to reduce its security and carry out further surveillance and attacks.

As Sni5Gect works in real-time, its creators have claimed, and can inject attack payloads, including multi-stage attacks, based on protocol state, it's suited to fingerprinting, denial-of-service attacks, and downgrading.

jelizondo writes:

The Guardian recently published a couple of articles about digital streaming services and piracy. The first one reads:

With a trip to Florence booked, all I want is to rewatch Medici. The 2016 historical drama series tells of the rise of the powerful Florentine banking dynasty, and with it, the story of the Renaissance. Until recently, I could simply have gone to Netflix and found it there, alongside a wide array of award-winning and obscure titles. But when I Google the show in 2025, the Netflix link only takes me to a blank page. I don't see it on HBO Max, Disney+, Apple TV+, or any of the smaller streaming platforms. On Amazon Prime I am required to buy each of the three seasons or 24 episodes separately, whereupon they would be stored in a library subject to overnight deletion. Raised in the land of The Pirate Bay, the Swedish torrent index, I feel, for the first time in a decade, a nostalgia for the high seas of digital piracy. And I am not alone.

For my teenage self in the 00s, torrenting was the norm. Need the new Coldplay album on your iPod? The Pirate Bay. The 1968 adaptation of Romeo and Juliet? The Pirate Bay. Whatever you needed was accessible with just a couple of clicks. But as smartphones proliferated, so did Spotify, the music streaming platform that is also headquartered in Sweden. The same Scandinavian country had become a hub of illegal torrenting and simultaneously conjured forth its solution.

"Spotify would never have seen the light of day without The Pirate Bay," Per Sundin, the then managing director of Universal Music Sweden, reflected in 2011 [in Swedish]. But music torrenting died out as we all either listened with ads or paid for the subscription. And when Netflix launched in Sweden in late 2012, open talk of torrenting moving images also stopped. Most of the big shows and a great collection of award-winning films could all be found for just 79 SEK (£6) a month. Meanwhile, the three founders of The Pirate Bay were arrested and eventually jailed. Pirating faded into the history books as far as I was concerned.

A decade and a half on from the Pirate Bay trial, the winds have begun to shift. On an unusually warm summer's day, I sit with fellow film critics by the old city harbour, once a haven for merchants and, rumour has it, smugglers. Cold bigstrongs in hand (that's what they call pints up here), they start venting about the "enshittification" of streaming – enshittification being the process by which platforms degrade their services and ultimately die in the pursuit of profit. Netflix now costs upwards of 199 SEK (£15), and you need more and more subscriptions to watch the same shows you used to find in one place.

A fellow film critic confides anonymously: "I never stopped pirating, and my partner also does it if he doesn't find the precise edition he is looking for on DVD." While some people never abandoned piracy, others admit they have recently returned – this time turning to unofficial streaming platforms. One commonly used app is legal but can, through community add-ons, channel illicit streams. "Downloading is too difficult. I don't know where to start," says one film viewer. "The shady streams might bombard me with ads, but at least I don't have to worry about getting hacked or caught."

According to London‑based piracy monitoring and content‑protection firm MUSO, unlicensed streaming is the predominant source of TV and film piracy, accounting for 96% [PDF] in 2023. Piracy reached a low in 2020, with 130bn website visits. But by 2024 that number had risen to 216bn. In Sweden, 25% of people [PDF] surveyed reported pirating in 2024, a trend mostly driven by those aged 15 to 24. Piracy is back, just sailing under a different flag.

"Piracy is not a pricing issue," Gabe Newell, the co-founder of Valve, the company behind the world's largest PC gaming platform, Steam, observed in 2011. "It's a service issue." Today, the crisis in streaming makes this clearer than ever. With titles scattered, prices on the rise, and bitrates throttled depending on your browser, it is little wonder some viewers are raising the jolly roger again. Studios carve out fiefdoms, build walls and levy tolls for those who wish to visit. The result is artificial scarcity in a digital world that promised abundance.

