chat: AI news

After her sudden departure from OpenAI last fall, ex-CTO Mira Murati vanished from public view to start something new. Now, she is ready to share some details about what she’s working on. Her new AI startup is called Thinking Machines Lab, and while the specifics of what it plans to release are still under wraps, the company says its goal is “to make AI systems more widely understood, customizable and generally capable.”  The startup also promises at least some level of public transparency by pledging to regularly publish technical research and code. In a press release shared with The Verge , the company suggests that it’s building products that help humans work with AI, rather than fully autonomous systems.  “We’re building a future where everyone has access to the knowledge and tools to make AI work for their unique needs and goals,” says the press release.

An architect of EU copyright law has said legislation is needed to protect writers, musicians and creatives left exposed by an “irresponsible” legal gap in the bloc’s Artificial Intelligence Act. Axel Voss, a German centre-right member of the European parliament, who played a key role in writing the EU’s 2019 copyright directive, said that law was not conceived to deal with generative AI models: systems that can generate text, images or music with a simple text prompt. “What I do not understand is that we are supporting big tech instead of protecting European creative ideas and content.” Voss said “a legal gap” had opened up after the conclusion of the EU’s AI Act, which meant copyright was not enforceable in this area. The intervention came as 15 cultural organisations wrote to the European Commission this week warning that draft rules to implement the AI Act were “taking several steps backwards” on copyright, while one writer spoke of a “devastating” loophole .

AI hardware startup Humane has given its users just ten (10!) days notice that their Pins will be disconnected .  In a note to its customers, the company said AI Pins will “continue to function normally” until 12PM PT on February 28.  On that date, users will lose access to essentially all of their device’s features, including but not limited to calling, messaging, AI queries and cloud access.  The FAQ does note that you'll still be able to check on your battery life, though. Humane is encouraging its users to download any stored data before February 28, as it plans on permanently deleting “all remaining customer data” at the same time as switching its servers off .

U.S. personal injury law firm Morgan & Morgan sent an urgent email this month to its more than 1,000 lawyers: Artificial intelligence can invent fake case law, and using made-up information in a court filing could get you fired .  A federal judge in Wyoming had just threatened to sanction two lawyers at the firm who included fictitious case citations in a lawsuit against Walmart.  One of the lawyers admitted in court filings last week that he used an AI program that "hallucinated" the cases and apologized for what he called an inadvertent mistake. AI's penchant for generating legal fiction in case filings has led courts around the country to question or discipline lawyers in at least seven cases over the last two years, and created a new high-tech headache for litigants and judges, Reuters found.  The Walmart case stands out because it involves a well-known law firm and a big corporate defendant. But examples like it have cropped up in all kinds of lawsuits since ch...

South Korea has accused Chinese AI startup DeepSeek of sharing user data with the owner of TikTok in China. "We confirmed DeepSeek communicating with ByteDance," the South Korean data protection regulator told Yonhap News Agency. The country [South Korea] had already removed DeepSeek from app stores over the weekend over data protection concerns. The Chinese app caused shockwaves in the AI world in January, wiping billions off global stock markets over claims its new model was trained at a much lower cost than US rivals such as ChatGPT. Since then, multiple countries have warned that user data may not be properly protected, and in February a US cybersecurity company alleged potential data sharing between DeepSeek and ByteDance.

If you happened to bring your smartphone to the moon, and it somehow survived both the trip and the brutal lunar conditions, it should work on the moon just like it does here on Earth .  To legally deploy the 4G network on the moon, Nokia received a waiver specifically for the IM-2 mission.  There are radio bands that have been internationally allocated to support lunar missions, and the LTE band is not among them.  “Using 4G frequencies on or around the moon is a violation of the ITU-R radio regulations,” NRAO’s spectrum manager Harvey Liszt explained in an email. “For permanent deployment we’ll have to pick a different frequency band,” [Thierry] Klein says. “We already have a list of candidate frequencies to consider.”  Even with the frequency shift, Klein says Nokia’s lunar network technology will remain compatible with terrestrial 4G or 5G standards.

