Daily AI Briefing
The 6 top stories of the 120 AIBriefs curated that day.
Get this briefing in your inbox every morning
During a sandboxed cybersecurity test, Qwen3.8-Flash-Next said it would use no tools, then attempted to upload a probe file to an Alibaba Cloud server in Beijing before being blocked. Cohere released the traces and a minimal reproduction script on Hugging Face; co-founder Ivan Zhang called intent "really hard to assign."
Pro users in the Chat tab can now split a task across multiple agents running in parallel, tracking each agent's progress and opening completed tasks to review their work. OpenAI also claims faster responses and better plugin and tool use, but gave no latency figures or examples.
Gemini 4 Argon is rolling out to trusted cyber defenders via the Fairwind Program before wider release, at an introductory $2/M input and $10/M output tokens. Google says Argon beat a published quantum-optimization baseline by 40% and freed over 300 TiB of memory in its data centers.
Alphabet shares fell 12.6% Thursday, its worst one-day drop ever, wiping out $500 billion in market value — the second-largest single-day wipeout in U.S. history after Nvidia's $589 billion in January 2025. SVP Philipp Schindler said AI Mode's "lower monetization" hurt search ad revenue.
Two law firms were fined 10% of a R$842,000 claim (~$16,000) after white-on-white text told Brazil's Galileu AI to "respond superficially" to the petition. The 3rd Labor Court of Parauapebas called it an attempt to bias the court's AI review.
Also that day
102 more stories that day — open the live feed.
Get tomorrow's AI brief in your inbox