Whether piracy today is rebellion or resignation is almost irrelevant; the sails are hoisted either way.

upstart writes:

Attorneys General, HCA Settle Over Nurse Training Repayment Provisions:

California Attorney General Rob Bonta has announced a settlement with HCA Healthcare Inc. and Health Trust Workforce Solutions LLC (together, HCA), resolving allegations that HCA unlawfully required entry-level nurse employees to repay the cost of a mandatory training program if they did not remain employed with the company for two years.

One of the nation's largest hospital systems, for-profit HCA has several hospitals in California.

Today's settlement is the result of a years-long investigation by attorneys general in California, Colorado and Nevada, working in partnership with the Biden Administration's Consumer Financial Protection Bureau. The states' investigation found that HCA violated California employment and consumer protection laws as well as the federal consumer financial protection laws by using training repayment agreement provisions (TRAPs) in nurses' employment contracts. These TRAPs are a form of employer-driven debt, or debt obligations incurred by individuals through employment arrangements.

Here is how the California attorney general' s office described HCA's nursing training program and the settlement: As a condition of employment at an HCA hospital, HCA generally requires that entry-level nurse employees complete the Specialty Training Apprenticeship for Registered Nurses (StaRN) Residency Program. The company has advertised StaRN as an avenue for entry-level RNs to get the education and training they need to land their first nursing jobs in an acute-care hospital setting, although StaRN does not provide nurses with education or training necessary for licensure as an RN.

Until the spring of 2023, HCA required that RNs hired through the StaRN program at facilities in several states, including California, sign a TRAP agreement in their new-hire paperwork. The TRAPs purported to require nurses to repay a prorated portion of the StaRN "value" if they did not work for HCA for two years. If a nurse left HCA before the end of the two-year period, then the TRAP loan was typically sent to debt collection.

Nothing much to say that hasn't already been said. We need submissions please.

An Anonymous Coward writes:

Politico reports on a hack affecting Federal Courts in the USA:

The identities of confidential court informants are feared compromised in a series of breaches across multiple U.S. states.

The electronic case filing system used by the federal judiciary has been breached in a sweeping cyber intrusion that is believed to have exposed sensitive court data across multiple U.S. states, according to two people with knowledge of the incident.

The hack, which has not been previously reported, is feared to have compromised the identities of confidential informants involved in criminal cases at multiple federal district courts, said the two people, both of whom were granted anonymity because they were not authorized to speak publicly about the hack.

The Administrative Office of the U.S. Courts — which manages the federal court filing system — first determined how serious the issue was around July 4, said the first person. But the office, along with the Justice Department and individual district courts around the country, is still trying to determine the full extent of the incident.

upstart writes:

Charged Drops Don't Splash:

When a droplet falls on a surface, it spreads itself horizontally into a thin lamella. Sometimes — depending on factors like viscosity, impact speed, and air pressure — that drop splashes, breaking up along its edge into myriad smaller droplets. But a new study finds that a small electrical charge is enough to suppress a drop's splash, as seen below.

The drop's electrical charge builds up along the drop's surface, providing an attraction that acts somewhat like surface tension. As a result, charged drops don't lift off the surface as much and they spread less overall; both factors inhibit splashing.* The effect could increase our control of droplets in ink jet printing, allowing for higher resolution printing.

*Note that this only works for non-conductive surfaces. If the surface is electrically conductive, the charge simply dissipates, allowing the splash to occur as normal.

Journal Reference:
Fanfei Yu, Aaron D. Ratschow, Ran Tao, et al. Why Charged Drops Do Not Splash, Physical Review Letters (DOI: 10.1103/PhysRevLett.134.134001)

Original Submission

upstart writes:

3D printing just went there:

Ever since the popularity of 3D-printing skyrocketed in the mid-aughts, people have manufactured everything from chocolate to rocket fuel—and that list now includes a microscopic elephant inside of a living cell (which you can see here). Technology has really leveled up since 2005.

As new biological opportunities for 3D printing keep emerging, a team of researchers—from the J. Stefan Institute, University of Ljubljana, and CENN Nanocenter in Slovenia—have found a way to pull the process off within a cell's cytoplasm. They were successfully able to print not only an elephant, but several other impossibly small structures using a liqiud polymer and a hyperfocused petawatt laser.

