// AI Briefing
May 16, 2026
OpenAI is going full Accenture with a $4B consulting army, ChatGPT now wants to see your bank statements, and Cerebras just proved that betting against Nvidia's dominance can still pay off spectacularly. Meanwhile, David Silver's RL dream lab gets Nvidia's blessing, and Europe quietly rewrites its AI rulebook to be less painful for startups.
OpenAI rolled out personal finance tools for ChatGPT Pro subscribers in the US, partnering with Plaid to connect over 12,000 financial institutions. Users get a portfolio dashboard, spending analysis, and financial planning all inside ChatGPT. The move follows OpenAI's acqui-hire of personal finance startup Hiro in April and signals a clear push to make ChatGPT a daily utility rather than just a chatbot.
Cerebras Systems soared on its first day of trading, closing at $311 after pricing at $185 and raising $5.55 billion in the largest US tech IPO since Uber in 2019. The company's wafer-scale chips offer an alternative architecture to Nvidia's GPUs, and investor appetite suggests the market believes there's room for more than one winner in AI silicon. With underwriter options, total proceeds could reach $6.38 billion.
Nvidia announced an engineering collaboration with David Silver's London-based Ineffable Intelligence to co-design infrastructure for large-scale reinforcement learning. The work starts on Nvidia Grace Blackwell and will be among the first to explore the upcoming Vera Rubin platform. This signals Nvidia's strategic bet that the next wave of AI progress comes from RL self-play systems that generate their own training data rather than relying on human-produced datasets.
OpenAI launched a separate entity called the OpenAI Deployment Company, backed by $4 billion from TPG, Advent, Bain Capital, and Brookfield, with 19 total partners including Capgemini. The venture will embed Forward Deployed Engineers into enterprises to rebuild workflows around AI. By acquiring Tomoro (150 AI engineers), OpenAI is copying Palantir's playbook of selling implementation alongside the product, signaling that raw model access alone isn't enough to capture enterprise value.
After talks collapsed in early May, EU negotiators rebounded with a provisional agreement on the AI Act Omnibus that pushes back high-risk compliance deadlines to late 2027/2028, extends SME-style regulatory breaks to small mid-cap companies, and adds a new prohibition on AI-generated non-consensual intimate imagery effective December 2026. The deal resolves the overlap between the AI Act and existing product safety laws, removing a major source of industry confusion.
Test Your Understanding
Quiz
1 / 8
Which financial connection service did OpenAI partner with for ChatGPT's personal finance feature?
Summary
A federal judge just told the Pentagon it cannot punish an AI lab for refusing to hand over its models, which is the first time a court has drawn a hard line around what safety policies cost you in Washington. Elsewhere the theme is AI learning by watching rather than being told: a robot foundation model that copies a ten-minute task from one video, and Claude grinding through algebra that stumped a lab for eighteen months. Plus Meta patching the dumbest privacy hole in its glasses, and Pew's number on how many Americans now ask a chatbot about their symptoms.
Top Stories
A Judge Just Ruled the Pentagon Illegally Punished Anthropic for Saying No to Mass Surveillance
Skild's Robot Model Watches One Video of You Making Pancakes and Then Makes Pancakes
Claude Did Eighteen Months of Algebra in Five Weeks, and Got Some of It Wrong Along the Way
Meta Is Finally Killing the Trick Where You Start Recording Then Cover the Light
A Third of Americans Now Ask a Chatbot About Their Symptoms, and Most Won't Share Their Health Data With It