Friday, Oct 30, 2026
AI;DR
2026-10-30 · 11 stories- anthropicCursor said a modified Claude Code extension in its marketplace exposed up to 5,000 users; it disabled and restored the marketplace, urging updates to 2.1.314. Anthropic disputed claims its official tool was compromised, while another report said Anthropic was hacked to deliver a remote-access trojan.+1
- googleBig Tech lost nearly $900B in market value as AI costs and weak capex expectations rattled investors. Alphabet lost about $500B after warning AI was hurting search ads, while Intel fell 23% and Amazon 7%; Microsoft also missed Q3 AI capex expectations.+2
- metaThe UK CMA opened a probe into whether Meta imposed exploitative terms on publishers and advertisers to use their content for AI and other services. Reuters reported the investigation on Friday.
- safetyEpoch AI found open-weight GLM 5.3 autonomously built exploits for six recent Firefox releases in 12–48 hours, costing about $1,000 each, without provider refusals. It failed against Chrome after tests costing about $200,000; Mozilla patched a Firefox flaw on Oct. 27.+1
- anthropicAnthropic and OpenAI said their models independently disproved the Exponential Time Hypothesis. Anthropic’s Claude found a subexponential 3-SAT algorithm on Oct. 16 using 14M output tokens without human input; the result also supports subexponential integer factorization, though encryption is not immediately broken.+1
- chipsChina’s CXMT reported Q3 revenue of about $16.7B, up 613.2% year on year, and net profit of about $8.17B. Reuters said AI-driven demand lifted memory-chip prices and pushed the company into profitability.
- ProductsSpotify aims to launch an AI product in Q1 2027 that lets subscribers create remixes and covers for about $5 a month, pending licensing deals. Sources told Anna Nicolaou of the Financial Times that the service could charge up to $5 monthly.
- codingCisco gave an engineer unlimited AI tokens to rewrite 98% of the company’s code. The item provides no further details about the project or its results.
- anthropicAnthropic said in its S-1 that it heavily depends on TSMC chips made in Taiwan and warned of risks from potential disruption amid tensions with China. The Wall Street Journal said the reliance makes Taiwan AI-chip production a major US foreign-policy interest.
- regulationWikipedia’s operator warned that a bad ruling in a landmark AI copyright case could threaten the open web. The group said both sides were taking the wrong approach.
- regulationAI-oriented PACs spent nearly $80M on the 2026 US elections, mostly before the primaries, according to FEC data reported by Puck. Pro-regulatory PACs outspent Leading the Future’s network, contrary to expectations.