"Intracellular 3D printing offers an unprecedented degree of control over the cellular interior, allowing the integration of synthetic structures with native biological functions," the team said in a study recently posted to the preprint server arXiv. "This platform could allow for reconfiguration of cellular architecture, embed logic or mechanical components within the cytoplasm, and design cells with enhanced or entirely new properties."

For this experiment, the team used a negative photoresist (a material that changes when exposed to certain wavelengths of energy), which became insoluble when exposed to light. It was also the most biocompatible formula possible. After a droplet of photoresist was injected into the cell, an object was printed using a process called two-photon photolithography, which involves targeting an area inside the droplet with a laserto create a microstructure. Anything zapped with two photons from the laser hardens, while any remaining photoresist that has not been lasered into a structure dissolves.

Along with the ironically tiny 10-micrometer elephant, the research team printed other microstructures, like barcodes and a sphere that acted as a micro-laser. The former could eventually allow scientists to track what is going on inside individual cells, and give experts much more detailed insight into cellular function than is currently possible. The latter could be produced in various sizes that all emit light slightly differently, labeling cells with specific light signatures.

Surviving cells continued to go on as if nothing had happened. When a few of them divided, the microstructure inside was passed down to one of the daughter cells. Viability was still an issue, however—even the biocompatible photoresist was still somewhat toxic, and injecting liquid polymer damaged the cell membrane and sometimes caused cell death. How likely cells were to survive depended on the type of cell, and in total, about half of the cells that had microstructures printed in them made it through the experiment.

See also:

• Two-photon 3D printing of functional microstructures inside living cells

Original Submission

upstart writes:

Wikipedia loses challenge against Online Safety Act verification rules:

Wikipedia has lost a legal challenge to new Online Safety Act rules which it says could threaten the human rights and safety of its volunteer editors.

The Wikimedia Foundation - the non-profit which supports the online encyclopaedia - wanted a judicial review of regulations which could mean Wikipedia has to verify the identities of its users.

But it said despite the loss, the judgement "emphasized the responsibility of Ofcom and the UK government to ensure Wikipedia is protected".

The government told the BBC it welcomed the High Court's judgment, "which will help us continue our work implementing the Online Safety Act to create a safer online world for everyone".

Judicial reviews challenge the lawfulness of the way in which a decision has been made by a public body.

In this case the Wikimedia Foundation and a Wikipedia editor tried to challenge the way in which the government decided to make regulations covering which sites should be classed "Category 1" under the Online Safety Act - the strictest rules sites must follow.

It argued the rules were logically flawed and too broad, meaning a policy intended to impose extra rules on large social media companies would instead apply to Wikipedia.

In particular the foundation is concerned the extra duties required - if Wikipedia was classed as Category 1 - would mean it would have to verify the identity of its contributors, undermining their privacy and safety.

The only way it could avoid being classed as Category 1 would be to cut the number of people in the UK who could access the online encyclopaedia by about three-quarters, or disable key functions on the site.

The government's lawyers argued that ministers had considered whether Wikipedia should be exempt from the regulations but had reasonably rejected the idea.

Strange Kaisen Linux 3.0 Announcement Comes With Message of Distro Being Discontinued

An Anonymous Coward writes:

https://distrowatch.com/?newsid=12524
https://archive.ph/pDIIb

Exactly one year after Kaisen Linux's most recent release candidate, the project has announced version 3.0 of its Debian-based, desktop distribution. In an unusual move, the release announcement also includes a report that the distribution is being discontinued. "I would like to begin this blog post by announcing the end of the Kaisen Linux project with this latest release. I wish to embark on other professional and personal projects that will take up a considerable amount of my time, and for this reason, I can no longer continue developing Kaisen Linux. This release will therefore be the last. However, security updates will still be provided for two years, giving you time to switch to another Linux system and familiarize yourself with your new environment." The announcement goes on to share highlights of the new version: "KDE is now the default interface for Kaisen Linux, and is in version 6. SDDM is now the default display manager instead of lightdm. Lightdm was used instead of SDDM due to some missing customization settings, which were introduced with KDE version 6. Xfce is now available in version 4.20...." [...]

