Google unveiled Gemini 2.0 yesterday, almost exactly one year after Google's initial Gemini launch. The new release offers enhanced multimodal capabilities like native image and audio output, real-time tool use, and advanced reasoning to enable agentic experiences, such as acting as a universal assistant or research companion. VentureBeat reports: During a recent press conference, Tulsee Doshi, director of product management for Gemini, outlined the system's enhanced capabilities while demonstrating real-time image generation and multilingual conversations. "Gemini 2.0 brings enhanced performance and new capabilities like native image and multilingual audio generation," Doshi explained. "It also has native intelligent tool use, which means that it can directly access Google products like search or even execute code."

The initial release centers on Gemini 2.0 Flash, an experimental version that Google claims operates at twice the speed of its predecessor while surpassing the capabilities of more powerful models. This represents a significant technical achievement, as previous speed improvements typically came at the cost of reduced functionality. Perhaps most significantly, Google introduced three prototype AI agents built on Gemini 2.0's architecture that demonstrate the company's vision for AI's future. Project Astra, an updated universal AI assistant, showcased its ability to maintain complex conversations across multiple languages while accessing Google tools and maintaining contextual memory of previous interactions. [...]

For developers and enterprise customers, Google introduced Project Mariner and Jules, two specialized AI agents designed to automate complex technical tasks. Project Mariner, demonstrated as a Chrome extension, achieved an impressive 83.5% success rate on the WebVoyager benchmark for real-world web tasks -- a significant improvement over previous attempts at autonomous web navigation. Supporting these advances is Trillium, Google's sixth-generation Tensor Processing Unit (TPU), which becomes generally available to cloud customers today. The custom AI accelerator represents a massive investment in computational infrastructure, with Google deploying over 100,000 Trillium chips in a single network fabric.

Social media giants confronted a familiar dilemma over user content moderation after murder suspect Luigi Mangione's arrest in the killing of UnitedHealthcare's CEO on Monday, highlighting the platforms' varied approaches to managing digital footprints of criminal suspects.

Meta quickly removed Mangione's Facebook and Instagram accounts under its "dangerous organizations and individuals" policy, while his account on X underwent a brief suspension before being reinstated with a premium subscription. LinkedIn maintained his profile, stating it did not violate platform policies. His Reddit account was suspended in line with the platform's policy on high-profile criminal suspects, while his Goodreads profile fluctuated between public and private status.

The New York Times adds: When someone goes from having a private life to getting public attention, online accounts they intended for a small circle of friends or acquaintances are scrutinized by curious strangers -- and journalists.

In some cases, these newly public figures or their loved ones can shut down the accounts or make them private. Others, like Mr. Mangione, who has been charged with murder, are cut off from their devices, leaving their digital lives open for the public's consumption. Either way, tech companies have discretion in what happens to the account and its content. Section 230 of the Communications Decency Act protects companies from legal liability for posts made by users.

An anonymous reader shares a report: BeReal, the in the moment social media platform, is far from its 2022 heyday, but that hasn't stopped one organization from going after it. Austrian advocacy group Noyb has filed a complaint surrounding the platform's data consent banner practices. The organization claims that the banner disappears if users accept that their personal data can inform advertising practices, but if they click reject then the banner appears daily.

Noyb filed its complaint with the French data protection authority (CNIL) as Voodoo, a French company, bought BeReal in June -- the practice in question started in July. "BeReal's daily attempt to pressure its users into accepting the tracking for personalised advertising has a significant impact on user behaviour. Consent given under these circumstances is not freely given, which means it doesn't meet the requirements established in Article 4(11) GDPR," Noyb argued in its complaint. It asked the CNIL to fine BeReal and force it to be compliant.

An anonymous reader shares a report: Former Intel chief Pat Gelsinger, who stepped down from his leadership post a week ago, is inviting people to join him in prayer and fasting for the struggling chipmaker's employees.

"Every Thursday I do a 24 hour prayer and fasting day," Gelsinger wrote on X on Sunday morning. "This week I'd invite you to join me in praying and fasting for the 100K Intel employees as they navigate this difficult period. Intel and its team is of seminal importance to the future of the industry and US."

FlatpanelsHD: LG has discontinued all Blu-ray players, including the UBK80 and UBK90 UHD Blu-ray players, with remaining units only available while stocks last.

