Tops GPTs 
OpenAI and Poland’s NASK signed an agreement on October 23, 2026 to explore joint work on AI safety, security and technical research. The deal will pair OpenAI’s capabilities with NASK’s AI Safety Research Center, which serves as the Polish AISI.
Anthropic circulated a draft bill that would block AI model releases judged to pose a “foreseeable, imminent” risk of more than 50 deaths or a billion dollars in property damage. Independent evaluators registered with the U.S. Commerce Department would alert authorities and prompt a D.C. judge revie
An independent franchisee operating three co-branded Schlotzsky’s and Cinnabon restaurants in New Mexico used ChatGPT and ChatGPT Work to speed menu repricing from five days to 1–2 hours and to trim monthly cash reconciliation from three hours to minutes.
FairStress evaluates demographic fairness beyond final choices by testing how models respond when decisions are challenged. The benchmark includes 300 scenarios across 29 domains and a 50K FairStress-Core subset, and reports asymmetric sensitivity to persuasive prompts across many LLMs.
NVIDIA presents Cosmos3-H-Dreams, a generative simulator trained on clinical cholecystectomy recordings and designed for live kinematic control. The system runs interactively on a single NVIDIA RTX PRO 6000 GPU and includes teacher-to-student distillation for low-latency rollouts.
Vercel published usage data showing TypeSafe AI’s Jev was adopted by 35% of paid teams within two weeks of its limited release on September 15, nearly doubling the reach of Sonnet 5.5. Vercel’s report also highlights rising use of open source models and growing spend on frontier labs.
Higgsfield AI says GPT‑6 Astra enables creators to produce videos longer than the previous ~1-minute limit. The startup also highlights GPT‑Image‑2.5 for readable promotional imagery, Codex for faster prototyping, and plans for voice-to-video with GPT‑Live‑1.
MyLens AI says it delivers live, private 1-1 presentations at scale that draw on users’ raw sources, answer audience questions and build slides in real time. The company lists this as its 4th launch, with 2 reviews, 1.1K followers and free options.
Attackers accessed an older OpenCode system overnight, impacting 1.4% of OpenCode Console users and exposing API keys that led to unauthorised charges. OpenCode said it has notified affected users, is issuing refunds and is working with AI providers to block exposed keys.
Amazon Web Services released TorchNeuron on November 4 to let PyTorch apps run on its Trainium chip with a single configuration change, AWS says. The company also plans to contribute changes to PyTorch to reduce framework ties to specific chips.