Semicolon usage in British literature has declined from once every 205 words in 2000 to [1]once every 390 words today , representing a nearly 50% drop, according to analysis commissioned by language learning company Babbel. The punctuation mark appeared once every 90 words in British literature from 1781, making the current frequency the lowest on record.

A survey of young learners in the London Student Network found that more than half of respondents could not correctly use semicolons, with only 11% describing themselves as frequent users. The average score on a semicolon knowledge quiz was 49%.



[1] https://www.smithsonianmag.com/smart-news/could-the-semicolon-die-out-a-recent-study-finds-a-marked-decline-in-its-usage-180986689/

Lawmakers in Britain have narrowly [1]approved a bill to legalize assisted dying for terminally ill people, capping a fraught debate in Parliament and across the country that cut across political, religious and legal divides. From a report:

> MPs passed the bill by 314 votes to 291, in their final say on the question. The bill -- which has split lawmakers and sparked impassioned conversations with their constituents the breadth of Britain -- will now move to the House of Lords for its final rounds of scrutiny.

>

> Friday's vote puts Britain firmly on track to join a small club of nations that have legalized the process, and one of the largest by population to allow it. It allows people with a terminal condition and less than six months to live to take a substance to end their lives, as long as they are capable of making the decision themselves. Two doctors and a panel would need to sign off on the choice. Canada, New Zealand, Spain and most of Australia allow assisted dying in some form, as do several US states, including Oregon, Washington and California.



[1] https://edition.cnn.com/2025/06/20/uk/uk-assisted-dying-commons-vote-gbr-intl

A senior Broadcom executive has defended VMware's controversial licensing changes by arguing that customers complaining about costs simply [1]weren't using the software bundles properly . VMware shifted away from selling perpetual licenses for individual products to subscription bundles after [2]Broadcom's acquisition . Some smaller and mid-sized customers claim their costs increased eight to 15 times under the new pricing structure, prompting migration plans to alternative platforms.

Joe Baguley, Broadcom's chief technology officer for EMEA, countered that 87% of VMware's top 10,000 customers have signed up for VMware Cloud Foundation, and that cost complaints "don't play out" when Broadcom sits down with customers directly.



[1] https://www.theregister.com/2025/06/20/vmware_price_hikes_excuse/

[2] https://slashdot.org/story/22/05/26/1748248/broadcom-to-acquire-vmware-in-massive-61-billion-deal

Australia's world-first social media ban for under-16s moved [1]closer to implementation after a key trial found that checking a user's age is technologically possible and can be integrated into existing services. From a report:

> The conclusions are a blow to Facebook-owner Meta Platforms, TikTok and Snap, which opposed the controversial legislation. Some platform operators had questioned whether a user's age could be reliably established using current technology.

>

> The results of the government-backed trial clear the way for the law to come into force by the end of the year. The findings also potentially allow other jurisdictions to follow Australia's lead as countries around the world grapple with ways to protect children from harmful content online. "Age assurance can be done in Australia and can be private, robust and effective," the government-commissioned Age Assurance Technology Trial said in a statement Friday announcing its preliminary findings.



[1] https://www.bloomberg.com/news/articles/2025-06-19/teen-social-media-ban-moves-closer-in-australia-after-tech-trial

Turning recently closed coalmines into solar energy plants could [1]add almost 300GW of renewable energy by 2030, converting derelict wastelands to productive use, according to a new report. From a report:

> In a first of its kind analysis, researchers from Global Energy Monitor (GEM) identified 312 surface coalmines closed since 2020 around the world, and 134 likely to close by the end of the decade, together covering 5,820 sq km (2,250 sq miles) -- a land area nearly the size of Palestine.

>

> Strip mining turns terrains into wastelands, polluted and denuded of topsoil. But if they were filled with solar panels and developed into energy plants, the report claims, they could generate enough energy to power as big and power hungry a nation as Germany.