The announcement echoes similar moves from Oppo in 2018 and Samsung in 2019, when both companies exited the optical disc player market. LG has now officially discontinued its Blu-ray and UHD Blu-ray players, as reflected on LG's online portals and confirmed by multiple sources to FlatpanelsHD.

Harvard University announced Thursday it's releasing a high-quality dataset of nearly one million public-domain books that could be used by anyone to train large language models and other AI tools. From a report: The dataset was created by Harvard's newly formed Institutional Data Initiative with funding from both Microsoft and OpenAI. It contains books scanned as part of the Google Books project that are no longer protected by copyright.

Around five times the size of the notorious Books3 dataset that was used to train AI models like Meta's Llama, the Institutional Data Initiative's database spans genres, decades, and languages, with classics from Shakespeare, Charles Dickens, and Dante included alongside obscure Czech math textbooks and Welsh pocket dictionaries. Greg Leppert, executive director of the Institutional Data Initiative, says the project is an attempt to "level the playing field" by giving the general public, including small players in the AI industry and individual researchers, access to the sort of highly-refined and curated content repositories that normally only established tech giants have the resources to assemble. "It's gone through rigorous review," he says.

Leppert believes the new public domain database could be used in conjunction with other licensed materials to build artificial intelligence models. "I think about it a bit like the way that Linux has become a foundational operating system for so much of the world," he says, noting that companies would still need to use additional training data to differentiate their models from those of their competitors.

Photobucket was sued Wednesday after a recent privacy policy update revealed plans to sell users' photos -- including biometric identifiers like face and iris scans -- to companies training generative AI models. From a report: The proposed class action seeks to stop Photobucket from selling users' data without first obtaining written consent, alleging that Photobucket either intentionally or negligently failed to comply with strict privacy laws in states like Illinois, New York, and California by claiming it can't reliably determine users' geolocation.

Two separate classes could be protected by the litigation. The first includes anyone who ever uploaded a photo between 2003 -- when Photobucket was founded -- and May 1, 2024. Another potentially even larger class includes any non-users depicted in photographs uploaded to Photobucket, whose biometric data has also allegedly been sold without consent.

Photobucket risks huge fines if a jury agrees with Photobucket users that the photo-storing site unjustly enriched itself by breaching its user contracts and illegally seizing biometric data without consent. As many as 100 million users could be awarded untold punitive damages, as well as up to $5,000 per "willful or reckless violation" of various statutes.

An anonymous reader shares a report: The news came by Slack message. Cruise CEO Marc Whitten, who took the top post in June, posted a message Tuesday afternoon in the company's announcements channel along with a link to a press release entitled "GM to refocus autonomous driving development on personal vehicles."

GM, which acquired the self-driving car startup in 2016, would no longer fund the company, ending a mission that hundreds of Cruise engineers had worked on for years. Minutes later, during an all-hands meeting, Cruise employees learned a few more details. The self-driving car company would be absorbed into parent company GM and combined with the automaker's own efforts to develop driver assistance features -- and eventually fully autonomous personal vehicles. Whether their jobs would be safe or cut was, and still is, unclear.

That meeting was short and unsatisfactory, according to one source, who noted that the senior leadership team was also surprised by this turn of events. Whitten, president and chief technology officer Mo Elshenawy, and chief administrative officer Craig Glidden, led the all-hands. Several Cruise employees who spoke to TechCrunch on condition of anonymity said they were "surprised" and "blindsided" by the decision. One source told TechCrunch that employees learned about GM's plans the same time the media did.

WP Engine just won a preliminary injunction against WordPress.com parent company Automattic. On Tuesday, a California District Court judge ordered Automattic to stop blocking WP Engine's access to WordPress.org resources and interfering with its plugins. From a report: The preliminary injunction comes after WP Engine, a third-party WordPress hosting service, filed a lawsuit that accused Automattic and its CEO, Matt Mullenweg, of "multiple forms of immediate irreparable harm." It later asked the court to stop Mullenweg from restricting WP Engine's access to WordPress.org.

Mullenweg waged a public campaign against WP Engine in September, accusing the service of misusing the WordPress trademark and not contributing enough to the WordPress community. After blocking WP Engine from WordPress.org's servers, Automattic took control of WP Engine's ACF Plugin.

After spending more than $10 billion on its robotaxi unit, General Motors is abandoning its Cruise driverless ride-hailing service. From a report: The Detroit automaker on Tuesday said it will no longer fund its Cruise division's robotaxi development and will instead fold the unit into its broader tech team. "Cruise was well on its way to a robotaxi business -- but when you look at the fact you're deploying a fleet, there's a whole operations piece of doing that," GM CEO Mary Barra said on a call Tuesday.

