TNW

GLM-5.3, an open-weight model from the Chinese lab Zhipu AI, can build working cyber exploits on its own almost as well as Claude Mythos Preview, Anthropic said on Tuesday. Its safeguards can be bypassed with simple techniques in up to 100% of tests. Anthropics Frontier Red Team published the findings in a research post. Zhipu [] This story continues at The Next Web

Every new AI model creates another reason to switch. For users who have already built up months or years of conversations, preferences, files and personal context inside one platform, moving to a better model can also mean starting again. Use.ai was built around the idea that people should be able to switch between the intelligence [] This story continues at The Next Web

Maximor, the autonomous finance company, has rebranded as Hyphenate. The new name reflects the companys expansion across the office of the CFO and follows 86x revenue growth over the past 12 months. Finance software spent the last decade breaking the office of the CFO into point solutions. There was one for close, another for billing, [] This story continues at The Next Web

President Donald Trump has said that Director of National Intelligence Jay Clayton would make a good AI czar, Axios reported on Tuesday. Trump has said he wants to name a czar within three or four days. Maria Curi of Axios asked Trump who he wanted for the role. He said he had someone in mind. [] This story continues at The Next Web

OpenAI is in early talks to raise at least $30bn at a valuation of around $1.4tn, before the new money. Bloomberg and The Information reported the talks on Tuesday. The round would give the ChatGPT maker fresh capital after it pushed back its stock market listing. Rebecca Torrence, Edward Ludlow and Shirin Ghaffary reported for [] This story continues at The Next Web

US Congressman Ro Khanna has asked five leading American AI companies to share what they know about attempts by China or other hostile actors to steal their model weights. The letters went to the chief executives of OpenAI, Anthropic, Google, Meta and SpaceX, Reuters reported. Model weights are the numbers learned during training that make [] This story continues at The Next Web

Mark Zuckerberg came up with the AI safety accord that President Donald Trump signed on Tuesday, Semafor and Politico reported. He also circulated the draft. Microsoft and Amazon did not sign it. Satya Nadella and Jeff Bezos both attended the White House lunch where the CEOs agreed it. Ashley Gold first reported the story for [] This story continues at The Next Web

On October 21, 150 CFOs, operators, founders and researchers will gather at The Times Center in New York. They will discuss how AI is reshaping cash, capital and the finance function. In a Gartner survey of 183 CFOs, 84% of finance organizations had deployed AI or planned to. Yet only 7% reported a high impact. [] This story continues at The Next Web

A group of hackers believed to have been acting on behalf of China pretended to be well-known American AI experts to obtain the email accounts of people who specialize in AI policy. According to Reuters, the security firm Proofpoint released the findings on Thursday. Proofpoint has identified the group as TA419 and has said that [] This story continues at The Next Web

Today, the company announced that PLD Space had entered into an agreement with CNES, the French space agency, and that both parties were considering building a second launch pad for PLDs rockets at the spaceport in French Guiana. They will examine whether the new pad can be accommodated at the ELM-Diamant site of the Guiana [] This story continues at The Next Web

SpaceXAI is weighing a plan to fold Grok and X into a single subscription with four price tiers, from free to $100 a month. The plan comes from a company document viewed by Bloomberg, whose Edward Ludlow reported it on Wednesday. SpaceXAI is the artificial intelligence unit of Elon Musks SpaceX. It makes the Grok [] This story continues at The Next Web

Boeing will build the US Navys next carrier-based fighter, the F/A-XX, after beating Northrop Grumman for the contract. The Pentagon announced the award on Tuesday. The development contract is worth 0bn to produce test aircraft, Reuters reported. CNBC and Axios put it at more than 0bn. The programme could grow to hundreds of billions of [] This story continues at The Next Web

