beat · 194 stories
A 10,000-query UK study tested ChatGPT, Claude, Copilot, Grok and Gemini on tax, pension and loan questions. The data also tells you where the chatbot answer is safe to trust.
Researchers built a benchmark from real business dashboards. Frontier AI still scored below 50%, and the tool that closed some of the gap shows how far 'AI replaces the analyst' really is.
Chinese AI developers lease Nvidia capacity in foreign data centers through non-Chinese proxy entities, a workaround the House and the Commerce Department's Bureau of Industry and Security (BIS) are now trying to cut off.
BYD's Denza sub-brand is rolling a cockpit AI assistant called Didi Xia to four models, with open agent protocols and 10+ partner integrations; NIO, Xpeng, and Huawei already ship similar systems.
Spear phishing pulls names from LinkedIn and company sites to look personal; a BYU study found AI versions now match or beat human ones, and people told them apart only half the time.
The Japanese tech investor swaps a short-term bridge loan for multi-year debt ahead of an Oct. 1 close on its third OpenAI follow-on payment.
Treasury Secretary Scott Bessent proposed a US-China AI incident-notification channel at Sunday's New York talks; Vice-Premier He Lifeng left without commenting before the Thursday summit.
AX offers sandboxing, repos, network policy, and model keys as one declarative runtime. The lock-in question is what every agent builder should walk away with.
Google's Loviisa nuclear deal is the first time a US hyperscaler crossed the Atlantic to underwrite three new Finnish data centers with up to half the output of a single plant.
Anthony Albanese will press Donald Trump on AI safeguards in New York this week, but the operative Australian move is a domestic standards bill and a copyright regime flagged as a hurdle for US labs.
The cost of running AI is on the same curve as electricity and broadband. The real question is who captures the value when it does.
AI that can read files, click buttons, and act on your behalf is now a browser feature, and one extension can take it over. Chrome and Edge have patched; three others have not.