The news desk
Recent developments in artificial intelligence, with the original sources one click away.
An Arizona appellate court vacated a manslaughter conviction sentence after an AI-generated video depicting the victim was shown during proceedings.
Sources · NYT Technology
Researchers propose novel RL techniques to enhance compositional reasoning and address training instabilities in large language models.
Sources · arxiv.org
OpenAI introduces visual advertising in ChatGPT alongside new measurement tools, attribution partnerships, and brand safety features for advertisers.
Sources · OpenAI
Three new papers present techniques for improving LLM training through reinforcement learning, addressing critic-free methods and inference-time scaling.
Sources · arxiv.org
President Trump signed an executive order renaming AI as 'super intelligence' and created a Super Intelligence Force to coordinate federal efforts.
Sources · techcrunch.com
New York Times reports legal experts find applying existing law to AI company liability complex and messy when systems go rogue.
Sources · NYT Technology
Millions have downloaded Meta's Muse AI agent, but using it requires sharing extensive personal data about friends and family.
Sources · WIRED AI
The seven-year-old Nvidia Shield TV streaming device has seen its price rise by $100, attributed to AI-driven demand for memory components.
Sources · WIRED AI
The New York Times conducted an experiment where five humans created music over a weekend to compare with AI-generated compositions.
Sources · NYT Technology
David Robinson, an OpenAI safety employee, has resigned while warning that the company's culture is broken.
Sources · TechCrunch AI
Anthropic released Opus 5.5 with Fable-level performance at lower prices, while OpenAI launched GPT-6 Sol and Luna models with reduced costs and errors.
Sources · Last Week in AI
Google shared its latest AI developments and features from September 2026 in a recent blog post.
Sources · Google AI Blog
Blackstone's Jas Khaira and Clay's Kareem Amin headline sessions on AI company financing and the emerging GTM engineer role.
Sources · techcrunch.com
AI startup Trillium Labs aims to publicly share research on risky topics like self-improvement and model behavior, contrary to typical frontier lab secrecy.
Sources · WIRED AI
President Trump signed an executive order rebranding AI as 'super intelligence' while tech CEOs signed a safety pledge, as consumer AI economics remain difficult.
Sources · techcrunch.com
Albertsons is using ChatGPT Enterprise and OpenAI API to accelerate team productivity and enhance customer shopping experiences.
Sources · OpenAI
Four new papers address collaboration between language models, asynchronous training methods, response masking techniques, and tool use in reasoning tasks.
Sources · arxiv.org
ChatGPT now allows users to virtually try on clothes using their photos and save items to a Favorites library.
Sources · TechCrunch AI
Four new papers tackle LLM efficiency through compute-aligned training, activation sparsity, environmental accountability, and adaptive agentic inference.
Sources · arxiv.org
A recently patched vulnerability in ChatGPT's Mac application could have allowed hackers to access sensitive user data, according to WIRED.
Sources · WIRED AI