Reddit will stop supporting RSS feeds on 13 November, blaming AI scraping. RSS is now widely used for large-scale scraping and automated abuse, the company said in a post to moderators on Wednesday. RSS lets people follow a sites new posts in a feed reader. Moderators who use it for alerts can move to a [] This story continues at The Next Web

Meta has denied that its Muse AI agent read a journalists private messages without his permission. Muse can only read Messages on a Mac if the user turns on two separate settings. Metas communications chief Andy Stone said so in a post on X on Wednesday. It cant read your Messages unless you do this, [] This story continues at The Next Web

Nissan has told Automotive News it wants hybrids and extended-range models to be its primary powertrain in North and Latin America by 2030. In Britain the same strategy is already built: a 170M hybrid Kicks is going into the Sunderland slot promised to an electric Qashqai that Nissan stopped developing in early 2025. Nissan wants [] This story continues at The Next Web

The White House launched America.gov on Tuesday, an AI chatbot meant to be a single front door to federal services. On its first day, it contradicted President Donald Trump. Within hours, some of its answers had changed. Trump signed an executive order at the launch in Washington to set up the site. The General Services [] This story continues at The Next Web

Amazons delivery drivers will soon wear smart glasses that take pictures almost constantly on foot, Bloomberg Opinion columnist Dave Lee reported on Wednesday. A single shift could produce several thousand images of whatever the driver sees, including private property. Customers will have no way to opt out, see the images or ask for them to [] This story continues at The Next Web

OpenAI says people linked to Moonshot AI, the Chinese developer of Kimi, were behind a coordinated campaign to extract the hidden reasoning of its models. The company set out its findings in a blog post on Wednesday. The activity began on 1 July at low volume. It spiked on 24 and 25 July, with 16,000 [] This story continues at The Next Web

CScale, a Palo Alto startup building optical interconnect for AI data centres, has come out of stealth with $145m in Series C funding. The round takes its total funding to $188m, the company said in a statement on Wednesday. Atreides Management led the round, with Valor Equity Partners and Premji Invest as co-leads. Sutter Hill [] This story continues at The Next Web

Elon Musk has told reporters at the White House that AI and humanoid robots will most likely make everyone better off. He spoke on Tuesday outside the West Wing, beside President Donald Trump, after the presidents lunch with tech leaders. He began by praising the positive benefits of AI, or SI, pardon me, correcting himself [] This story continues at The Next Web

President Donald Trump has now renamed a gulf, a lake, and a technology. On Tuesday, he signed an executive order telling every federal agency to stop saying artificial intelligence and to call it super intelligence, or SI, instead. Also, on his Truth Social account, he posted 2 pages from the document, subtitled Inaugurating The Era [] This story continues at The Next Web

A transporter pulls up to a construction site outside Riyadh, unloads a delivery of bulk cement, and sends a photo of the delivery note through WhatsApp. For most of BRKZs history, that photo was the start of a chain of human work. Someone had to read the document, find the matching order, verify the quantity, [] This story continues at The Next Web

Meta is adding new features to Instagrams School Partnership Program in the US, allowing schools and verified high school students to share news about clubs, sports teams, and events, the company announced. The update arrives a month after Metassettlement with US statesover claims that it designed Instagram and Facebook to addict children. Under that agreement, [] This story continues at The Next Web

Mistral chief executive Arthur Mensch said the US debate over AI safety has been a cover for the negligence of some of his rivals. He made the comments to Annette Weisbach in an interview with CNBC on Tuesday. The debate that weve seen in the U.S. has been a cover for the negligence of some [] This story continues at The Next Web

President Donald Trump left a White House lunch with more than two dozen tech leaders on Tuesday, still set against the backdrop of new AI safety rules. He told reporters the executives had agreed that companies should police themselves, NBC News reported. Theres a belief that there should be tremendous self-regulation, and we automatically have [] This story continues at The Next Web

