Good morning. It’s Friday, May 15th.

You read. We listen. Let us know what you think by replying to this email.

Your current trading platform is probably letting you down

Liquid is one of the fastest growing trading platforms, allowing users to trade stocks, commodities, FX, and more 24/7/365 from their phone and computer.

Trading on Liquid is as simple as:

The best part is that Liquid markets never close. So no matter what is going on in the world, you are able to keep your portfolio positioned properly.

Start trading 24/7 today

OpenAI is pushing Codex into the ChatGPT mobile app, giving users a way to steer live coding sessions from iOS and Android. The idea is simple: start a coding agent on your machine, then keep managing it from your phone while it works. You can review diffs, approve commands, inspect screenshots, look at terminal output, and switch models without being chained to your desk.

This matters because coding agents are becoming less like chatbots and more like remote employees. OpenAI also published details on its Windows sandbox for Codex, while GitHub launched a Copilot desktop app that gives developers isolated agent sessions tied to issues, pull requests, and repo work.

GitHub also added the ability to start Copilot cloud agent tasks through the REST API, which is a bigger deal than it sounds. That turns agentic coding from something you manually open into something internal tools can trigger automatically. Google is heading in the same direction with Genkit Middleware, a new layer for approvals, retries, fallbacks, and observability in agentic apps.

The boring interpretation is “more developer tools.” The correct interpretation is that AI coding is turning into a work queue. Read more.

Anthropic announced a $200 million partnership with the Gates Foundation to push Claude into global health, education, agriculture, and economic mobility projects. The program will run over four years and includes grants, Claude credits, technical support, connectors, and benchmarks.

Reuters framed the deal as a major attempt to bring AI into health and education work in lower income countries. That is the idealistic version. The strategic version is that Anthropic is trying to make Claude feel less like another enterprise chatbot and more like civic infrastructure.

This comes as Anthropic is also gaining ground in business adoption. According to Business Insider’s writeup of Ramp’s AI spending index, Anthropic recently passed OpenAI in business AI adoption among Ramp customers. Anthropic is also publishing policy arguments about how the U.S. could build a one to two year AI lead over China.

The pattern is clear. Anthropic does not just want to win chat. It wants to be the serious AI company for governments, enterprises, nonprofits, and institutions. Read more.

Cerebras opened 89% above its IPO price in its U.S. market debut, giving investors one of the first clean public-market ways to bet on the AI compute story outside Nvidia. TechCrunch reported that Cerebras raised $5.5 billion by selling 30 million shares at $185.

Cerebras is a proxy for the entire question underneath the AI economy: who gets paid when everyone needs more compute? Business Insider called it the opening shot in a potential year of AI mega-IPOs.

The timing is perfect and ugly. On one side, investors want more AI infrastructure exposure. On the other, the public is starting to push back against the physical footprint of AI. A new Gallup survey found that 70% of Americans oppose AI data centers being built near them, and The Verge noted that the backlash is becoming a real political constraint.

AI is software until it needs land, water, power, chips, debt, and zoning approval. Then it becomes industry. Read more.

OpenAI is reportedly exploring legal options against Apple after its high-profile Apple partnership failed to deliver the kind of ChatGPT integration OpenAI expected. The partnership was supposed to put ChatGPT deep into Apple Intelligence. Instead, Apple’s AI rollout has been slow, awkward, and underwhelming.

Bloomberg reported that OpenAI may issue a breach notice or take other legal action. The Verge framed it bluntly: the deal has not worked out as planned. TechCrunch added that Apple has a pattern of making partners feel like they are doing work for a platform that never fully shows up for them.

The bigger story is that AI distribution is getting tense. OpenAI wants access to Apple’s billion-device surface area. Apple wants optionality, and is reportedly testing Claude and Gemini as it tries to rebuild Siri into something competitive. Meanwhile, 9to5Mac says the timing is especially awkward with Apple’s next developer event coming soon.

AI companies need distribution. Platform companies need models. Nobody wants to be the replaceable layer. Read more.

SynthBoard is a decision intelligence platform using AI advisory boards to conduct stress tests and boardroom simulations.

Liminary is a research retrieval tool using siloed AI to help strategists access and trace relevant data.

display.dev hosts agent-generated artifacts behind company auth, enabling secure HTML publishing and team feedback with one command.

Latitude for Claude Code traces coding sessions with one command, surfacing system prompts, tool calls, and per-turn token costs.

Spellar 3.0 captures meetings to build context-aware memory, retrieving cross-session details and decisions across multiple trusted AI models.

Your feedback is valuable. Respond to this email and tell us how you think we could add more value to this newsletter.

Interested in reaching smart readers like you? To become an AI Breakfast sponsor, reply to this email or DM us on X!

Thinking of starting your own newsletter? AI Breakfast readers who sign up with Beehiiv receive a 14-day free trial and 20% off for 3 months.