OpenAI unveiled ChatGPT Images 2.0 via a noon PT livestream, its next-generation image model boasting 99% typography accuracy, 2K resolution output, and generation speeds twice its predecessor. The model handles previously challenging tasks — small text, UI elements, dense compositions, and subtle stylistic constraints — with unprecedented fidelity. Available immediately to all ChatGPT users, with an API (gpt-image-2) rolling out to developers.
A stealth lab founded by Richard Socher (ex-Salesforce chief scientist), Tim Rocktäschel (ex-Google DeepMind), and others with OpenAI/DeepMind pedigrees has raised over $500 million from GV and Nvidia, with the round reportedly oversubscribed to nearly $1 billion. The roughly 20-person team is building systems that recursively improve themselves across evaluation, data selection, training, and research direction without human-in-the-loop at each step.
Founded by Ohio State professor Yu Su and colleagues, NeoCognition raised $40M co-led by Cambium Capital and Walden Catalyst, with angels including Intel CEO Lip-Bu Tan and Databricks co-founder Ion Stoica. The lab is building AI agents that self-learn to become domain experts the way humans do, targeting enterprise SaaS companies as customers. The 15-person team is majority PhDs.
YouTube announced today it is opening its AI likeness detection technology to the entertainment industry, giving talent agencies like CAA, UTA, and WME the ability to scan for and request removal of AI-generated deepfakes of their clients. The system works like Content ID but for faces, scanning uploads for visual matches of enrolled participants. Audio detection is planned next.
Anthropic is now requiring select Claude users to submit government-issued photo IDs and live selfies via partner Persona Identities, triggered by detected potentially fraudulent or abusive behavior. Neither OpenAI nor Google require this for consumer use, making Anthropic the most restrictive of the Big Three. Users are pushing back hard on privacy grounds, though Anthropic says data won't be used for training.
India constituted the AI Governance and Economic Group, a high-level inter-ministerial body chaired by IT Minister Ashwini Vaishnaw, tasked with mapping which job profiles AI will displace first and developing region-specific transition plans. A companion Technology and Policy Expert Committee will feed technical guidance into AIGEG's decisions. It's the most structured governmental AI governance apparatus any major emerging economy has stood up to date.