Tails Linux OS 7.0~Rc1 Release Candidate 1 Available for Testing

An Anonymous Coward writes:

https://tails.net/news/test_7.0-rc1/

Test 7.0~rc1
2025-08-07

We are very excited to present you with a release candidate of the upcoming Tails 7.0.

We plan to release Tails 7.0 officially on October 16. You can help us by testing this release candidate already.

Tails 7.0 will be the first version of Tails based on Debian 13 (Trixie) and GNOME 48. It will bring new versions of many applications included in Tails.

We have tested 7.0~rc1 with the same extensive automatic and manual test suites that we use for regular releases. But, Tails 7.0~rc1 might still contain undiscovered issues.

We will provide automatic security upgrades for Tails 7.0~rc1, like we do for regular versions.

What is Tails OS?
- https://en.wikipedia.org/wiki/Tails_(operating_system)

Announcement:
https://blog.torproject.org/tails-7_0-rc1-testing/

Changes and updates

*SPOILER* (click to show) *SPOILER* (click to hide)

        Replace GNOME Terminal with GNOME Console. (#20161)

        We broke the Root Terminal while working on this change.

        To open a root terminal, execute the following command in a regular Console.

$sudo -i

        Replace GNOME Image Viewer with GNOME Loupe (#20640)

        Remove Kleopatra from the Favorites menu. (#21072)

        To start Kleopatra choose Apps ▸ Accessories ▸ Kleopatra.

        Remove the obsolete Network Connection option from the Welcome Screen. (#21074)

Included software

        Update the Tor client to 0.4.8.17.

        Update Thunderbird to 128.13.0esr.

        Update the Linux kernel to 6.1.14.

        This improves support for newer hardware: graphics, Wi-Fi, and so on.

        Update Electrum from 4.3.4 to 4.5.8.

        Update OnionShare from 2.6.2 to 2.6.3.

        Update KeePassXC from 2.7.4 to 2.7.10.

        Update Kleopatra from 4:22.12 to 4:24.12

        Update Inkscape from 1.2.2 to 1.4.

        Update GIMP from 2.10.34 to 3.0.4.

        Update Audacity from 3.2.4 to 3.7.3.

        Update Text Editor from 43.2 to 48.3.

        Update Document Scanner from 42.5 to 46.0.

Removed software

        Remove unar. (#20946)

        Remove aircrack-ng. (#21044)

        Remove sq. (#21042)

Fixed problems

        Fix selecting the correct keyboard for certain languages. (#12638)

For more details, see the list of closed issues on the 7.0 milestone in GitLab.

Known issues

        Tails 7.0~rc1 requires 3 GB of RAM instead of 2 GB to run smoothly. (#18040)

        We estimated that less than 2% of current users will be affected.

        Tails 7.0~rc1 takes longer to start.

        We plan to fix this in the final Tails 7.0.

For more details, see the list of issues on the 7.0 milestone in GitLab.

Send your feedback

Please, report any new problem to either:

        tails-testers@boum.org (public mailing list)

        support@tails.net (private email)

Original Submission

Original Submission

upstart writes:

Polluted water was released into loch near Glasgow because Royal Navy failed to maintain 1,500 water pipes, says watchdog:

Radioactive water from the base that holds the UK's nuclear bombs was allowed to leak into the sea after old pipes repeatedly burst, official files have revealed.

The radioactive material was released into Loch Long, a sea loch near Glasgow in western Scotland, because the Royal Navy failed to properly maintain a network of 1,500 water pipes on the base, a regulator found.

The armaments depot at Coulport on Loch Long is one of the most secure and secretive military sites in the UK. It holds the Royal Navy's supply of nuclear warheads for its fleet of four Trident submarines, which are based nearby.

Files compiled by the Scottish Environment Protection Agency (Sepa), a government pollution watchdog, suggest that up to half the components at the base were beyond their design life when the leaks occurred.