Barra said GM would instead focus on the development of autonomous systems for use in personal vehicles. GM cited the increasingly competitive robotaxi market, capital allocation priorities and the considerable time and resources necessary to grow the business as reasons for its decision.

Cable industry lobbyists have urged the Federal Communications Commission to avoid regulating data caps and overage charges, comparing broadband plans to restaurant menus in a filing last week.

NCTA - The Internet & Television Association argued that usage-based pricing benefits low-income consumers by providing cheaper options, pushing back against advocacy groups who say data caps disproportionately harm price-sensitive users. The group likened different pricing models to restaurants offering tasting menus, buffets, or unlimited soup and salad.

Consumer advocates, including Public Knowledge and Free Press, countered that low-income households often have no choice but to accept data caps since lower-priced plans typically include usage limits. They cited examples of users like Gloria Simmons, a Georgia retiree who pays $60 monthly for internet service plus $10 for every 50 gigabytes over her data allowance.

An anonymous reader shares a report: Amazon on Tuesday said it is piloting a quick commerce service in India that will see the U.S. tech giant delivering grocery and other items in 15 minutes or less.

[...] The quick-commerce model -- delivering items to customers within 10 to 15 minutes -- hasn't worked in most parts of the world, but it's increasingly finding success in India, where a range of retailers and internet firms, from food delivery giant Swiggy to online cosmetics platform Nykaa, are gearing up their supply chain ecosystems to accommodate for faster deliveries.

As tech giants struggle to profit from AI, a growing industry of specialized AI training firms is emerging by hiring doctors, radiologists and other experts to develop commercially viable applications. The $20 billion data services sector, projected to grow 20% annually, is attracting major investment by focusing on high-value, specialized AI applications.

Companies like iMerit and Centaur Labs are recruiting specialists worldwide, from radiologists in Kazakhstan to agricultural experts in Bhutan, paying premium rates for domain expertise rather than basic data processing. While Microsoft and Alphabet post losses on AI development, this specialized sector is finding profitability by bridging the gap between raw AI capabilities and practical business applications.

"Mercedes-Benz is researching a new type of solar modules that could be seamlessly applied to the bodywork of electric vehicles," according to a recent Mercedes-Benz press release.

They describe the 5-micrometer coating as "similar to a wafer-thin layer of paste... significantly thinner than a human hair" — but creating an active photovoltaic surface with an efficiency of 20%. An area of 118.4 square feet (equivalent to the surface of a mid-size SUV) could produce energy for up to 7,456 miles per year under ideal conditions [based on daylight conditions from their testing in Stuttgart]. The energy generated by the solar cells is used for driving or fed directly into the high-voltage battery...

Solar paint has a high level of efficiency and contains no rare earths or silicon — only non-toxic, readily available raw materials. It is easy to recycle and considerably cheaper to produce than conventional solar modules. The Mercedes-Benz research department is currently working to enable use of the new solar paint on all exterior vehicle surfaces — regardless of shape or angle.
Solar paint could power 62% of the travel for a typical Stuttgart driver, their announcement notes. But in a sunnier city like Los Angeles, "It could be used for 100% of their driving, on average." (And "the surplus of energy could be fed directly into the home network via bidirectional charging.")

Mercedes-Benz researchers "initially thought the tech had limited scope for mass production," reports EV Central, "until experiments were carried out with prototypes coated with the paint in real-world scenarios. Instead of just coating the roof and bonnet to form a 1.8-square-metre surface area, one scientist suggested covering an entire car with the new solar paint, ramping up the surface area to more than 11m2. Another difference to the [Mercedes-Benz 2022 Vision EQXX concept] is instead of wiring the body panels to the 12-volt system, scientists hardwired the body panels to the Benz's high-voltage battery and the performance of the paint was well beyond expectations... Available in all colours, engineers admit the solar paint work best in darker shades. When it's launched, the tech should be as durable as regular paint. The photovoltaic surface is protected by at least two protective lacquer finishes to ensure it isn't damaged in daily use.
Mercedes-Benz says the solar paint could mean "increased electric range and fewer charging stops." And this is significant, because "Electric vehicle charging and infrastructure are two major obstacles to EV adoption on a mass scale," writes Autoblog — arguing that Mercedes-Benz "may have a solution... " Alternative methods of energy harnessing could help alleviate range anxiety, increase an EV's driving distance, and reduce charging costs across the board. Not only that but considering the cost of producing Mercedes' solar coating and the lack of rare earth metals, it could be the leading solution to charging concerns... While the German automaker says the solar paint isn't ready for production on a mass scale, research, and development are progressing at a steady rate. If all goes well, we'll hopefully see solar coating as an accessory EV charging solution within the next decade.
Thanks to long-time Slashdot reader schwit1 for sharing the news.

