24 September 2026

Meta unveils Muse Charm, a keychain for talking to an AI agent

At the Connect conference, Mark Zuckerberg showed Muse Charm - a keychain that starts a conversation with the Muse agent when you touch its fingerprint reader. It is due to go on sale before the holidays; Meta did not give a price. The company also said it will offer AI glasses in more than 100 styles by the end of 2026, and its stock rose 4.5 percent on Thursday.

CNBC ↗

From recent days

Google will limit model access in free Gemini - only Flash-Lite from October 9

Starting October 9, Google will begin limiting model access in the Gemini app on personal accounts: users without a paid plan will keep only the Flash-Lite model and lose access to the Flash and Pro models, according to a Google help document reported by 9to5Google. Subscribers to the cheapest paid plan, AI Plus, will lose access to the Pro model, and Google is to notify them of the date by email. AI Pro subscribers, on the other hand, will get Deep Think mode, so far reserved for the AI Ultra plan.

9to5Google ↗ Google ↗

GPT-6 Astra drove a real Toyota Corolla through a cone course, Claude and Grok dropped out after a few meters

Three computer scientists from the San Francisco area, working as DrivingBench, connected GPT-6 Astra, GPT-5.6 Sol, Claude Fable 5.1 and Grok 4.6 to a rented Toyota Corolla and handed them the steering wheel, gas and brake, 404 Media reports. Only GPT-6 Astra completed the cone course in a parking lot; the other models drove a few meters. To get the models to agree to drive at all, the team had to call the real parking lot a "sandbox."

404 Media ↗ DrivingBench ↗

Three fired OpenAI employees urge the board to keep monitoring models' reasoning

Tomek Korbak, Mikita Balesni and Jasmine Wang, fired from OpenAI last week for violating rules on access to confidential information they allegedly passed to an AI safety organization, wrote a letter to the company's board and safety committees about chain-of-thought monitoring, Gizmodo reports, citing The Wall Street Journal. "As an industry, we don't yet know how to safely develop and deploy models we cannot monitor," the authors write, urging OpenAI and other companies not to pursue work that would further limit this ability. OpenAI said the employees were not fired for raising safety concerns, and in an internal memo wrote that it "strongly agrees" with the letter's recommendations.

Gizmodo ↗

All news →