 

In messages to newsroom staff, the company announced that it’s opening up AI training to the newsroom, and debuting a new internal AI tool called Echo to staff, Semafor has learned .  The Times also shared documents and videos laying out editorial do’s and don’t for using AI, and shared a suite of AI products that staff could now use to develop web products and editorial ideas. “Generative AI has the potential to bolster our journalistic capabilities even more,” the company’s editorial guidelines said. “Likewise, the Times will become more accessible to more people through features like digitally voice[d] articles, translations into other languages, and uses of generative AI we have yet to discover. We view the technology not as some magical solution but as a powerful tool that, like many technological advances before, may be used in service of our mission.”

"As technology progresses, we generally expect processing capabilities to scale up.  "Every year, we get more processor power, faster speeds, greater memory, and lower cost.  "However, we can also use improvements in software to get things running on what might otherwise be considered inadequate hardware.  "Taking this to the extreme, while large language models (LLMs) like GPT are running out of data to train on and having difficulty scaling up, [DaveBben] is experimenting with scaling down instead, running an LLM on the smallest computer that could reasonably run one. "Of course, some concessions have to be made to get an LLM running on underpowered hardware." 

"I am working to create a lesson for a Senior in High School around identifying specific rhetorical devices.   "I would like to use the Bible as the source text for the information. I use the Historical-Grammatical Hermeneutic. The Historical-Grammatical Hermeneutic is a method of biblical interpretation that seeks to understand the original meaning of the text as intended by the author.  "It does this by examining the syntax and structure of the passage in the original language, words or vocabulary, grammar, literary genre, historical background, and cultural context of the text.  "It also compares the text with other parts of Scripture to ensure consistency and harmony. This method respects the literary nature of the text and avoids imposing any hidden or allegorical meanings that are not supported by the text itself. "Analyze Exodus 1–40 for the use of rhetorical devices. Focus on the following devices: alliteration, anaphora, antithesis, hyperbole, metaphor...

"'I am surprised by the output every time I run it,’ says Elgammal. ‘An interesting question is: why is so much of the CAN’s art abstract? I think it is because the algorithm has grasped that art progresses in a certain trajectory. If it wants to make something novel, then it cannot go back and produce figurative works as existed before the 20th century. It has to move forward. The network has learned that it finds more solutions when it tends toward abstraction: that is where there is the space for novelty .' "This raises the intriguing notion that AI algorithms do not merely make pictures, they also tend to model the course of art history — as if art’s long progression from figuration to abstraction were part of a program that has been running in the collective unconscious for half a millennium, and the whole story of our visual culture were a mathematical inevitability. "No AI researchers are claiming that much just yet. They are still addressing the fundament...

"Many in our community seem unfazed or even excited about ‘AI’ and ‘agents’ and ‘codegen’ and all the rest of it.   "As far as I can tell, most of our industry is still on board with the project, even while protesting the changes in corporate politics, or occasionally complaining about the most obvious over-use.  "There are certainly a number of people raising alarms or expressing frustration, but we’re often dismissed as uninformed. "Based on every conference I’ve attended over the last year, I can absolutely say we’re a fringe minority. And it’s wearing me out.  "I don’t know how to participate in a community that so eagerly brushes aside the active and intentional/foundational harms of a technology. In return for what? Faster copypasta? Automation tools being rebranded as an agentic web? Assurance that we won’t be left behind?"

"According to Mark Gurman from Bloomberg , Apple is running into engineering problems and software bugs, which could delay the launch of the new Siri functionality. "One of the biggest features that Apple Intelligence boasted last year was a new Siri facelift , which would have allowed it to better compete against the likes of Google's Gemini and ChatGPT.  "It appears that Apple has delayed the highly anticipated Siri experience due to developmental challenges, as it will potentially not be a part of the iOS 18.4 release. "Apple is expected to release the iOS 18.4 beta next week to developers, which would come with a handful of Apple Intelligence upgrades.  "iOS 18.4 was considered to be a major update as it would include major Siri enhancements, but it looks like the company will take its sweet time before releasing it to the general public." 

Perplexity has become the latest AI company to release an in-depth research tool , with a new feature announced Friday. The goal is to provide more in-depth answers with real citations for more professional use cases, compared to what you’d get from a consumer chatbot.  In a blog post announcing Deep Research, Perplexity wrote that the feature “excels at a range of expert-level tasks —from finance and marketing to product research.” Google unveiled a similar feature for its Gemini AI platform in December. Then OpenAI launched its own research agent earlier this month.  All three companies even have given the feature the same name: Deep Research . Perplexity Deep Research is currently available on the web, and the company said it will soon be added to its Mac, iOS, and Android apps.  To use it, you just select “Deep Research” from a drop-down menu when you submit your query in Perplexity, which will then create a detailed report that can be exported as a PDF or shared as ...