Google's app store claims that their text-messaging app Google Messages means "conversations are end-to-end encrypted".

"That is some serious bullshit," argues tech blogger John Gruber: It's shamefully misleading regarding Google Messages's support for end-to-end encryption... Google Messages does support end-to-end encryption, but only over RCS and only if all participants in the chat are using a recent version of Google Messages. But the second screenshot in the Play Store listing flatly declares "Conversations are end-to-end encrypted", full stop...

I realize that "Some conversations are end-to-end encrypted" will naturally spur curiosity regarding which conversations are encrypted and which aren't, but that's the truth. And users of the app should be aware of that. "RCS conversations with other Google Messages users are encrypted" would work.

Then, in the "report card" section of the listing, it states the following:

Data is encrypted in transit
Your data is transferred over a secure connection

Which, again, is only true sometimes. It's downright fraudulent to describe Google Messages's transit security this way.... [D]epending who you communicate with — iPhone users, Android users with old devices, Android users who use other text messaging apps — it's quite likely most of your messages won't be secure... E2EE is never available for SMS, and never available if a participant in the chat is using any RCS client (on Android or Apple Messages) other than Google Messages. That's an essential distinction that should be made clear, not obfuscated.
Gruber's earlier blog post had pointed out that the RCS standard "has no encryption; E2EE RCS chats in Google Messages use Google's proprietary extension and are exclusive to the Google Messages app, so RCS chats between Google Messages and other apps, most conspicuously Apple Messages, are not encrypted."

And in his newer post, Gruber adds, "While I'm at it, it's also embarrassing that Google Voice has no support for RCS at all. It's Google's own app and service, and Google has been the world's most vocal proponent of RCS messaging."

This week a Florida-based Bitcoin-tech company named MARA Holdings announced it had bought a 114-megawatt Texas wind farm, reports Chron.com, "and will subsequently take it off the power grid and use it to energize its mining operations."

MARA's CEO tells the site they're "leveraging renewable resources that would have otherwise been curtailed" while "reducing our bitcoin production costs through vertical integration, and demonstrating MARA's commitment to environmental stewardship." The wind farms were not a part of the Electric Reliability Council of Texas (ERCOT) grid, but instead they were located within the Southwest Power Pool, which manages the market for the central U.S., including but not limited to most or parts of Oklahoma, Kansas, Nebraska, South Dakota and North Dakota... A 114-MW facility could power somewhere between 20,000 and 100,000 homes, depending on who you ask...

Historically, the facilities use up a lot of power and have generated backlash from neighbors who have complained about the noise of the machines inside. Texas has been a haven for cryptocurrency tech companies, primarily because of the state's space, deregulated power market and friendly business climate. Two weeks ago, the Public Utilities Commission adopted a rule requiring crypto and other virtual currency miners within the ERCOT grid to register their locations, ownership information and electricity demands, to further ensure that they could be watchful of this emerging source of energy consumption.
"Crypto mining operations currently consume around 2.3 percent of US electricity, and it requires roughly 155,000kWh to mine one Bitcoin," notes the site Data Centre Dynamics. This is the second off-grid power deal MARA has signed over the last few months. In October, it launched a 25MW micro data center operation across oil wellheads in Texas and North Dakota. The data center will be powered exclusively by excess natural gas from oilfield production that would have otherwise been flared. The operation will be distributed across wellheads in Texas and North Dakota, with operational status expected by January 2025.
Some context from Bloomberg: A few years ago Bitcoin miners took part in a global scramble for electricity to power their specialized computers... But the rise of AI, with its insatiable demand for electricity, dwarfed the needs of crypto and upended energy markets worldwide. Miners must now compete with much-larger tech firms for connections to electrical grids and power contracts. "Bitcoin miners are being forced to go look at marginal generation," said [MARA CEO Fred] Thiel. "The AI guys can afford to pay a much higher amount for energy than a Bitcoin miner"... MARA's plan to mine only when the wind is blowing makes economic sense because its mine will house last-generation computers that would otherwise have been retired, Thiel said.
"Thiel said he'd be interested to potentially buy more wind farms over time."

