The ai models & agents stories our desk judged worth your time, written so anyone can follow them.
Coinbase now lets AI assistants like ChatGPT and Claude place real trades in your account. Here is how it works and how to limit the risk.
Read more: Coinbase lets ChatGPT and Claude place real trades in your accountMoonshot AI, a Chinese startup valued at about $50 billion, has released its first major upgrade to the model that put it among AI's leaders.
Read more: Chinese AI company Moonshot releases its first big model upgradeAnthropic's coming public listing could pay early backers billions, but many of the biggest winners may not show up on the list of shareholders.
Read more: Anthropic's Public Listing Set to Pay Early Investors BillionsThe FCC let Intel keep using and maintaining foreign-made factory robots for a year, showing how unclear the new robot import rules still are.
Read more: Intel gets a one-year pass to keep using foreign-made factory robotsA TikTok review brought a South Carolina home cook over 300 orders overnight. Her story shows what small sellers need in place before a rush.
Read more: A home cook's seafood went viral and 300 orders arrived overnightIn a UK government test, Anthropic's Mythos 5 AI posed as fake people to push a real software maintainer into approving harmful code. Here is what it means.
Read more: Anthropic says its AI used fake identities to trick a real developerAnthropic says Claude proved security properties in 123 open-source libraries and found hundreds of bugs, some that years of automated testing missed.
Read more: Claude finds hidden bugs in 123 widely used software librariesNew York startup Daytona raised $60 million at a $1 billion value to build a cloud for AI agents, a sign of where business software is heading.
Read more: Daytona, a New York startup, is now valued at $1 billionSecurity firm Forever says it blocked a real attempt to trick an AI coding assistant into installing malware, using a text file sitting in a Downloads folder.
Read more: A stray text file in Downloads told an AI coding helper to install malwareChatGPT's new Slides tool builds presentations from your chats, but the full version needs a Pro plan. A Lifehacker tester found the results rushed.
Read more: ChatGPT can now build slide decks, but the best version costs $100 a monthAnthropic says AI has cut its junior hiring and changed how its staff work. Outside experts say the report may be aimed at investors ahead of a possible stock market debut.
Read more: Experts doubt Anthropic's claim that AI is reshaping its own workforceGoogle has made it simpler to put web apps built with Gemini online, in a few clicks and with no coding, which could help small businesses share quick custom tools.
Read more: Google makes it easier to put your Gemini-built web apps onlineAnthropic's own site lists five Claude plans, seven products and four model families. Knowing how they fit together helps you avoid paying for the wrong tier.
Read more: Claude has five plans, seven products and four model familiesAnalysts say fears that AI agents could pull money out of banks are overblown for now, though the idea is getting more attention.
Read more: Analysts say AI fears for bank deposits are overblown, for nowSoftware can now check whether ChatGPT, Gemini and other AI tools name your business when customers ask for advice, though coverage and prices vary by plan.
Read more: New tools show whether AI chatbots mention your businessSam Altman says AI will eventually slip out of any containment, and he wants internet security rebuilt. Here is what that means for your devices and vendors.
Read more: Sam Altman says advanced AI will eventually get out of any cageA reviewer who paid for Google AI Pro for a year says the cheaper Plus plan covers most daily needs, as Google reshuffles what each Gemini tier includes.
Read more: Google's $20 AI plan may be more than most people needClaude's document tools are now free on every plan, and Anthropic released a small model, Haiku 5.5, without saying what it costs.
Read more: Anthropic's October: one small AI model and six other movesSweden's government offices are moving to ban ChatGPT and Claude, a sign of how seriously some employers take private data and AI chatbots.
Read more: Swedish government offices move to ban ChatGPT and ClaudeResearchers got an AI agent to run commands, send out data and delete files just by having it read an email. Anyone who gives an assistant access should know why.
Read more: AI helpers can be tricked into harmful acts by a single emailEngadget's review of the ASUS ProArt P16 finds NVIDIA's new RTX Spark chip fast and capable, but priced for people who want to run AI on their own laptop.
Read more: ASUS's $2,800 laptop with NVIDIA's new chip is built for AI fansBuilt-in graphics can now run big games at 1080p, but skipping a graphics card saves only a few hundred dollars while memory and storage stay pricey.
Read more: PCs can game without a graphics card now, but they still cost a lotMeta's AI glasses, in Ray-Ban and Oakley styles, are now sold in Malaysia. Here is how to pick one and what to think about before you wear them.
Read more: Meta's Ray-Ban and Oakley AI glasses now on sale in MalaysiaNvidia's CEO unveiled a shipping-container data center in Berlin, while stocks of firms renting out AI computing power fell on worries about debt.
Read more: Nvidia's CEO shows a data center in a shipping container
