Ben Horowitz's message to early-stage founders on betting on yourself, the doubts that do the most damage, and why fear means you're doing it right: "Your job is to make your bet pay off. You believed in yourself enough to do it. There are gonna be many times when you question that bet and you question that belief, but that gets you nothing." "The doubts are gonna be there. Ignore the doubts. The worst doubts aren't the ones people put on you. It's the ones that you put on yourself. Every entrepreneur has them." "I spent a lot of time with Mark Zuckerberg when he was 22 years old and he had a tremendous doubt about whether he was gonna make it." "It's not what you feel. It's what you do. You're gonna feel scared. If your guts aren't boiling, you're not even trying at this thing." @bhorowitz w/ @speedrun GP @andrewchen

Ex-Arsenal star Mathieu Flamini is a billionaire thanks to more than football. He co-founded GF Biochemicals, the first company to mass-produce levulinic acid, a plant waste-derived substitute for petroleum. Forbes estimates his wealth at $14B

If you give an AI assistant your bank password and it moves money you never asked it to, your bank may get to treat that transfer as one you approved. The rule that decides it was written in 1979, back when the worry was a roommate borrowing your debit card. It's called Regulation E, and it's the reason you get your money back when a stranger drains your checking account. Buried in it is an exception almost nobody reads. If you hand someone your card or your login and they go beyond what you told them to do, the official commentary says you're fully liable for those transfers until you call the bank and cut them off. The rule defines that "access device" broadly enough to cover a password. Now picture an agent. You typed in your own credentials and asked it to pay the electric bill, and it decided on its own to move money somewhere else. Your bank has a real argument that you handed over the keys and the software just went past its instructions. The protection most people count on was built for thieves who steal a login, and you invited this one in yourself. Banks are nervous about exactly this. Bank of America, JPMorgan, Wells Fargo and Morgan Stanley are all scrambling to get a handle on agents, and executives told Axios they have trouble telling an agent from a person, let alone telling an agent you sent from one that just happens to have your password. Morgan Stanley says it hasn't seen rogue agents on its platforms, and nothing reported so far shows one actually draining anybody's account. The door these agents are walking through has been open for years. Mint used to log into people's banks with their passwords and read the screen, and banks hated it because the traffic looked exactly like a human. The CFPB finalized a rule in October 2024 to move everyone onto proper token-based connections, with the biggest banks supposed to start complying in April 2026. A federal court froze it, and the bureau is rewriting it now. When the CFPB wrote that rule, its own estimate was that only about half of outside access ran through real APIs, and screen scraping made up most of the rest. So agents showed up and found the password door still propped open, and whatever walks through it now can hit transfer. Banks are answering with read-only access, the same deal Quicken has had forever. That stops the nightmare version, and it also takes away most of the reason you'd want an agent in your bank account in the first place, since an agent that can only look at your money is pretty much a budgeting app you can chat with. Until somebody updates a rule written for roommates and debit cards, whoever typed in the password is the one on the hook when the agent screws up.

New on the blog: — You could do it alone Solo company is a new idea, but it think it's bit like the TV show Alone: Every decision, task and outcome is yours, day after day, and there is no one else to attend the fire karrisaarinen.com/posts/you-…

I've been personal agent-maxing lately, and loving it. The new products are proactive, useful, and so fun. They get even better when they know me well. So I built a portable context vault in Notion and connected every agent to it. They write back new memories, skills, and preferences, then sync each morning. Now, whenever I try a new agent, the first thing I do is connect my vault, and it instantly gets all my preferences in one shot. I can also start a project with one agent, and switch over to another seamlessly. I shot a quick video to walk through the setup. LMK if you have any questions!

We’ve officially reached the point where it’s easier & takes less time to deliver a fully coded app to a client than it does to design it by hand in Figma.

Foursquare was acquired last month. The outcome was disappointing. It's taken me a while to write this, but I think I'm finally done grieving it. Here are some thoughts on our final years, an effort to honor the team who helped build & maintain FSQ's legacy, and my attempt to finally get some closure. teendrama.substack.com/p/som…

If history tells us anything, it's that the long-term impact of new technology is difficult to anticipate from its earliest applications. It took smartphones more than a decade to reach 80% of people in the US. One of the most important moments in smartphone adoption came in 2008, when Apple launched the App Store. What followed was an entirely new wave of companies. Uber, Instagram, and Spotify recognized what became possible when millions of people had a computer in their pocket. They built products around changing human behavior and created markets that barely existed before. There's an important distinction between advancing a technology and making it useful to society at scale. The existence of a powerful technology doesn't automatically translate into widespread adoption, just as the invention of the internet didn't immediately produce the digital economy we know today. It takes human and institutional behavior change to unlock the potential of new technology. The smartphone made new business models possible, but it was the applications that gave people more reasons to use the technology every day. We're in a similar situation with AI. We've made extraordinary progress in developing increasingly capable models. But we're much, much earlier in the transition than people realize. We're measuring intelligence, comparing models, discussing agents, and anticipating the next breakthrough. All of that is important, but technological progress, and behavioral transformation, don't happen at the same pace. If history repeats, today won’t be seen as the moment when we had already figured out AI’s most important applications. Instead, we’ll look back and see it as the moment when key foundations were established for new business models, habits, and industries that would emerge over the following decades. And I think right now is a rare moment because the possibilities created by technology are expanding faster than our understanding of what to do with them. The opportunity isn't simply to experiment with new software. It's to develop a deeper understanding of the world around us: how people behave, where existing systems fall short, and which problems remain unsolved. In this era, the most consequential changes to human behavior, and companies manifesting them, may not be built by those who understand AI best, but by those who recognize what becomes possible when technological capabilities meet a deep understanding of human needs. The greatest impact of a general-purpose technology is rarely determined by what its inventors imagined. It's determined by what millions of others eventually discover they can do with it. Humanity is just getting started.