[1] https://www.theguardian.com/environment/2025/jun/18/turning-coalmines-into-solar-energy-plants-renewables-gem-report

The DOJ has [1]filed a civil complaint to [2]seize $225.3 million in cryptocurrency linked to pig butchering scams -- long-con frauds where victims are tricked into fake crypto investments. The funds were laundered through a blockchain network, and the DOJ says recovered money will go toward reimbursing victims. The Verge reports:

> The [3]75-page complaint (PDF) filed in the US District Court for the District of Columbia lays out more detail about the seizure. According to it, the US Secret Service (USSS) and Federal Bureau of Investigation (FBI) tied scammers to seven groups of Tether stablecoin tokens. The fraud fell under what's typically known as "pig butchering": a form of long-running confidence scam aimed at tricking victims -- sometimes with a fake romantic relationship -- into what they believe is a profitable crypto investment opportunity, then disappearing with the funds. Pig butchering rings often traffic the workers who directly communicate with victims to Southeast Asian countries, something the DOJ alleges this ring did.

>

> The DOJ says Tether and crypto exchange OKX first alerted law enforcement in 2023 to a series of accounts they believed were helping launder fraudulently obtained currency through a vast and complex web of transactions. The alleged victims include Shan Hanes (referred to in this complaint as S.H.), the former Heartland Tri-State Bank president who was sentenced to 24 years in prison for embezzling tens of millions of dollars to invest in one of the best-known and [4]most devastating pig butchering scams. The complaint lists a number of other victims who lost thousands or millions of dollars they thought they were investing (and did not commit crimes of their own). An [5]FBI report (PDF) cited by the press release concluded overall crypto investment fraud caused $5.8 billion worth of reported losses in 2024.



[1] https://www.justice.gov/opa/pr/united-states-files-civil-forfeiture-complaint-against-225m-funds-involved-cryptocurrency

[2] https://www.theverge.com/news/689956/doj-tether-cryptocurrency-investment-pig-butchering-scam-seizure

[3] https://storage.courtlistener.com/recap/gov.uscourts.dcd.281742/gov.uscourts.dcd.281742.1.0_1.pdf

[4] https://www.nytimes.com/2025/02/19/magazine/cryptocurrency-scam-kansas-heartland-bank.html

[5] https://www.ic3.gov/AnnualReport/Reports/2024_IC3Report.pdf

[1]alternative_right shares a report from ScienceAlert:

> The colossal black hole lurking at the center of the Milky Way galaxy is [2]spinning almost as fast as its maximum rotation rate . That's just one thing astrophysicists have discovered after developing and applying a new method to tease apart the secrets still hidden in supermassive black hole observations collected by the Event Horizon Telescope (EHT). The unprecedented global collaboration spent years working to give us the first direct images of the shadows of black holes, first with M87* in a galaxy 55 million light-years away, then with Sgr A*, the supermassive black hole at the heart of our own galaxy. [...]

>

> Their results show, among other things, that Sgr A* is not only spinning at close to its maximum speed, but that its rotational axis is pointed in Earth's direction, and that the glow around it is generated by hot electrons. Perhaps the most interesting thing is that the magnetic field in the material around Sgr A* doesn't appear to be behaving in a way that's predicted by theory. M87*, they discovered, is also rotating rapidly, although not as fast as Sgr A*. However, it is rotating in the opposite direction to the material swirling in a disk around it -- possibly because of a past merger with another supermassive black hole.

The findings have been detailed in three papers published in the journal Astronomy & Astrophysics. They can be found [3]here , [4]here , and [5]here .



[1] https://slashdot.org/~alternative_right

[2] https://www.sciencealert.com/our-galaxys-monster-black-hole-is-spinning-almost-as-fast-as-physics-allows

[3] https://doi.org/10.1051/0004-6361/202553784

[4] https://doi.org/10.1051/0004-6361/202553785

[5] https://doi.org/10.1051/0004-6361/202553786

The first macOS Tahoe beta [1]appears to drop support for legacy FireWire 400 and 800 , making it impossible to sync or mount older iPods and external drives that rely on the standard. MacRumors reports:

> Unlike on macOS Sequoia and earlier versions, the first macOS Tahoe beta does not include a FireWire section in the System Settings app. Of course, this could all end up being a false alarm. It is still early in the macOS Tahoe beta testing cycle, and FireWire support could return in a later beta version, or in time for the final release.

