ChatGPT goes into your bank account + Circle wires agents to USDC

ChatGPT Pro users can now connect more than 12,000 financial institutions — including Schwab, Fidelity, Chase, Robinhood, American Express, and Capital One — through Plaid and ask the model to analyze spending, portfolio performance, and financial planning. The integration appears in the sidebar or as an @Finances, connect my accounts slash command, with OpenAI saying synced data is removed within 30 days of disconnect and that financial memories can be deleted separately. It's OpenAI's most aggressive consumer expansion since launching Codex on mobile yesterday and pulls ChatGPT directly into the territory occupied by Intuit's Mint, YNAB, and bank-native assistants — with all the privacy and liability questions that come with reading every transaction in a user's life.

Circle Internet Group released an open-source set of agent primitives — Circle CLI, Agent Wallets, an Agent Marketplace, and Nanopayments powered by Circle Gateway — designed to let autonomous agents hold assets, discover services, and transact in USDC across supported blockchains. The headline feature is gas-free transfers as small as one-millionth of a dollar, aimed at the high-frequency machine-to-machine commerce that traditional payment rails make uneconomic. The launch arrived alongside Circle's new Arc blockchain and a $222 million ARC token presale led by BlackRock and Andreessen Horowitz. Coming the same week OpenAI wired ChatGPT into 12,000 banks, the two announcements bracket the question of how agents move money — through regulated bank rails or through stablecoins designed for them in the first place.

A May 13 virtual town hall organized by the Better Path Coalition and No False Climate Solutions PA pulled in roughly 225 viewers and over 20 speakers from multiple counties, all describing the same set of grievances: surging electricity bills, heavy water draws, noise, industrial-scale builds next to homes and schools, and what one speaker called being "bulldozed over" by projects whose permitting moves faster than community input. State Senator Katie Muth is pushing a three-year moratorium on hyperscale development, and Chester County's North Coventry Township is now drafting a local data-center ordinance. The Pennsylvania backlash echoes the NAACP's suit against xAI's Mississippi turbines from this week — community pushback against AI compute siting is becoming a named political issue, not just a permitting headache.

Oregon-based Panthalassa raised $140 million led by Peter Thiel's personal fund — with Founders Fund a returning backer from 2018 — pushing the company's valuation to roughly $1 billion. The pitch is unsubtle: pair massive floating orbs that harvest wave energy with on-board AI compute, use cold ocean water for cooling, and uplink results to land via low-Earth-orbit satellites. CEO Garth Sheldon-Coulson is targeting a northern Pacific Ocean-3 pilot deployment later this year and commercial nodes in 2027. The bet — placed the same week Pennsylvanians filled a town hall demanding onshore data centers slow down — is that ocean siting bypasses transmission costs, water-rights fights, and the community veto that keeps blocking gigawatt-scale builds on land.

PwC said it will train and certify 30,000 of its professionals on Claude, stand up a joint Anthropic-PwC Center of Excellence, and roll Claude Code and Claude Cowork out across its global workforce of hundreds of thousands. PwC is also anchoring a new Office of the CFO finance practice on Anthropic's stack, with Anthropic citing "delivery improvements of up to 70%" in current production work across professional sports, insurance underwriting, mainframe modernization, HR transformation, and cybersecurity. The deal is Anthropic's most explicit Big Four distribution lever yet — pairing its enterprise sales reach with PwC's audit and consulting relationships at the moment OpenAI is building parallel Deloitte and Accenture motions.

US District Judge Araceli Martinez-Olguin declined to grant final approval at Thursday's San Francisco fairness hearing, pressing both sides for more detail on attorneys' fees and lead-plaintiff payments before signing off on what would be the largest known US copyright settlement and the first major US AI-training case to settle. The deal covers more than 480,000 works — claimants representing 92% of them have already filed — but a group of more than 25 opt-outs including Dave Eggers and Vendela Vida filed a separate California complaint against Anthropic on May 13, arguing the per-work payout is too low. Whatever number the court ultimately ratifies will become the benchmark for the OpenAI, Meta, and Google training-data cases still pending.

ICLR 2026 work from Tao Yu and Youngsuk Park at AWS AI Labs extends Chinchilla with a conditional scaling law that ties three architectural knobs — hidden size, the ratio of MLP-to-attention parameters, and grouped-query attention — to model loss, letting designers pick a Pareto-optimal configuration before committing to a full training run. The team's Surefire-1B reference model matched or beat LLaMA-3.2-1B accuracy while hitting up to 47% higher generation throughput on H200 GPUs under SGLang serving. For a hyperscaler footing the compute bill, that's the kind of architecture-search artifact that quietly becomes load-bearing inside Bedrock and EC2 inference pricing.