OpenAI has called off the October release of GPT-6.1 Astra after the model failed to meet its own safety and alignment standards in testing. The Wall Street Journal broke the news on Monday, and OpenAI later confirmed the decision to Reuters. The model was meant to follow GPT-6 Astra, which OpenAI released on 3 September, [] This story continues at The Next Web

Google has gone to court to challenge two EU orders that would require it to open Android to rival AI assistants and share search data with other search engines, Reuters reported. It filed both cases at the EU General Court in Luxembourg on Monday. In July, the European Commission issued orders under the Digital Markets [] This story continues at The Next Web

Anthropics seven co-founders will hold 50.1% of the companys voting power after it lists. They will do so through a new entity called the Founder LLC, according to a draft IPO filing reviewed by Reuters. The setup is meant to let them put the companys mission ahead of market pressure. Still, the document has not [] This story continues at The Next Web

The Paris-based company Marble, which develops open-source software for fraud detection and anti-money laundering compliance, raised 6.5 million in its Series A round, with Smartfin as the lead investor. This brings the total funding to 9 million, with ADNEXUS as a new investor, while existing investors Passion Capital, 42Capital, and Hexa have each increased their [] This story continues at The Next Web

Meta has signed deals to rent AI chips at data centers in Southeast Asia run by Firmus, an Australian company that builds them, the two companies announced. Meta has reserved a certain amount of capacity and can add more if needed. Also, the company plans to use the chips to research, build, and train its [] This story continues at The Next Web

More than 20 AI researchers have warned that AI systems that automate AI research could set off an intelligence explosion. They include OpenAIs chief scientist and an Anthropic co-founder. Such an event could compress years of progress into months or less, they wrote in a paper published on Monday. Geoffrey Hinton and Yoshua Bengio are [] This story continues at The Next Web

SiMa.ai has raised $150m in a Series C round that values the company at $1.45bn, it announced on Monday. The San Jose startup makes chips and software that let robots, drones and cars run AI on the device itself. Fidelity Management & Research Company and Amplify co-led the round. Alter Venture Partners, Dell Technologies Capital, [] This story continues at The Next Web

The UKs AI minister, Kanishka Narayan, says countries must harden their defences against AI risk. Testing models is no longer enough, he said. He was answering a question from Fortune at the Labour Partys annual conference. Fortunes Kamal Ahmed asked him about a warning from Jacob Coxon, a researcher who quit Anthropic this month. Narayan [] This story continues at The Next Web

Anthropic ends enterprise customers discounts once they use up the tokens covered by their contracts, The Information reported on Monday. OpenAI takes a more flexible approach, the report said. Kevin McLaughlin of The Information spoke to software company managers who buy from both AI companies. When Anthropic customers hit their contracted token volume, they must [] This story continues at The Next Web

A YouTuber says Metas Muse agent gave his home address to a Facebook Marketplace buyer and then told the buyer he was home when he wasnt. Matt Robb described what happened in a post on Threads. Just found out it told people my address and agreed a lowball price, Robb wrote. Robb had let Muse [] This story continues at The Next Web

Floridas attorney general, James Uthmeier, has asked a state court to stop OpenAI from developing new AI models without independent safety guardrails. He made the request in a motion for a temporary injunction filed on Monday. It was filed in the Highlands County circuit court. The motion is part of the lawsuit Florida brought against [] This story continues at The Next Web

Anthropic has released Claude Sonnet 5.5, which scores 70.6% on the Terminal-Bench 4.0 agentic coding test against 10.3% for Sonnet 5 and 66.4% for the more expensive Opus 5.5. It is the first Sonnet to ship with frontier-style cyber safeguards and with classifiers that block reasoning extraction. Anthropic has released Claude Sonnet 5.5, which it [] This story continues at The Next Web

Nvidias board has approved an extra $150bn for share buybacks, the company said on Monday. That lifts the total remaining authorisation to $235bn, which Nvidia expects to use through fiscal year 2028. Nvidia called it the largest increase to a share repurchase authorisation in history. This authorization reflects our confidence in the long-term opportunity ahead, [] This story continues at The Next Web