>

> FireWire was primarily developed by Apple, but it was later standardized as IEEE 1394 and licensed for use in non-Apple devices. iPods started to transition from FireWire to USB for data transfer in 2003, so the standard is very outdated, but it would still be the end of an era if macOS Tahoe drops it. The last Mac with a FireWire port was released in 2012, so connecting older iPods and FireWire drives to newer Macs has long required the use of adapters.



[1] https://www.macrumors.com/2025/06/19/macos-tahoe-beta-drops-firewire-support/

Longtime Slashdot reader [1]fahrbot-bot shares a report from the Cool Down:

> A team of chemical engineers at the Massachusetts Institute of Technology has [2]invented a new process to separate crude oil components , potentially bringing forward a replacement that can cut its harmful carbon pollution by 90%. The original technique, which uses heat to separate crude oil into gasoline, diesel, and heating oil, accounts for roughly 1% of all global energy consumption and 6% of dirty energy pollution from the carbon dioxide it releases.

>

> "Instead of boiling mixtures to purify them, why not separate components based on shape and size?" said Zachary P. Smith, associate professor of chemical engineering at MIT and senior author of the [3]study , as previously [4]reported in Interesting Engineering. The team invented a polymer membrane that divides crude oil into its various uses like a sieve. The new process follows a similar strategy used by the water industry for desalination, which uses reverse osmosis membranes and has been around since the 1970s. [The membrane excelled in lab tests. It increased the toluene concentration by 20 times in a mixture with triisopropylbenzene. It also effectively separated real industrial oil samples containing naphtha, kerosene, and diesel.]



[1] https://slashdot.org/~fahrbot-bot

[2] https://www.thecooldown.com/green-tech/crude-oil-separation-mit-chemical-engineers/

[3] https://www.science.org/doi/10.1126/science.adv6886

[4] https://interestingengineering.com/energy/crude-oil-energy-reduction-membrane

[1]alternative_right shares a report from ScienceAlert:

> A team of biologists from Northeastern University and the University of Kentucky has [2]found one of the key molecules involved in axolotl regeneration . It's a crucial component in ensuring the body grows back the right parts in the right spot: for instance, growing a hand, from the wrist. "The cells can interpret this cue to say, 'I'm at the elbow, and then I'm going to grow back the hand' or 'I'm at the shoulder... so I'm going to then enable those cells to grow back the entire limb'," biologist James Monaghan explains.

>

> That molecule, retinoic acid, is arranged through the axolotl body in a gradient, signaling to regenerative cells how far down the limb has been severed. Closer to the shoulder, axolotls have higher levels of retinoic acid, and lower levels of the enzyme that breaks it down. This ratio changes the further the limb extends from the body. The team found this balance between retinoic acid and the enzyme that breaks it down plays a crucial role in 'programming' the cluster of regenerative cells that form at an injury site. When they added surplus retinoic acid to the hand of an axolotl in the process of regenerating, it grew an entire arm instead.

>

> In theory, the human body has the right molecules and cells to do this too, but our cells respond to the signals very differently, instead forming collagen-based scars at injury sites. Next, Monaghan is keen to find out what's going on inside cells -- the axolotl's, and our own -- when those retinoic acid signals are received.

The research is [3]published in Nature Communications .



[1] https://slashdot.org/~alternative_right

[2] https://www.sciencealert.com/axolotl-discovery-brings-us-closer-than-ever-to-regrowing-human-limbs

[3] https://doi.org/10.1038/s41467-025-59497-5

An anonymous reader quotes a report from the BBC:

> The Earth could be [1]doomed to breach the symbolic 1.5C warming limit in as little as three years at current levels of carbon dioxide emissions. That's the stark warning from more than 60 of the world's leading climate scientists in the most up-to-date assessment of the state of global warming. [...] At the beginning of 2020, scientists estimated that humanity could only emit 500 billion more tonnes of carbon dioxide (CO2) -- the most important planet-warming gas -- for a 50% chance of keeping warming to 1.5C. But by the start of 2025 this so-called "carbon budget" had shrunk to 130 billion tonnes, according to the new study.

