This website uses cookies

Read our Privacy policy and Terms of use for more information.

⏱ 4-min read

OpenAI's model just disproved a math problem Paul Erdős posed in 1946 — on its own. Google unveiled a 24/7 agent that runs while your laptop is closed. Anthropic's CFO admitted Claude now writes 90% of their code. And Harvard just named the burnout hitting AI's heaviest users.

Today's thread: AI stopped being a tool you use. It's doing the work itself.

Let's get into it. 👇

🧠OpenAI's Model Just Cracked a Math Problem Humans Couldn't Solve in 80 Years

A problem so famous Paul Erdős offered prize money for it. Untouched since 1946. Today, an OpenAI model handed in the answer.

Image source: OpenAI

That's the planar unit distance conjecture. For 80 years, mathematicians believed square grids were the best possible answer. OpenAI's internal model just disproved that — finding an infinite family of constructions that beat the grid, using algebraic number theory no human had connected before. Nine top mathematicians, including Tim Gowers, already verified it.

The part most people are missing: this wasn't AI assisting. It wasn't checking work. The model got the problem and produced a novel proof on its own.

The uncomfortable question: If a model can autonomously crack what the greatest combinatorialist of the 20th century left open for 80 years, what does "original research" mean in 2027?

In Partnership with Cuey

Have you gotten burned by a confident-sounding AI answer yet?

Claude, ChatGPT, and Gemini are brilliant and for casual work, one answer is plenty. But for high stakes research and decisions, a confident-sounding answer is a risk most users can't afford to take.

Cuey sends one prompt to ChatGPT, Claude & Gemini in a single tab. Spot hallucinations. Cross-check reasoning side-by-side.

🧪The "AI Brain Fry" Hitting 14% of Workers Right Now

Harvard just named what high performers have been feeling: AI brain fry — mental fatigue and decision overload from heavy AI use. The people closest to AI are burning out first. 70% are hiding it.

This Wednesday's Mental Health edition of AIHealthTech Insider unpacks the research and the 4-minute reset.

Get the Wednesday Mental Health edition free in your inbox?

Login or Subscribe to participate

🤖Google Just Gave You an AI Employee That Works While You Sleep

Open your laptop tomorrow. Your inbox is triaged. Your weekly report is drafted. A hidden fee on your credit card bill got flagged at 2 a.m. You didn't do any of it.

Image source: Google

That's Gemini Spark, Google's new 24/7 personal AI agent unveiled at I/O yesterday. Spark runs on dedicated Cloud VMs — it keeps working when your phone is locked and your laptop is closed. Deep hooks into Gmail, Docs, Calendar, plus Canva, OpenTable, and Instacart on day one.

Teach it once. It runs forever. Then asks before sending anything sensitive.

Personal assistants used to be a luxury for executives. Google is putting one in your laptop next week,if you're willing to pay for it.

The real question: what happens to the VA industry, and your to-do list, once an always-on agent gets good enough to trust?

Stop Paying for 6 Tools. One AI Does It All.

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.

🔥This week in AI

Image source: Google

🔹Google Pics Wants to Replace Canva

Google launched Pics at I/O — an AI image app on Nano Banana 2 that isolates objects, edits text inside images (preserving font), and resizes anything with a prompt. Plugs into Slides and Drive. Canva built an empire on "design without skills." Google just dropped that pitch into 4 billion Workspace inboxes.

🔹Anthropic Says 90% of Its Code Is Now Written by AI

Anthropic's CFO just told the room what most labs are quietly thinking: "90%-plus" of the company's code is now written by Claude. Google, Microsoft, and Snap have made similar claims, but Anthropic is leading the pack. The company building Claude is letting Claude build the company. If the people training the model trust it to write 9 out of 10 lines of production code, what's your excuse?

🔹Antigravity 2.0 Built a Working OS in 12 Hours

Google's Antigravity 2.0 isn't a coding assistant. It's an agent orchestrator. In a live demo, sub-agents built a functional OS core in 12 hours, fixed missing keyboard drivers mid-run, and got Doom booting on it for under $1,000. One developer shipping what a team used to is no longer a thesis. It's a Tuesday.

🛠️Build a Telegram AI Chatbot That Books Appointments (Zero Code)

A clinic front desk wastes half its day on the same five questions. Now imagine a Telegram bot that handles all of it, understands natural language, books appointments, saves them to a Google Sheet, and pings the patient with a confirmation.

The full tutorial walks through the build with n8n, OpenAI, and SerpAPI. Total cost: under $5/month.

The kicker? A patient can say "book me for a broken bone, I'm free around 5ish tomorrow, name's Alex" and the agent maps it to Orthopedic at 17:00 and asks only for the missing phone number.

Same pattern works for salons, gyms, any appointment business. The full prompt, JSON trick, and exact node setup are in the build.

The World's Biggest Dev Event Hits Silicon Valley

From AI and cloud to DevOps and security — WeAreDevelopers World Congress brings the entire modern stack to San Jose. 500+ speakers. 10,000+ developers. One epic September. Use code GITPUSH26 for 10% off.

🏆Want to actually build with this stuff instead of just reading about it?

This Sunday Special: The USAII Global AI Hackathon (virtual, June 14–21) just opened with a $15,000+ prize pool and AI certification scholarships for students grades 9 through graduate. The Web Data UNLOCKED Hackathon runs May 25–31 with $5,000 cash plus up to $20,000 in Bright Data startup credits. Plus: free AWS AI & ML Scholars applications close June 24 — 4,500 winners get a fully funded Udacity Nanodegree.

New here? Get the next issue Sunday.

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

Reply

Avatar

or to participate

Keep Reading