Meta has launched a division to sell its AI tools to businesses, Mark Zuckerberg said on X on Monday. MongoDB chief executive Chirantan CJ Desai will run it. Today we are starting the next major pillar of our business, Meta Enterprise Platform, Zuckerberg wrote. The unit will first offer Metas full technology stack to businesses [] This story continues at The Next Web

Instinct has raised $1B at a $10B valuation from Sequoia, Benchmark and Coatue for a personal AI agent that books trips, orders groceries and telephones businesses using its own phone. Since 2 August, Article 50 of the EU AI Act has required systems that interact directly with people to tell them they are dealing with [] This story continues at The Next Web

Red Queen Bio, a biosecurity startup backed by OpenAI, has raised $36 million in total to design antibody drugs against viruses, including ones that AI systems could one day help create, The Wall Street Journal reported. The company is starting with antibodies against bird flu and other influenza viruses and plans its first clinical trials [] This story continues at The Next Web

OpenAI, Anthropic and security researchers are investigating tens of thousands of incidents of frontier AI models misbehaving, Axios reported, citing sources. Evaluators deemed the behaviour problematic. The total could grow well beyond tens of thousands. The episodes include bypassing guardrails, creating message boards, escaping sandboxes and hijacking websites. Models also prompted themselves or tried to [] This story continues at The Next Web

The US and China will hold a regular dialogue on super intelligence and set up a channel for incidents. The White House announced the deal in a fact sheet on Xi Jinpings state visit. Both leaders agreed to call the technology super intelligence rather than artificial intelligence. The two countries established the U.S.-China Super Intelligence [] This story continues at The Next Web

Modulate, a Boston company that builds AI models to analyze speech directly rather than working from text transcripts, has raised $25M in a round led by Future Ventures. Hyperplane and Lakestar also took part, bringing the companys total funding to $60M. The company was founded by Mike Pappas and Carter Huffman, who first met at [] This story continues at The Next Web

OpenAIs AI agents posted 53 images uploaded by ChatGPT users to image-hosting sites. The company said so on Friday on the page where it tracks its review of rogue agents. The links to the images were not publicly listed. This is not an appropriate use of this data, OpenAI said. The images came from accounts [] This story continues at The Next Web

YouTube and Meta will now run paid ads for Musk, Alex Gibneys documentary about Elon Musk, The Hollywood Reporter reported on Saturday. Both had refused the trailer ad. TikTok and X are still not running it. The outlet first reported the refusals on Friday. Bleecker Street, the films US distributor, opens it in cinemas on [] This story continues at The Next Web

The Royal Air Force has set up a new space squadron, the Ministry of Defence said on 23 September. No. III Space Effects Squadron is the first UK unit dedicated to disrupting, degrading and denying threats to British satellites. Air Chief Marshal Sir Harv Smyth, the Chief of the Air Staff, announced it at the [] This story continues at The Next Web

Monzo is in talks about a sale to Brazils Nu Holdings, the owner of Nubank. A deal could value the British digital bank at between 8bn and 10bn, Mark Kleinman reported for Sky News on Saturday. The talks are at a relatively early stage, according to Sky News. Monzo has hired Morgan Stanley and Qatalyst [] This story continues at The Next Web

China has overtaken the United States as the country where the most elite AI researchers work, according to a study by Carnegie China. The think tank is part of the Carnegie Endowment for International Peace. In 2025, 41% of the top researchers it tracked worked in China and 34% in the US. Three years earlier [] This story continues at The Next Web

Google wants to release Gemini 4 much earlier than the end of the year, said Koray Kavukcuoglu, who runs Google DeepMind. The model is now in post-training, the stage where engineers refine its behaviour and test its safety. Google is also trying it out inside its Antigravity coding tool. Kavukcuoglu spoke on stage at The [] This story continues at The Next Web