>

> That reduction is largely due to continued record emissions of CO2 and other planet-warming greenhouse gases like methane, but also improvements in the scientific estimates. If global CO2 emissions stay at their current highs of about 40 billion tonnes a year, 130 billion tonnes gives the world roughly three years until that carbon budget is exhausted. This could commit the world to breaching the target set by the Paris agreement, the researchers say, though the planet would probably not pass 1.5C of human-caused warming until a few years later.

>

> Last year was the first on record when global average air temperatures were more than 1.5C above those of the late 1800s. A single 12-month period isn't considered a breach of the Paris agreement, however, with the record heat of 2024 given an extra boost by natural weather patterns. But human-caused warming was by far the main reason for last year's high temperatures, reaching 1.36C above pre-industrial levels, the researchers estimate. This current rate of warming is about 0.27C per decade -- much faster than anything in the geological record. And if emissions stay high, the planet is on track to reach 1.5C of warming on that metric around the year 2030. After this point, long-term warming could, in theory, be brought back down by sucking large quantities of CO2 back out of the atmosphere. But the authors urge caution on relying on these ambitious technologies serving as a get-out-of-jail card.

"For larger exceedance [of 1.5C], it becomes less likely that removals [of CO2] will perfectly reverse the warming caused by today's emissions," warned Joeri Rogelj, professor of climate science and policy at Imperial College London.

"Reductions in emissions over the next decade can critically change the rate of warming," he added. "Every fraction of warming that we can avoid will result in less harm and less suffering of particularly poor and vulnerable populations and less challenges for our societies to live the lives that we desire."



[1] https://www.bbc.com/news/articles/cn4l927dj5zo

Publishers face [1]an existential threat in the AI era and need to take action to make sure they are fairly compensated for their content, Cloudflare CEO Matthew Prince told Axios at an event in Cannes on Thursday. From a report:

> Search traffic referrals have plummeted as people increasingly rely on AI summaries to answer their queries, forcing many publishers to reevaluate their business models. Ten years ago, Google crawled two pages for every visitor it sent a publisher, per Prince.

>

> He said that six months ago:

> For Google that ratio was 6:1

> For OpenAI, it was 250:1

> For Anthropic, it was 6,000:1

>

> Now:

>

> For Google, it's 18:1

> For OpenAI, it's 1,500:1

> For Anthropic, it's 60,000:1

>

> Between the lines: "People aren't following the footnotes," Prince said.



[1] https://www.axios.com/2025/06/19/ai-search-traffic-publishers

An anonymous reader quotes a report from the Hollywood Reporter:

> Bruce Lee, Jackie Chan and Jet Li and a legion of the all-time greats of martial cinema are about to get an AI makeover. In a sign-of-the-times announcement at the Shanghai International Film Festival on Thursday, a collection of Chinese studios revealed that they are [1]turning to AI to re-imagine around 100 classics of the genre . Lee's classic Fist of Fury (1972), Chan's breakthrough Drunken Master (1978) and the Tsui Hark-directed epic Once Upon a Time in China (1991), which turned Li into a bone fide movie star, are among the features poised for the treatment, as part of the "Kung Fu Movie Heritage Project 100 Classics AI Revitalization Project."

>

> There will also be a digital reworking of the John Woo classic A Better Tomorrow (1986) that, by the looks of the trailer, turns the money-burning anti-hero originally played by Chow Yun-fat into a cyberpunk, and is being claimed as "the world's first full-process, AI-produced animated feature film." The big guns of the Chinese industry were out in force on the sidelines of the 27th Shanghai International Film Festival to make the announcements, too. They were led by Zhang Pimin, chairman of the China Film Foundation, who said AI work on these "aesthetic historical treasures" would give them a new look that "conforms to contemporary film viewing." "It is not only film heritage, but also a brave exploration of the innovative development of film art," Zhang said.

>

> Tian Ming, chairman of project partners Shanghai Canxing Culture and Media, meanwhile, promised the work -- expected to include upgrades in image and sound as well as overall production levels -- while preserving the storytelling and aesthetic of the originals -- would both "pay tribute to the original work" and "reshape the visual aesthetics." "We sincerely invite the world's top AI animation companies to jointly start a film revolution that subverts tradition," said Tian, who announced a fund of 100 million yuan ($13.9 million) would be implemented to kick-start the work.