Sepa said the flooding at Coulport was caused by "shortfalls in maintenance", resulting in the release of "unnecessary radioactive waste" in the form of low levels of tritium, which is used in nuclear warheads.

In one report in 2022, the agency blamed the leaks on the navy's repeated failure to maintain the equipment in the area devoted to storing the warheads, and said plans to replace 1,500 old pipes at risk of bursting were "sub-optimal".

The leaks are revealed in a cache of confidential inspection reports and emails given to the investigative website the Ferret and shared with the Guardian, which Sepa and the Ministry of Defence fought to keep secret.

[...] The Sepa files show there had been a pipe burst at Coulport in 2010 and a further two in 2019. One leak in August 2019 released "significant amounts of water" that flooded a nuclear weapons processing area, where it became contaminated with low levels of tritium and passed through an open drain that fed into Loch Long.

While Sepa said radioactivity levels in that incident were very low and did not endanger human health, it found there were "shortfalls in maintenance and asset management that led to the failure of the coupling that indirectly led to the production of unnecessary radioactive waste".

hubie writes:

Toothpaste made from your own hair may offer a sustainable and clinically effective way to protect and repair damaged teeth:

In a new study published today [13 August 2025], scientists discovered that keratin, a protein found in hair, skin and wool, can repair tooth enamel and stop early stages of decay.

The King's College London team of scientists discovered that keratin produces a protective coating that mimics the structure and function of natural enamel when it comes into contact with minerals in saliva.

Acidic foods and drinks, poor oral hygiene, and ageing all contribute to enamel erosion and decay, leading to tooth sensitivity, pain and eventually tooth loss.

While fluoride toothpastes are currently used to slow this process, keratin-based treatments were found to stop it completely. Keratin forms a dense mineral layer that protects the tooth and seals off exposed nerve channels that cause sensitivity, offering both structural and symptomatic relief.

The treatment could be delivered through a toothpaste for daily use or as a professionally applied gel, similar to nail varnish, for more targeted repair. The team is already exploring pathways for clinical application and believes that keratin-based enamel regeneration could be made available to the public within the next two to three years.

In their study, published in Advanced Healthcare Materials, the scientists extracted keratin from wool. They discovered that when keratin is applied to the tooth surface and comes into contact with the minerals naturally present in saliva, it forms a highly organised, crystal-like scaffold that mimics the structure and function of natural enamel.

Over time, this scaffold continues to attract calcium and phosphate ions, leading to the growth of a protective enamel-like coating around the tooth. This marks a significant step forward in regenerative dentistry.

[...] Dr Elsharkawy concluded: "We are entering an exciting era where biotechnology allows us to not just treat symptoms but restore biological function using the body's own materials. With further development and the right industry partnerships, we may soon be growing stronger, healthier smiles from something as simple as a haircut.

Journal Reference: Sara Gamea, Elham Radvar, Dimitra Athanasiadou, et al., Biomimetic Mineralization of Keratin Scaffolds for Enamel Regeneration, Advanced Healthcare Materials [OPEN], First published: 12 August 2025 https://doi.org/10.1002/adhm.202502465

Original Submission

An Anonymous Coward writes:

https://www.osnews.com/story/143044/firefox-new-ai-features-cause-cpu-spikes-and-battery-drain/

Almost three weeks ago, Mozilla released Firefox 141 that, among other features like memory optimizations for Linux and a built-in unit converter, brought controversial AI-enhanced tab groups.

Powered by a local AI model, these groups identify related tabs and suggest names for them. There is even a "Suggest more tabs for group" button that users can click to get recommendations.

Now, several users have taken to the Firefox subreddit to complain about high CPU usage when using the feature, as well as express their disappointment in Mozilla for adding AI to the browser.

[...] If you are also dealing with CPU spikes and battery drain from Firefox's new AI features, you can disable them through the browser's advanced settings. Head to about:config in a new tab, accept the risk warning, and use the search bar to find the controls. To kill the AI chatbot feature, search for browser.ml.chat.enabled and set it to false. To stop smart tab grouping, search for browser.tabs.groups.smart.enabled and set it to false.

Original Submission