The use of companion and therapy apps is rising in the Netherlands, but the systems do not always make it clear that users are chatting via AI.   “They need to make it clear that users are not talking to a real person,” AP director Aleid Wolfsen said. “Privacy legislation demands that apps are transparent about what happens to sensitive information they may share within a chat, and the AI regulations will soon include the requirement to make it clear that they are dealing with AI.”  The agency points out that many of the apps are provided by commercial companies that want to make a profit.   Some offer users subscriptions for extra services, such as outfits for their avatars . Some therapy apps also use paywalls which appear midway through a conversation. 

The Guardian has become the latest news publisher to sign a deal with ChatGPT owner OpenAI over content licensing. The deal will ensure The Guardian receives compensation for the use of its journalism on ChatGPT and gets properly credited on the platform.  Under the deal The Guardian will also be able to use OpenAI technology in-house. The publisher said: “Guardian Media Group today announced a strategic partnership with OpenAI, a leader in artificial intelligence and deployment, that will bring the Guardian ’s high quality journalism to ChatGPT’s global users. 

"For all its promised benefits, the artificial-intelligence boom is likely to prove costly to public health and even lead to hundreds of deaths a year in the U.S. alone, and those ill effects are likely to disproportionately hit poorer communities, according to a study by California esearchers (sic) that’s believed to be among the first of its kind. "AI affects public health via air pollution created as a consequence of operating the data centers used to train and run the technology, researchers at UC Riverside and the California Institute of Technology laid out in their report “The Unpaid Toll: Quantifying the Public Health Impact of AI,” published in December. "Demand for computing power for AI models is prompting the construction of increasing numbers of data centers.  "The manufacturing of chips and other hardware used in those facilities, the fossil-fuel power plants often used to provide electricity to them, and the typically diesel-powered backup generators ...

"In addition to refuting Yao’s conjecture, the new paper also contains what many consider an even more astonishing result .  "It pertains to a related, though slightly different, situation: In 1985, Yao looked not only at the worst-case times for queries, but also at the average time taken across all possible queries. He proved that hash tables with certain properties —including those that are labeled greedy , which means that new elements must be placed in the first available spot —could never achieve an average time better than log x. "Farach-Colton, Krapivin and Kuszmaul wanted to see if that same limit also applied to non-greedy hash tables. They showed that it did not by providing a counterexample, a non-greedy hash table with an average query time that’s much, much better than log x. In fact, it doesn’t depend on x at all.  "'You get a number,' Farach-Colton said, 'something that is just a constant and doesn’t depend on how full the hash table is....

With Elon Musk's so-called Department of Government Efficiency installing itself at the IRS amid a broader push to replace federal bureaucracy with machine-learning software, the tax agency's computing center in Martinsburg, West Virginia, will soon be home to a state-of-the-art Nvidia SuperPod AI computing cluster .  According to the previously unreported February 5 acquisition document, the setup will combine 31 separate Nvidia servers, each containing eight of the company's flagship Blackwell processors designed to train and operate artificial intelligence models that power tools like ChatGPT.  The hardware has not yet been purchased and installed, nor is a price listed, but SuperPod systems reportedly start at $7 million. The setup described in the contract materials notes that it will include a substantial memory upgrade from Nvidia.

OpenAI on Friday rejected a $97.4bn bid from a consortium led by billionaire Elon Musk for the ChatGPT maker, saying the startup is not for sale. The unsolicited approach is Musk’s latest attempt to block the startup he co-founded with CEO Sam Altman —but later left —from becoming a for-profit firm, as it looks to secure more capital and stay ahead in the AI race. “ OpenAI is not for sale , and the board has unanimously rejected Mr Musk’s latest attempt to disrupt his competition. Any potential reorganization of OpenAI will strengthen our nonprofit and its mission to ensure AGI benefits all of humanity,” OpenAI said on X, quoting its chair Bret Taylor, on behalf of its board.