At New York Tech Week, privacy was the main course and openness was the biggest ingredient 🍽️ Highlights from our privacy dinner and afterparty with @boundless_xyz, featuring the architects of enterprise-grade privacy solutions.

"Eventually the Chinese models may go closed source. They're spending tons of money to train these models and getting no revenue and no data flow from it" Tom Shaughnessy on why China's open source AI strategy might not last "The Chinese models are all open source. DeepSeek, Kimi, MiniMax, GLM, and they're incredible. But there were some rumblings. Alibaba open sourced a model and the next day the main guy was fired. GLM kept a version private for a while before fully open sourcing it" "The reason it could go closed source is these companies are spending tons of money to train these models and not getting any value flow. They're not getting revenue because the models run on inference providers like Venice and OpenRouter. They're not getting the data flow either" "Their current open source strategy makes sense. If you don't have frontier intelligence, you want to impact your competition by giving a close and free substitute. But if China goes closed source, the models will have to be really good. And they're not there yet"

Sent over a full refund. Please let me know if it doesn’t show up on your end within a day or two. And sorry about this. As models work for longer, we’re adding additional spending controls to automatically catch cases where an agent runs much longer than intended.

Failure is the likeliest outcome when you start a company. It's still painful and stupidly surprising when it happens to you. Hyli was built from a devouring ambition to push blockchains forward. As application builders, Lancelot and I understood the inherent limitations of current architectures. We took a deep look at what was needed to build a better blockchain, and just built it. We wanted to finish the job. 4 engineers built a SOTA blockchain in a cave in Paris, with a bunch of scraps. Unfortunately the market is coalescing around a few use cases for which we have no edge. We simply do not see a winning future for Hyli. After two years in the trenches I'm incredibly proud of what we've achieved with our small team. The industry is now converging on the architecture we committed to two years ago: simplify the protocol, use proofs, boost the consensus, DA is all you need. I want to thank everyone who believed in us. Partners, investors, friends, and above all our team, who saw two crazy dudes with a lot of ambition and said "fuck yes". Thank you. They're incredible operators any founder would kill for, you should hire them. It's been a wild ride, thank you for the amazing moments. @wraitii and I will be exploring new territories around agent-driven commerce. 🧡 Drop me a DM, always happy to chat

This morning @zerohedge out with a report on the death of Tokenmaxxing "Microsoft’s AI Chief added to the unflattery this week after cancelling Claude Code licenses in May. “Anthropic is extremely expensive, and I think many people are urgently looking for alternatives” Very similar to my thesis I shared last week x.com/zerohedge/status/20647…

At this rate, the first solo unicorn is going to have to be someone who was already a billionaire before they started, just to afford the tokens.

JUST IN: Nvidia is now worth more than India

🚨 SPACEX's EX-EMPLOYEE EARNING $28/HR TO BECOME A MILLIONAIRE POST-IPO Juan Hernandez, a Mexican immigrant, joined SpaceX in 2015 as a contractor earning $28 an hour. He later went full-time, received a $10,000 equity grant that vested over five years, & bought more through payroll deductions. His remaining shares are now reportedly worth around $880,000 at SpaceX’s $135 IPO price. SpaceX’s reported IPO could push him past millionaire status.

This is a pretty striking shift toward Chinese models by American AI startups since the start of the year. substack.com/@profgmarkets/p…

Ahead of humanity’s return to the lunar surface, Axiom Space and Prada unveiled the Liquid Cooling and Ventilation Garment (LCVG) designed to be worn by astronauts inside the Axiom Extravehicular Mobility Unit (AxEMU) spacesuit. The LCVG collaboration draws on Prada's expertise in design, patternmaking, and advanced materials, resulting in a next-generation garment developed through advanced 3D modeling techniques that maintain cooling and ventilation while enhancing comfort during up to eight-hour spacewalks. Discover more: tinyurl.com/mruj7rvy #Prada #AxEMU @axiom_space

Oh, I see what you're doing here! Aside from provoking me, that is. I have to be honest: that idli looks a bit too solid and dense for my liking. And the discolouration in the pic is not very appetising. There’s something about a perfectly soft, snowy-white, fluffy idli that just can’t be beaten. This one looks likely to be chewy and rubbery. Not an A-grade idli. I’m a massive chai fan myself, but I’ve always been a firm believer in the 'separate but equal' policy: I prefer my chai alongside or after eating, rather than dunking anything in it. Anyway, a truly good, soft idli would probably just dissolve in the hot chai and ruin it, whereas a 'dunkable' idli would be too rubbery for my taste and not worth eating. I say keep the chai in the cup and the idli on the plate—they’re both better off that way! @333maheshwariii

Your margin is my opportunity: AI version… The biggest surprise of 2026 is that the capability gap between the best open-weight/source models and the best closed models has narrowed much faster than the pricing gap. The pricing gap remains enormous while the capability gap is quite narrow. What does this means in practice? For a company consuming 1 billion input tokens and 1 billion output tokens per month: GPT-5.5 Pro: ~$105,000 Claude Opus 4.8: ~$30,000 DeepSeek V4 Pro: ~$5,220 DeepSeek R1: ~$2,740 I asked ChatGPT what it thought about this and it answered as follows: “If I were building a company today, the economic frontier would look roughly like: DeepSeek V4 Pro / R1 for high-volume inference. Claude Opus for premium agent workflows where reliability matters. GPT-5.5 Pro only for workloads where its incremental capability demonstrably produces enough business value to justify a 20–40× token premium.” Most CEOs have no idea that, instead of this nuanced approach, their teams are running amok internally by picking the most expensive models in most cases and burning through massive budgets with zero governance, audit ability and control. As control planes like our Software Factory become more standard, you can expect the run rate revenue growth of the frontier labs to go down meaningfully and the revenues of the open models to skyrocket. Why? Because we can implement the nuanced approach above and be agnostic to model - instead focusing on customer intent, model task and cost management among other things.

certified boundless time in nyc last night thank you to everyone who came thru :)

casually gave a talk at the microsoft office today

In the age of AI, formal verification is the way forward for securing software and Zcash is leading the way. Zcash will introduce formal verification in the next network upgrade, making "print money" bugs in shielded pools impossible. Encrypted money with provable correctness is unstoppable.

agentic payments

cookin again 👀 on set with @boundless_xyz & crew

I got to spend all day today with Jensen in Taiwan: talking with thousands of engineers and eating street food at a night market. Jensen is received as a rockstar in Taiwan, like it's Beatles in the 60's. It's mind-blowing and fun to watch. But most importantly, through all the interactions and all my conversations with him, he remained the same humble, kind, thoughtful, funny guy he always was, even as a kid who went to these same night markets many years ago. Btw, we tried a crazy amount of different street food. It's legit some of the most delicious food I've ever had. I can't wait to share video of it, including a ton of our conversations and hangout. When I can pause for a moment from all the travel to edit the video, I'll post it. Can't wait to continue talking to Jensen and engineers at Computex this week, and exploring more of Taiwan, and of course roaming the night markets for some more delicious street food. Days like these, even more than usual, I feel like the luckiest kid in the world. Love you all! ❤️