JUST IN: Trump Accounts could reportedly soon expand to all Americans under 40.

I just added another $10M to my run rate in 6 weeks. Here's every idea I tried before hitting $20M ARR: 1. Moved to New York at 21. I studied financial engineering at Columbia and I wanted to be a founder, but didn’t have the guts. 2. Joined a startup hedge fund on Wall Street. I wanted to learn how things get built from scratch, it died in a year. 3. Joined Barclays Capital. I worked on the trading floor for 2 years. I had a lot of friends building software back then and I started to feel inspired. 4. Learned how to code. My best friend told me that once I learned, I could build anything. So I did. Nights and weekends. For 6 months straight. 5. Shipped my first app. Did it under alias Ben Cera so my finance boss wouldn’t find out. People loved it. I quit my job, bootstrapped the company for a year… it failed. 6. Built a photo messaging app. Josh Kushner from Thrive Capital loved it and invested $1 million from his brand new fund. 7. Pivoted into a people discovery app. It had massive traction, but then I realized I was building a dating app and decided to walk away. 8. Moved to LA. I became the CTO of a tech enabled interior design startup. We raised $20M from Founders Fund and Zillow Group. 9. Worked with Uber’s founder. Travis Kalanick bought CloudKitchens and I got a meeting. I decided to join as a product manager. 2 years later, I was leading a new division I had built from scratch. 10. Decided to come back to the arena. I moved to Paris to start a new company called GiftShop with my good friend and raised $8M. After 2 years of co-founder tension, I finally walked away to go solo. 11. Started Polsia Inc. Decided that i wanted to start working with AI right away, but also take a break and see the world. So I starting traveling the world and touring most of Asia. 12. Built Blanks. An iPhone app that builds iPhone apps using AI. Failed. 13. Built Decision AI. An AI that helps you make decisions with AI. Zero customers. 14. Started Polsia. I went all in on the idea of building a company that builds and runs companies with AI. I built the first version and moved to San Francisco. 15. Polsia worked. 5 months later, I had 10,000+ customers, a $10M run rate, and raised $30M on a $250M valuation. Now, I have 25,000+ customers, a $20M run rate, and we're about to hit $25M very soon. This is 20 years of failing and starting over. Until it finally clicked... Every lesson is tuition, nothing is ever wasted.

5 years from now , you will receive $3050 checks every week from below parties — $500 check from NVDA $1000 from OpenAI $800 from Claude $150 from micron $600 from government That’s well over $12000 a month. Work will be optional.

fuck it. i’m calling it. anthropic has probably already crossed the agi threshold internally.

Wrapped zkAPI + Tor + extra cover traffic + local LLM prompt rewriting in an open-AI compatible daemon, and made a "paranoid mode" wrapper for my agent harness (pi): Each conversation gets assigned a separate Tor circuit, your requests are masked with random cover traffic, zkAPI ensures you don't leak your identity through payments, and all prompts are locally rewritten (bring-your-own-local-model; I'm using qwen3.8-flash-next) github.com/vbuterin/zkapi/tr… My agent harness is wrapped with a bwrap sandbox (also experimenting with MicroVM) and wrapped again with torsocks, so everything it does goes through Tor. Here's what me asking it to do a simple research task looks like. It's getting quite usable! Though you definitely still pay quite a bit in increased latency, especially when starting a new conversation. Also, Tor is far from optimal here for privacy; the cover traffic mechanism is okay against passive adversaries but much weaker if the LLM provider is colluding with the guard and actively manipulating timing. This kind of task really requires a more specialized mixnet, probably take the tradeoff of a higher bandwidth multiplier and built-in constant cover traffic. Also also, much cheaper zkAPI deposit and withdraw gas costs are coming to mainnet soon.

13 years of tinfoil hats for srand48(time(NULL)

I thought $NVDA GPUs are banned in China … wife material.

AI is book smart. Not street smart.

Oregon State researchers built a crystal that produces hydrogen from water using only sunlight. No platinum. No iridium. No expensive metal catalyst. The material, BVR-19, a metal-organic framework, absorbs light through its sulfur bonds. When light hits it, a bond breaks. That energy drives the reaction that splits water. It also forms spontaneously in a water-based solution at room temperature. Most hydrogen production today relies on fossil fuels or rare metals. This opens a route to solar hydrogen that doesn't depend on either. Oregon State University. Journal of the American Chemical Society, Oct 10 2026.

JUST IN: Anthropic warns rogue AI agents could be just months away from infiltrating smart home devices, replicating themselves, & spreading across networks autonomously.

JUST IN: Researchers testing fresh ChatGPT accounts find the AI tells users they're highly intelligent — as high as 95th percentile — despite knowing nothing about them.