[01]
OpenAI launches GPT-6 Astra, its first model to cross a critical cybersecurity threshold
OpenAI's GPT-6 Astra became the first model to cross its "Critical" cybersecurity threshold, finding two zero-day exploits.
// daily-digest · fri 4 sep 2026 · morning refresh · last 72h focus
Ten items on the GPT-6 Astra launch and its cybersecurity implications, Nvidia's Hugging Face deal, AI funding and IPO news, and a staff-engineering essay, plus a wildcard on fruit fly brain connectomics.
— refreshed for you, in 10 items.
OpenAI's GPT-6 Astra became the first model to cross its "Critical" cybersecurity threshold, finding two zero-day exploits.
Ben Thompson interviews OpenAI president Greg Brockman on the company's history, the Astra launch, and alignment.
Simon Willison reviews GPT-6 Astra's benchmarks, pricing, and security strengths against Anthropic's Claude Fable 5.1.
Nvidia agreed to acquire AI platform Hugging Face for $12.9 billion while pledging to keep it framework-agnostic.
Anthropic plans to file its IPO prospectus after Labor Day, reportedly targeting a valuation above $2 trillion.
xAI rolled out Grok Bot to enterprise customers with admin controls and a two-week free trial.
AI security startup HiddenLayer raised a $100 million Series B to secure enterprise agentic and coding AI.
The US is pushing lighter AI oversight while the EU enforces its AI Act through new compliance probes.
Sean Goedecke argues "radical responsibility" as a leadership tactic treats employees as instruments, not moral agents.
Google Research and HHMI Janelia mapped a fruit fly's full brain: 166,000 neurons and 125 million synapses.