[1] https://www.hollywoodreporter.com/movies/movie-news/china-ai-remakes-kung-fu-films-bruce-lee-jackie-chan-jet-li-1236295093/

Starting mid-July 2025, Microsoft 365 will [1]begin blocking legacy authentication protocols like Remote PowerShell and FrontPage RPC to enhance security under its "Secure by Default" initiative. Admins must now grant explicit consent for third-party app access, which could disrupt workflows but aims to reduce unauthorized data exposure. The Register reports:

> First in line for the chop is legacy browser authentication to SharePoint and OneDrive using the Remote PowerShell (RPS) protocol. According to Microsoft, legacy authentication protocols like RPS "are vulnerable to brute-force and phishing attacks due to non-modern authentication." The upshot is that attempting to access OneDrive or SharePoint via a browser using legacy authentication will stop working.

>

> Also being blocked is the FrontPage Remote Procedure Call (RPC) protocol. Microsoft FrontPage was a web authoring tool that was discontinued almost two decades ago. However, the protocol for remote web authoring has lived on until now. Describing legacy protocols like RPC as "more susceptible to compromise," Microsoft will block them to prevent their use in Microsoft 365 clients.

>

> Finally, third-party apps will need administrator consent to access files and sites. Microsoft said: "Users allowing third-party apps to access file and site content can lead to overexposure of an organization's content. Requiring admins to consent to this access can help reduce overexposure."

"While laudable, shifting consent to the administrator could disrupt some workflows," writes The Register's Richard Speed. "The Microsoft-managed App Consent Policies will be enabled, and users will be unable to consent to third-party applications accessing their files and sites by default. Need consent? A user will need to request an administrator to consent on their behalf."



[1] https://www.theregister.com/2025/06/19/the_microsoft_axeman_is_coming/

SpaceX's Starship [1]exploded on its test stand in South Texas ahead of an engine test, marking the fourth loss of a Starship this year. "In three previous test flights, the vehicle [2]came apart or [3]detonated during its flight ," notes the Washington Post. No injuries were reported but the incident highlights ongoing technical challenges as SpaceX races to prove Starship's readiness for deep-space travel. From the report:

> In [4]a post on the social media site X, SpaceX said that the explosion on the test stand, which could be seen for miles, happened at about 11 p.m. Central time. For safety reasons, the company had cleared personnel from around the site, and "all personnel are safe and accounted for," it said. The company is "actively working to safe the test site and the immediate surrounding area in conjunction with local officials," the post continued. "There are no hazards to residents in surrounding communities, and we ask that individuals do not attempt to approach the area while safing operations continue."

>

> Starship comprises two stages -- the Super Heavy booster, which has 33 engines, and the Starship spacecraft itself, which has six. Before Wednesday's explosion, the spacecraft was standing alone on the test stand, and not mounted on top of the booster, when it blew up. The engines are test-fired on the Starship before it's mounted on the booster. SpaceX had been hoping to launch within the coming weeks had the engine test been successful. [...] In a post on X, Musk [5]said that preliminary data pointed to a pressure vessel that failed at the top of the rocket.

You can watch a recording of the explosion [6]on YouTube .

SpaceX called the incident a " [7]rapid unscheduled disassembly ," which caught the attention of Slashdot reader [8]hambone142 . In a story [9]submitted to the Firehose, they commented: "I worked for a major computer company whose power supplies caught on fire. We were instructed to cease saying that and instead say the power supply underwent a 'thermal event.' Gotta love it."



