⏱ 4-min read

This issue: Codex can now inspect live browser issues, Perplexity is turning research into full workflows, and OpenAI is building cloud infrastructure for agents that keep working after your laptop closes.

At the same time, patients are asking AI before doctors, nonprofits are getting Claude builders, and the job market is shifting again.

The main story: AI is no longer just answering questions. It is starting to do the work around them.

OpenAI shipped developer mode for browser use, giving Codex controlled access to the Chrome DevTools Protocol in both Chrome and the in-app browser.

Highlights

Why it matters: Debugging usually involves switching between code, browser, and devtools. Codex can now autonomously monitor the network tab, read the console, and trace slowdowns to address issues directly.

TLDR is the free daily email that does the reading for you. Each issue summarizes the most interesting stories in startups, tech, and programming, curated by ex-Google and Anthropic engineers, so you get the signal without the noise.

It's one email instead of 12 open tabs. A 5-minute read with your morning coffee, and you walk in already caught up.

Tech is just the start. We also cover AI, marketing, dev, and more, so you can follow the topics that matter to your work.

Free, daily, and read by 7M+ subscribers. Subscribe and get your mornings back.

Subscribe for Free

A new survey finds 52% of patients now use AI to look up conditions and symptoms — yet most still expect their doctor to double-check whatever it tells them. Patients aren't replacing physicians; they're walking into appointments already briefed.

The latest AIHealthTech Insider also covers Mayo Clinic and Microsoft building a healthcare-only AI model, GE's faster AI-powered scans, and new WHO guidance on AI in health policy — all circling one unanswered question: when AI enters the exam room, who's checking its answers?

Your AI reads text. It doesn't read documents. Tables flatten, handwriting breaks, mixed layouts fail. This webinar shows engineers and ops leads what Document Intelligence actually means and how to build systems that don't fail in production.

Sign Up Now

🔥Big News in AI

Perplexity launched Deep Research inside Computer. It takes one complex question, breaks it into subtasks, routes each part to the right model, searches across files and the live web, then returns a report, deck, or dashboard. That means research tools are moving from “find me sources” to “run the whole research workflow.”

OpenAI is rolling out saved rate limit resets for Codex users. Instead of using a reset the moment it arrives, Go, Plus, Pro, and Business users can now save one and use it later. That turns Codex access from a fixed waiting game into something developers can time around real work.

Kimi-K2.7-Code marks a shift: coding models aren’t just getting smarter — they’re becoming faster, more efficient systems for long‑horizon software work. The gains are clear: sharper benchmarks, leaner reasoning, and higher success rates on real end‑to‑end tasks. The direction is moving from “assist with snippets” to “execute sustained development workflows.”

OpenAI announced it will acquire Ona, a cloud execution and orchestration company, to expand Codex beyond single-device sessions. The goal is to let agents keep working over hours or days inside secure, customer-controlled cloud environments.

Most e-commerce sellers juggle 6–8 tools and pay hundreds monthly to keep operations running. StoreClaw replaces the stack with one autonomous AI engine that monitors competitors, optimizes listings, automates marketing, and tracks profit 24/7. Connect your store and let AI handle the work — no prompts, no complex setup, no credit card required.

Try if for free today

Anthropic introduced Claude Corps, a 12-month fellowship for early-career AI talent in U.S. nonprofits. Fully funded by Anthropic with support from CodePath and Social Finance, it requires no formal education—only applicants over 18, early in their career, and eager to develop AI solutions for social issues.

This model sends trained builders to nonprofits, removing the need for AI budget and talent, to boost social impact and test responsible AI. The key question is whether it will significantly impact AI adoption in under-resourced sectors or stay a minor initiative.

ChatGPT can make a beautiful image in seconds. Then you ask for one small change, and the only option is to regenerate the whole thing and lose what worked. Canva's Magic Layers fixes that — it takes the flat image and rebuilds it as a fully editable design, with real moveable layers and live text.

The workflow covers the exact ChatGPT prompt format that gives Magic Layers something it can actually unpack, the step-by-step from upload to edit, and how one base image becomes a LinkedIn post, an Instagram square, a YouTube thumbnail, and a blog cover — without ever opening a new chat.

One image. Six platforms. Zero regenerating.

Read the full guide →

Adding a second database was supposed to fix things. Now you manage sync, drift, and pipelines on top of queries that are still slow.

TimescaleDB extends Postgres instead. Hypertables, 95% compression, continuous aggregates. One database. No pipeline.

Start with $1000 In Credits

45% of companies have already restructured around it — and 55% have moved entry-level hiring budget straight into AI tools. The job your degree was supposed to land may not exist by graduation. But the skill replacing it can be learned in a weekend.

This Sunday Special shows where the new door is — three live hackathons (one closes June 15), free AI courses to start this weekend, and six AI roles paying up to $485K.

New here? Get the next issue Sunday.

70,000+ AI builders read The AI Entrepreneurs every week.