Saturday, 10 October 2026
This dated roundup collects the most interesting AI and technology developments found for Saturday, 10 October 2026.
Research & Products
Chinese AI developers published safety results for 3.6% of reviewed releases, study finds
The review measures public disclosure, not whether testing happened privately. It also finds that China’s binding rules focus on applications rather than duties triggered by advanced model capabilities.
Read moreMicrosoft releases AI decision model, claiming 35-times faster responses than GPT-6 Sol
Decision-1 is available in Foundry for routing, classification and workflow checks. Its speed and accuracy pitch rests on Microsoft’s own structured-task benchmarks.
Read moreAWS introduces open-source tools to train, simulate and deploy intelligent machines
The Physical AI Toolchain connects cloud training to machines in the field. Customers can adopt the whole workflow or keep their existing tools and select individual components.
Read moreA new feature for my blog, built using my voice
I shipped a new feature for my blog today: the Newsletters page, which offers an index of all of the newsletters I've sent out, both my free weekly Substack and my monthly sponsors-only updates. I built the feature almost entirely using my voice, chatting away to my laptop while I cooked dinner. Codex voice mode I used the ChatGPT desktop app for this, in the Codex tab, using the voice conversation mode, running against a local development environment. Here's what that looks like: I started the session against my local simonwillisonblog checkout by typing: Start dev server and open in...
Read moreAHRQ launches $7.95 million AI prize initiative for medical evidence and data
The agency wants much faster evidence analysis, but speed is only one judging factor. Two coordinated competitions will accept submissions in 2027.
Read moreTags: announcement, acquisition
SoftBank seeks up to $100 billion from Gulf investors for AI acquisition fund, FT says
The proposed fund would apply AI and robotics to acquired businesses. Investor discussions remain preliminary, with no financing announced.
Read moreTags: ai_safety
OpenAI fires three safety researchers for "mishandling research information"
Three fired OpenAI safety researchers dispute allegations of mishandling sensitive information, warning in an open letter that their dismissals are creating a chilling effect on the company’s AI safety culture.
Read moreTags: ai_agents
Rein Security raises $25 million to block harmful AI-agent actions as they happen
The startup says it can stop individual malicious actions without shutting down an agent. Its financing announcement offers customer names and growth multiples, but no absolute revenue figures.
Read moreTags: reasoning
Researchers retrofit AI models to read bytes using under 1% of a typical training budget
The Nature study reuses existing models rather than rebuilding them. Its reported gains include character-level reasoning and some coding tasks, not a universal improvement across every benchmark.
Read more