[1] https://www.washingtonpost.com/technology/2025/06/19/starship-spacex-explosion-musk/

[2] https://science.slashdot.org/story/25/05/28/024205/spacex-starship-blasts-off-in-ninth-test-flight

[3] https://science.slashdot.org/story/25/03/07/0422209/spacexs-latest-starship-test-flight-ends-with-another-explosion

[4] https://x.com/SpaceX/status/1935572705941880971

[5] https://x.com/elonmusk/status/1935660973827952675

[6] https://www.youtube.com/watch?v=VcJoZdfbKQw

[7] https://x.com/SpaceX/status/1880033318936199643

[8] https://slashdot.org/~hambone142

[9] https://slashdot.org/submission/17337109/rapid-unscheduled-disassembly-of-a-starship-rocket

An anonymous reader quotes a report from Cybernews:

> Several collections of login credentials reveal one of the largest data breaches in history, [1]totaling a humongous 16 billion exposed login credentials . The data most likely originates from various infostealers. Unnecessarily compiling sensitive information can be as damaging as actively trying to steal it. For example, the Cybernews research team discovered a plethora of supermassive datasets, housing billions upon billions of login credentials. From social media and corporate platforms to VPNs and developer portals, no stone was left unturned.

>

> Our team has been closely monitoring the web since the beginning of the year. So far, they've discovered 30 exposed datasets containing from tens of millions to over 3.5 billion records each. In total, the researchers uncovered an unimaginable 16 billion records. None of the exposed datasets were reported previously, bar one: in late May, Wired magazine reported a security researcher discovering a "mysterious database" with 184 million records. It barely scratches the top 20 of what the team discovered. Most worryingly, researchers claim new massive datasets emerge every few weeks, signaling how prevalent infostealer malware truly is.

>

> "This is not just a leak -- it's a blueprint for mass exploitation. With over 16 billion login records exposed, cybercriminals now have unprecedented access to personal credentials that can be used for account takeover, identity theft, and highly targeted phishing. What's especially concerning is the structure and recency of these datasets -- these aren't just old breaches being recycled. This is fresh, weaponizable intelligence at scale," researchers said. The only silver lining here is that all of the datasets were exposed only briefly: long enough for researchers to uncover them, but not long enough to find who was controlling vast amounts of data. Most of the datasets were temporarily accessible through unsecured Elasticsearch or object storage instances.

Key details to be aware of:

> - The records include billions of login credentials, often structured as URL, login, and password.

> - The datasets include both old and recent breaches, many with cookies, tokens, and metadata, making them especially dangerous for organizations without multi-factor authentication or strong credential practices.

> - Exposed services span major platforms like Apple, Google, Facebook, Telegram, GitHub, and even government services.

> - The largest dataset alone includes 3.5 billion records, while one associated with the Russian Federation has over 455 million; many dataset names suggest links to malware or specific regions.

> - Ownership of the leaked data is unclear, but its potential for phishing, identity theft, and ransomware is severe -- especially since even a - Basic cyber hygiene -- such as regularly updating strong passwords and scanning for malware -- is currently the best line of defense for users.



[1] https://cybernews.com/security/billions-credentials-exposed-infostealers-data-leak/

Hackers have stolen hundreds of millions of dollars from cryptocurrency holders and disrupted major retailers by [1]targeting outsourced call centers used by American corporations to reduce costs, WSJ reported Thursday. The attackers exploit low-paid call center workers through bribes and social engineering to bypass two-factor authentication systems protecting bank accounts and online portals.

Coinbase [2]faces potential losses of $400 million after [3]hackers compromised data belonging to 97,000 customers by [4]bribing call center workers in India with payments of $2,500. The criminals also used malicious tools that exploited vulnerabilities in Chrome browser extensions to collect customer data in bulk.

TaskUs, which handled Coinbase support calls, shut down operations at its Indore, India facility and laid off 226 workers. Retail attacks [5]targeted Marks & Spencer and Harrods with hackers impersonating corporate executives to pressure tech support workers into providing network access. The same technique [6]compromised MGM Resorts systems in 2023 . Call center employees typically possess sensitive customer information including account balances and recent transactions that criminals use to masquerade as legitimate company representatives.



