Chamath Palihapitiya

When the red line catches the blue line, if the cost of the blue products/companies don’t go down to be the same cost of the red products/companies, the red companies will win. https://t.co/dP2ycCX136

Using AI to code, without a clear intent upfront, is just AI slop waiting to happen.

Institutions are inefficient. Experts are overrated. Capital should flow toward builders. Technology compounds. Independent thinking wins.

I still believe this… https://t.co/fbgJBinI6y

He foretold it…

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 lo…

Capital investment cycles are not always extremely short - like they are in tech. In many non tech businesses, they can be extremely long and sneak up on you. And it can affect investor returns massively. Look at Airbnb. I was curious to understand why the company was range bound despite the underlying economics. What few hosts realized when they joined Airbnb is that after many years of guests, their homes would become “well used” and sometimes dilapidated and require upgrades, updates and general maintenance/TLC. Otherwise guest quality would trend down and their pricing power would go down with it. Even though Airbnb is asset light, aft…

I still believe this.

Beast…Dan Loeb.

SpaceX just paid for the https://t.co/KCtjqTdin6 acquisition many times over.

Deep Dive: How Close Are We to Living Longer?

Cancer, heart disease, dementia, and type 2 diabetes are funded, researched, and treated as separate problems. However, they share one underlying driver: age. It is the single largest risk factor for many of the diseases that kill people in the modern world, ahead of smoking, blood pressure, or cholesterol. Cure cancer outright, and average life expectancy rises only about three years, because the underlying risk keeps compounding. This is one of the premises behind longevity biotech: if aging is the one biological process that drives many diseases at once, then slowing it could delay all of them together, and let people live longer in good…

Control plane! Control plane! Control plane! You will hear this term of art a lot going forward. Why? Because in this next phase of AI, companies will want something to sit above the models. They will want control over their AI spend. They will want the flexibility to pick certain models over others, to have flexibility on spending their budgets and more tightly tying that to measurable outcomes. But most importantly, leaders will want to have INTENT around their AI spend. Vibe coding is dead. Now comes the serious ROI driven analysis that is associated with every other serious business initiative. This is why we built Software Factory. …

Walking the floor with Thomas Keller was an experience of a lifetime. It’s amazing to see a 🐐 in action and deep in his element. Very thankful. 🙏🏽🙏🏽

Our team will be doing a live walkthrough of Software Factory on June16 at 9am PST if you want to tune in and learn more.

If Chapter 1 was all about broad, open-ended experimentation, Chapter 2 may well be about realism and rationalizing the costs of Chapter 1 and what is sustainable moving forward.

We just added multi-repo indexing. Now all your projects can live inside of Software Factory. Why would you want that? Most large organizations have a complex application stack with tons of microservices - some are mono-repo, some services and products sit across repos. Managing this is difficult and the models can easily get confused when you point an agent at a problem. Software Factory allows you to create one holistic knowledge graph across all your code. Then apply powerful collaboration, governance and control features and a large cross functional team can now build software in a very powerful, new way. See why many enterprises ar…

I still believe this.

You will be hearing a lot more about governance and a “control plane” for AI over the next few quarters. This is what 8090 built Software Factory for. Find out more at https://t.co/UAxNYUWgf0

A year ago, I announced that I had gotten a California liquor license! Why? So @Sommjoshua could work with wineries to buy wine directly and bypass much of the middleman fuss and markup. We had almost 50 Founding members join us with the goal of making sure that winemaking wasn’t lost in this age of ever-marching technological process. Our Founding Members buy wine together. We travel the world to have dinners together. We sell wine to each other. Always at cost, never with a markup and usually with a discount that can be as high as 30-40% on even the most exclusive wines. We also wanted to become a safe harbor for multi-generational win…

AI doomerism exists, in part, because solutions to our most important problems still remain unsolved while examples of superficial process automation abound. We’ve lost the script. If AI wants to be celebrated, it should put solutions to some of society’s greatest problems at the front of the line. Case in point…why hasn’t anyone focused on an AI for fixing the growing illiteracy of our children?

On every metric, American kids are so much dumber than they were 5 years ago and we are grade inflating away the reality.

Agreed! This is what Software Factory helps enable - it organizes and creates the artifacts that guide all downstream productivity. Increasingly we are finding that large enterprises want a control plane that organizes the tribal knowledge while also providing governance, safety and control. This seems to be accelerating as the models asymptote on quality. Try it here: https://t.co/fkfTXgdfXK

All Categories - 12th in the country after posting a day ago… 🫡 https://t.co/rLisRyNFP1

Incredibly interesting take from @bgurley

Pod!!!! Absolute banger with @bgurley

Very amazing email to get: “Hi JCal, Chamath, 55 days ago you two had a 4 minute conversation about EAs and Athena. Since then we have: closed $29M in new revenue 🤯 seen a -53% decrease in blended CaC, placed 683 Athena executive partners ❤️ 683 new jobs!!” If you are looking for an EA, try Athena - they are excellent, only cost $3k/mo and you are helping some amazing people around the world upgrade their quality of life. https://t.co/WYgRR4aLBP

This is awesome. I’ve worked with Sue for more than a decade now. She started Brilliant - a learning site for advanced learning - which has now taken all of that decade-long training data and created a helper to make your kids smarter. It doesn’t matter their level, this adapts to them and brings them along. Please consider trying it.

A controversial but rational tax policy would index taxes directly to the proportion each of labor and capital actually makes in relation to each other. As one cohort “over earns” their taxes should naturally go up while the other’s should naturally go down. Specifically today, this would mean lower personal taxes and higher corporate taxes. But instead we have the opposite!

In a nutshell: “ Do we let the fox into the henhouse or not?” As the frontier models continue to converge and asymptote, it will be increasingly valuable for enterprises to govern which models get used and in which ways and prevent lock-in. Software Factory enables this for large teams and organizations. Try it here: https://t.co/fkfTXgdfXK

Yup.