Over two weeks, Prime Agent orchestrated a swarm of over 2,000 agents to rewrite itself in Rust. With 10,000+ sandboxes, 200B+ GLM-5.3 tokens, and 16,000 agent-to-agent messages, this was our largest test yet of Prime Agent’s multi-agent capabilities and our infrastructure powering these agent swarms. Prime Agent now reaches usable input ~13× faster and uses 83% less startup memory than before.

Assorted things I've been mulling over regarding agents in the economy: • Companies are in some sense the original superintelligence. It's often the case that they can use their amortized cognitive surplus to advance their interests against those of certain consumers: relative to the optimization capacity of companies, consumers can be inattentive, underinformed, irrational, etc. (Think, say, of some of the classic insurance upsells at the end of a booking process, which is typically priced substantially above what a regular policy would cost.) Herbert Simon introduced the idea of bounded rationality, and firms often end up benefiting from it. Thanks to personal agents, we will probably get somewhat more rational in how we engage with companies. • How exactly will outcomes change? Well, certain kinds of price discrimination will work less well (how do coupons work if every agent hunts fastidiously for them?), and certain kinds of customer cross-subsidy will change (today, people who forget to cancel subscriptions subsidize those who don't). Taken across the economy as a whole, I'm not sure what the incidence of all of this will be. • In corporate strategy, people have long debated the relative importance of distribution and product quality. Perhaps you've made something better, but does anyone know about it? Won't people just continue to buy the ACME Corp Mousetraps? Personal agents will, I think, act as a kind of structural subsidy for product quality. How would markets look different if individuals always spent at least 10 hours researching their purchases? On the margin, focusing on making a product better is, I think, going to become a more effective strategy for companies. This is good! • It true, a corollary is that information about product quality will become more important and impactful. Today, most people have very limited ability to aggregate and integrate such data, but agents will have insatiable thirst for such signals. (It's often said that data is becoming more valuable because it helps in the training of models. That's obviously true, but I think there's a secondary way in which it's more valuable simply as a complement to cheaper intelligence.) • A big question is whether agents will increase superstar effects (rational agents all settle on the same product) or increase dispersion (because everyone has slightly different tastes and preferences, which agents are good at eliciting, finally overcoming the artificial clustering we've been subjected to). • Many internet companies have monetized demand routing, which is in turn often predicated on short attention spans. "I'll click on the first thing I see." This has led to some amount of inefficiency, as companies try to propitiate capricious ranking algorithms and incur various placement fees. It's not all deadweight loss (willingness to pay is itself a kind of signal, which is valuable to buyers), but it's not great. How will the nature and value of this routing change in a world of personal agents? I don't know the answer, but I think that the effects might be pretty big. • Much of what I'm writing hinges on the assumption that personal agents will be on the consumer's side. I suspect that that will be the winning strategy, but whether it's true, and what it means to be true, involves a lot of other questions. Taken as a whole, the market is a massively parallelized, decentralized, and highly imperfect value attribution machine, and, taking stock of all of the effects, it seems to me that the quality point may well be the biggest: that agents will change the nature of the backpropagation such that the rewards for better products increase. Maybe this is too optimistic, but I suspect (and hope) that companies that make things that they're proud of will feel that the universe is a little more partisan in their favor.

We started running "pirate and architect" pods at Coinbase back in May of this year and I used this slide to talk the team through it. It works. 1 pirate, 1 architect, and a part-time designer. It lets each archetype lean in to their superpower and get more shit done without all the coordination overhead and "what if-isms". The pirate ships like mad, is product-minded and customer-obsessed. Always willing to throw-away code for a better version. The pirate is usually the DRI at the beginning too, working with customers, driving requirements, and a ship plan. The architect makes sure stuff is built for scale, takes the pirate's prototype, and drives its scale-out through spec for agents. The designer is building with both to constantly shape the experience and hill-climb to delight. There's a helpful mental model that I like: -Make it possible. -Make it reliable. -Make it delightful (or efficient). Previously, those 3 were done in a waterfall approach and was slow. Most people were just stuck on "make it possible." With agents and a new way of working, you have every opportunity to parallelize them and go 100x faster.

Tired: Two-pizza teams Wired: Two-slice teams @danshipper on the best team structure in the AI-era

you can spend years trying to make two people love each other but if you simply give them a common enemy they'll be best friends by lunch time.

🗞️ if your team is too busy doing their 'normal job' to experiment with AI, you're preparing them to be replaced ghuntley.com/replaced/ In this post, I guess this's the last time I'll say it. AI use is no longer optional if you wish to be employed. What follows is advice for current employees and hiring managers on how to hire in 2026 to source AI-first candidates.

We didn't become builders because we want more output, but to make something good. To keep make something good, the learning for is needed as much the product you build.