This year Marvell's stock rose 95%, giving it a $100 billion market capitalization, according to the Wall Street Journal.

"The latest gains have even put Marvell's market cap ahead of much-beleaguered Intel, which still generates 10 times as much annual revenue." Marvell's recent trajectory suggests that the revenue gap will continue to narrow. The explosive growth of its data center business has finally reached a point where it can more fully offset weakness in the company's more legacy segments, which sell chips used in goods such as telecommunications gear, cable TV boxes and autos. Data center sales nearly doubled year over year to $1.1 billion in the just-ended quarter, and Marvell's projection for the current period indicates the company will end its fiscal year in January with the data center unit encompassing about 72% of its total revenue, up from 40% in the previous year.

The next year is looking bright as well. Marvell's latest deal with Amazon is a five-year "multigenerational" agreement that has Marvell helping Amazon design its own artificial intelligence chips. Amazon, which runs the world's largest cloud computing service, has been expanding its internal chip efforts significantly, in part to reduce its reliance on Nvidia for crucial AI components. Amazon announced the next generation of its largest AI chip, called Trainium, at its annual developers conference this week. Analysts believe Trainium will play a role in Marvell's AI custom revenue more than doubling in the next fiscal year ending January of 2026.

That is expected to help propel Marvell's annual revenue to more than $8 billion in fiscal 2026, up 40% from what is expected for this year, according to consensus estimates from Visible Alpha. In addition, 20% growth is expected for the following year, when Marvell expects to be in production of custom AI chips for another unnamed big tech customer that analysts believe to be Microsoft. Analyst Mark Lipacis of Evercore ISI projects that the industry for custom AI chips will reach $30 billion to $50 billion in sales by 2030. In a note to clients last week, he said Marvell "has the potential to capture one-third of that market."
Marvell's CEO "has been among the few names floated as potential replacements for the recently ousted Pat Gelsinger at Intel's corner office," the article points out — which meant he had to reassure investors on an earnings call that he was staying at Marvell.

"The company is outstanding. The technology is best-in-class. I can't think of a better place to work than Marvell."

Not every tech "advent calendar" involves programming puzzles. Instead the geek tradition of programming-language advent calendars "seems to have started way back in 2000," according to one history, "when London-based programmer Mark Fowler launched a calendar highlighting a different Perl module each day."

So the tradition continues...

• Nearly a quarter of a century later, there's still a Perl Advent Calendar, celebrating tips and tricks like "a few special packages waiting under the tree that can give your web applications a little extra pep in their step."

• And of course there's a separate advent calendar for Raku programmers.

• Since 2009 web performance consultant (and former Yahoo and Facebook engineer) Stoyan Stefanov has been pulling together an annual Web Performance calendar with helpful blog posts.

• There's also a JVM Advent calendar with daily helpful hints for Java programmers.

• Another advent calendar promises daily posts about C#.

• The HTMHell site — which bills itself as "a collection of bad practices in HTML, copied from real websites" — is celebrating the season with the "HTMHell Advent Calendar," promising daily articles on security, accessibility, UX, and performance.

• There's even an advent calendar with tips for the reverse-engineering framework Radar.

• There's still a lovely web-design themed calendar at Designcember.com.

And meanwhile developers at the Svelte frontend framework are actually promising to release something new each day, "whether it's a new feature in Svelte or SvelteKit or an improvement to the website!"

But not every tech advent calendar is about programming...

• Adafruit's managing director is publishing a Retrocomputing Advent Calendar — daily looks at the ghosts of computers past.

• The Atlantic continues its 17-year tradition of a Space Telescope advent calendar, featuring daily images from both NASA's Hubble telescope and James Webb Space Telescope

• The gaming blog Rock Paper Shotgun has been counting down their favorite games of 2024...

The Washington Post details a vision of home security "pitched by Sauron, a Silicon Valley start-up boasting a waiting list of tech CEOs and venture capitalists." In the future, your home will feel as safe from intruders as a state-of-the-art military base. Cameras and sensors surveil the perimeter, scanning bystanders' faces for potential threats. Drones from a "deterrence pod" scare off trespassers by projecting a searchlight over any suspicious movements. A virtual view of the home is rendered in 3D and updated in real time, just like a Tesla's digital display. And private security agents monitor alerts from a central hub.... By incorporating technology developed for autonomous vehicles, robotics and border security, Sauron has built a supercharged burglar alarm [argued Sauron co-founder Kevin Hartz, a tech entrepreneur and former partner at Peter Thiel's venture firm Founders Fund]...