[1] https://www.msn.com/en-us/money/news/hackers-are-using-apple-and-microsoft-to-scam-millions-here-s-how-to-protect-yourself/ar-AA1rtYms

[2] https://yro.slashdot.org/story/25/05/20/2052220/coinbase-data-breach-will-lead-to-people-dying-techcrunch-founder-says

[3] https://it.slashdot.org/story/25/05/15/1415242/coinbase-offers-20-million-bounty-to-catch-data-thieves-after-extortion-attempt

[4] https://yro.slashdot.org/story/25/06/03/0135257/coinbase-breach-linked-to-customer-data-leak-in-india

[5] https://it.slashdot.org/story/25/05/14/1050202/marks-and-spencer-confirms-hackers-stole-customers-personal-data-cyber-insurance-payout-to-be-worth-up-to-133-million

[6] https://it.slashdot.org/story/23/09/14/0120204/hackers-claim-it-only-took-a-10-minute-phone-call-to-shut-down-mgm-resorts

Google is using its expansive library of YouTube videos to [1]train its AI models , including Gemini and the Veo 3 video and audio generator, CNBC reported Thursday. From the report:

> The tech company is turning to its catalog of 20 billion YouTube videos to train these new-age AI tools, according to a person who was not authorized to speak publicly about the matter. Google confirmed to CNBC that it relies on its vault of YouTube videos to train its AI models, but the company said it only uses a subset of its videos for the training and that it honors specific agreements with creators and media companies.

>

> [...] YouTube didn't say how many of the 20 billion videos on its platform or which ones are used for AI training. But given the platform's scale, training on just 1% of the catalog would amount to 2.3 billion minutes of content, which experts say is more than 40 times the training data used by competing AI models.



[1] https://www.cnbc.com/2025/06/19/google-youtube-ai-training-veo-3.html

Simon Willison, [1]commenting on the recent paper from Apple researchers that found state-of-the-art large language models [2]face complete performance collapse beyond certain complexity thresholds:

> I thought this paper got way more attention than it warranted -- the title "The Illusion of Thinking" captured the attention of the "LLMs are over-hyped junk" crowd. I saw enough well-reasoned rebuttals that I didn't feel it worth digging into.

>

> And now, notable LLM skeptic Gary Marcus has saved me some time by aggregating the best of those rebuttals [3]together in one place!

>

> [...] And therein lies my disagreement. I'm not interested in whether or not LLMs are the "road to AGI". I continue to care only about whether they have useful applications today, once you've understood their limitations.

>

> Reasoning LLMs are a relatively new and interesting twist on the genre. They are demonstrably able to solve a whole bunch of problems that previous LLMs were unable to handle, hence why we've seen [4]a rush of new models from OpenAI and Anthropic and Gemini and DeepSeek and Qwen and Mistral.

>

> They get even more interesting when you [5]combine them with tools .

>

> They're already useful to me today, whether or not they can reliably solve the Tower of Hanoi or River Crossing puzzles.



[1] https://simonwillison.net/2025/Jun/15/viral-apple-reasoning-paper/

[2] https://apple.slashdot.org/story/25/06/09/1151210/apple-researchers-challenge-ai-reasoning-claims-with-controlled-puzzle-tests

[3] https://garymarcus.substack.com/p/seven-replies-to-the-viral-apple

[4] https://simonwillison.net/tags/llm-reasoning/

[5] https://simonwillison.net/2025/Jun/6/six-months-in-llms/#ai-worlds-fair-2025-43.jpeg

ChatGPT-assisted writing dampened brain activity and recall in [1]a controlled MIT study [PDF] of 54 college volunteers divided into AI-only, search-engine, and no-tool groups. Electroencephalography recorded during three essay-writing sessions found the AI group consistently showed the weakest neural connectivity across all measured frequency bands; the tool-free group showed the strongest, with search users in between.

In the first session 83% of ChatGPT users could not quote any line they had just written and none produced a correct quote. Only nine of the 18 claimed full authorship of their work, compared with 16 of 18 in the brain-only cohort. Neural coupling in the AI group declined further over repeated use. When these participants were later asked to write without assistance, frontal-parietal networks remained subdued and 78% again failed to recall a single sentence accurately.

The pattern reversed for students who first wrote unaided: introducing ChatGPT in a crossover session produced the highest connectivity sums in alpha, theta, beta and delta bands, indicating intense integration of AI suggestions with prior knowledge. The MIT authors warn that habitual reliance on large language models "accumulates cognitive debt," trading immediate fluency for weaker memory, reduced self-monitoring, and narrowed neural engagement.



[1] https://arxiv.org/pdf/2506.08872v1