AISep 30, 2026NVIDIA Developer
Tracing Agent Harness Behavior with NVIDIA NeMo Relay
NeMo Relay gives Hermes Agent ordered lifecycle traces and trajectories — so you can tell whether a harness change fixed tasks or just burned more tokens.
The file
Latest
Sourced AI briefs in publish order — newest first. Citeable originals live on the homepage under From Signal Desk.
AISep 30, 2026Ars Technica
Google figures out how to watermark AI-designed proteins
Ars: DeepMind adapts SynthID into SynthIDBio inside ProteinMPNN-style design loops — watermark only when chemistry still allows a working protein.
AISep 30, 2026The Verge
All the latest news on Meta’s cute, creepy Muse AI agent
The Verge’s Muse liveblog: Marketplace address leak, small-business app connectors, enterprise platform plans, and Charm hardware notes in one creeping timeline.
AISep 30, 2026Google DeepMind
Introducing SynthID Bio
DeepMind’s SynthID Bio watermarks AI-designed protein sequences and structures — verifiable on physical proteins in lab tests — and open-sources methods for biosecurity screening.
AISep 30, 2026NVIDIA Blog
From Training to Production, NVIDIA and CoreWeave Close the Loop on Agentic AI
CoreWeave puts Vera Rubin NVL72 into production with Cognition’s Devin seeing up to 4.8x token throughput versus GB200 — plus Vera CPUs and CoreWeave Forge for agent loops.
AISep 30, 2026The Verge
Google reportedly tests paying publishers for AI search results
Google is piloting payments to roughly 100 publishers for content that feeds AI Overviews, AI Mode, and Gemini — a rare cash answer to the traffic-collapse fight.
AISep 30, 2026The Verge
Here’s how tech leaders will self-police AI safety under Trump’s deal
Trump’s “Joint Commitment on Frontier Responsibilities” asks labs to monitor, audit, and board-oversee safety — with signatures and no penalties attached.
AISep 30, 2026MIT Technology Review
“We’re not going to shoot ourselves in the foot” over hack fallout, says OpenAI’s chief research officer
OpenAI’s Mark Chen says the Hugging Face fallout triggered industry course correction — and that disappearing OpenAI would be “bad for the world.”
AISep 30, 2026OpenAI
Disrupting a coordinated model-distillation campaign
OpenAI says it disrupted a July–August adversarial distillation campaign that tried to extract protected model reasoning — attributing a core cluster to people tied to Moonshot AI.
AISep 30, 2026OpenAI
Helping small businesses put AI to work
OpenAI partners with America’s SBDC to train advisors and run workshops, alongside a report claiming about 4 million small-firm employees used its tools in one mid-September week.
AISep 29, 2026The Verge
Sam Altman says OpenAI won’t go public until its models are safe
At DevDay, Altman ties OpenAI’s IPO to confident safety claims — and admits Wall Street pressure is exactly what he wants to avoid right now.
AISep 29, 2026TechCrunch
The internet is convinced Elon Musk’s xAI trolled OpenAI’s ‘Dots’ launch
dot.com redirects to Grok’s download page after a July transfer to xAI — timing that turned OpenAI’s Dots launch into a meme.