For many tech elites, security is both a national priority and a growing concern in their personal lives... After the presidential election last month, the start-up incubator Y Combinator put out a request for "public safety technology" companies, such as those that produce tools that facilitate a neighborhood watch or technology that uses computer vision to identify "suspicious activities or people in distress from video feeds...." Sauron has raised $18 million in funding from executives behind Flock Safety and Palantir, the data analytics firm, [and] defense tech investors such as 8VC, a venture firm started by Palantir co-founder Joe Lonsdale... Sauron is targeting homeowners at the high end of the real estate market, beginning with a private event at Abraham's home on Thursday, during Art Basel Miami Beach, the annual art exhibition that attracts collectors from around the world. The company plans to launch in San Francisco early next year, before expanding to Los Angeles and Miami...

Big Tech companies haven't deployed tools such as facial recognition as aggressively as Hartz would like. "If somebody comes onto my property, I feel like I should know who that is," Hartz said... In recent years massive investments have driven down the cost of drones, high-resolution cameras and lidar sensors, which use light detection to create 3D maps. Sauron uses lower-cost hardware and tools like facial recognition, combined with custom-built software adapted for residential use. For facial recognition, it will use a third-party service called Paravision... Sauron is still figuring out how to incorporate drones, but it is already imagining more aggressive countermeasures, Hartz said. "Is it a machine that could take out a bad actor with a bullet or something?"

"OpenAI is partnering with defense tech company Anduril," wrote the Verge this week, noting that OpenAI "used to describe its mission as saving the world." It was Anduril founder Palmer Luckey who advocated for a "warrior class" and autonomous weapons during a talk at Pepperdine University, saying society's need people "excited about enacting violence on others in pursuit of good aims." The Verge notes it's OpenAI's first partnership with a defense contractor "and a significant reversal of its earlier stance towards the military." OpenAI's terms of service once banned "military and warfare" use of its technology, but it softened its position on military use earlier this year, changing its terms of service in January to remove the proscription.
Hours after the announcement, some OpenAI employees "raised ethical concerns about the prospect of AI technology they helped develop being put to military use," reports the Washington Post. "On an internal company discussion forum, employees pushed back on the deal and asked for more transparency from leaders, messages viewed by The Washington Post show." OpenAI has said its work with Anduril will be limited to using AI to enhance systems the defense company sells the Pentagon to defend U.S. soldiers from drone attacks. Employees at the AI developer asked in internal messages how OpenAI could ensure Anduril systems aided by its technology wouldn't also be directed against human-piloted aircraft, or stop the U.S. military from deploying them in other ways. One OpenAI worker said the company appeared to be trying to downplay the clear implications of doing business with a weapons manufacturer, the messages showed. Another said that they were concerned the deal would hurt OpenAI's reputation, according to the messages...

OpenAI executives quickly acknowledged the concerns, messages seen by The Post show, while also writing that the company's work with Anduril is limited to defensive systems intended to save American lives. Other OpenAI employees in the forum said that they supported the deal and were thankful the company supported internal discussion on the topic. "We are proud to help keep safe the people who risk their lives to keep our families and our country safe," OpenAI CEO Sam Altman said in a statement...

[OpenAI] has invested heavily in safety testing, and said that the Anduril project was vetted by its policy team. OpenAI has held feedback sessions with employees on its national security work in the past few months, and plans to hold more, Liz Bourgeois, an OpenAI spokesperson said. In the internal discussions seen by The Post, the executives stated that it was important for OpenAI to provide the best technology available to militaries run by democratically-elected governments, and that authoritarian governments would not hold back from using AI for military uses. Some workers countered that the United States has sold weapons to authoritarian allies. By taking on military projects, OpenAI could help the U.S. government understand AI technology better and prepare to defend against its use by potential adversaries, executives also said.
"The debate inside OpenAI comes after the ChatGPT maker and other leading AI developers including Anthropic and Meta changed their policies to allow military use of their technology," the article points out. And it also notes another concern raised in OpenAI's internal discussion forum.

The comment said "that defensive use cases still represented militarization of AI, and noted that the fictional AI system Skynet, which turns on humanity in the Terminator movies, was also originally designed to defend against aerial attacks on North America.