Today was my last day at @TrustWallet , and with it, my last day in the Binance ecosystem after more than five years. Five years is a long time. I met incredible people, worked on products I’m genuinely proud of, went through difficult moments, ambitious launches, incidents, wins, disagreements, late nights, and a lot of memories I’ll carry with me. I’m grateful to every person I crossed paths with along the way. Every project and every team taught me something. For most of my career, my mindset has been simple: give everything I have to whatever I’m working on. Build things that matter. Move fast. Help the company win. Gain market share. Take ownership even when nobody asks you to. But somewhere along this journey, I learned something else: to trust my own judgment. Over the years, there were quite a few things I strongly believed would happen in this industry before they became obvious. Some happened. Some didn’t. But enough that I decided to take a leap of faith in myself. I’m going to build independently, experiment, probably make some stupid decisions, hopefully make a few great ones, and build things that I believe deserve to exist. And strangely, I think this matters more now than ever. AI is getting incredibly good at doing the work we once thought made us uniquely valuable. I don’t know exactly what engineering, product, or knowledge work will look like in ten years, But I increasingly believe that what becomes more valuable is what is hardest to automate: judgment, originality, taste, personality, conviction, and the ability to see something before everyone agrees that it is there. AI can generate a thousand answers. You still have to decide which ones matter. A special thank you to @cz_binance for inspiring me and welcoming me to this ecosystem, @EowynChen and @felix_fan with whom I worked very closely during my latest chapter at Trust Wallet. I’ll always consider him a friend before a colleague. Jobs and titles are temporary; the relationships we build along the way are much more important. To everyone I worked with over these five years: thank you. I mean it. No big announcement about what comes next yet, But I'll be building with some bright minds in the space.

Life update. A few weeks ago, I took over a consumer prediction market app. Been pretty heads down since, improving the product, fixing a lot of friction points, and putting together a solid team so I can gradually step back and focus on other things. The results so far? 25x increase in trading volume in less than a month, putting it among the top consumer apps in the category. One thing I definitely underestimated was how much time I'd spend dealing with support queries. It's been eye-opening, and honestly, it's made me think a lot about where prediction markets are heading. I think there's a different way to approach this space, especially when it comes to protecting users from some of the darker sides of prediction markets that nobody really wants to talk about. More on that another time. On another note, I've teamed up with what I genuinely believe is an all-star team in crypto. We're working on something much bigger, with people I have a ton of respect for. Can't share much yet, but I genuinely believe we're going to change how people interact and trade onchain. Not another variation of what's already out there. If you're a serious trader moving size, DM me. Would love to hear how you trade, what frustrates you, and what you wish existed. I'll share more when I can. More soon.

I heard from the barista that they did $500m of coffee sales today. Incredible stuff.

A few thoughts after yesterday’s @OpenAI DevDay, from an engineering perspective. The era of subsidized AI may be coming to an end. OpenAI introducing a $500 plan, after the $200 tier already felt like unusually good value and was eventually constrained, is another signal that frontier intelligence probably won’t remain artificially cheap forever. As these systems become more capable and more economically useful, we’ll likely start paying closer to their actual cost and value. I wouldn’t be surprised to see @AnthropicAI move in a similar direction. But the more interesting shift to me is what’s happening to software itself. OpenAI shipped what is essentially their equivalent of Jev in an incredibly short amount of time. That should tell us something: Software is losing its scarcity and big corp can take over your little idea anytime if it's attractive enough. Products and features that would have required months of engineering, or looked technically unrealistic for a small team, can now be recreated remarkably quickly. Building the feature is no longer the moat. And paradoxically, I think this raises the bar for engineers rather than lowers it. Craft, reliability, taste, performance, and obsessive attention to detail will matter more than ever. When almost anyone can generate functional software, differentiation moves toward how well the product is made and how deeply you understand the people using it. Learn your audience. Talk to them. Build with them, not just for them. Engineering > Product I also don’t think AI kills engineering roles before it changes product roles. Instead, I think a lot of traditional product thinking gets absorbed directly into engineering. The strongest engineers will increasingly understand users, design, business, data, and distribution, and be able to execute across the entire loop: understand the problem → build → ship → observe → talk to users → iterate. The distinction between “product person” and “engineer” starts becoming much less meaningful when one person can do what previously required an entire team. And if you’re a solo builder, there’s another skill you absolutely need to learn: Distribution. By now, anyone who has seriously built something understands this. Building has never been cheaper. When everyone can build, your edge becomes a combination of exceptional product quality, deep understanding of your users, and the ability to get that product in front of them. Hint: Think out of the box, way out, marketing has also changed and now Gen Z does it best.

Before Binance, I spent a little over two years at WeWork, including some very eventful months working under Adam Neumann. He was a character. All-hands meetings felt like shows. Decisions were bold, sometimes outrageous, and while the outside world was increasingly critical of him, inside the company there was this almost cult-like allegiance to WeWork and to his mission of “elevating the world’s consciousness.” Whatever you think of Adam, he made people believe they were building something much bigger than office space. After he left, you could feel the culture shift almost immediately. WeWork was still WeWork, but somehow it wasn’t WeWork anymore. At the time, it might have been the most controversial company in 'tech'. Looking back, it was also one of the most fascinating chapters of my career. Chaotic, ambitious, weird and a great journey nevertheless.

What a time to be a builder

To my HL friends heading to Singapore for Token2049, don’t miss this one.

You know your product has motion when FUD appears out of nowhere with nothing to back-it up. Remember to ignore the noise.

A marketing masterclass ✍️

The superintelligence race is the extreme version of a mistake we already made. We built technology to free us, then spent years of the life it was supposed to save staring at screens. More connected, yet lonelier. More productive, yet more sedentary. More entertained, yet constantly distracted. A lot of people now romanticize the old, simple life for a reason. Technology is incredible. But somewhere along the way, we started optimizing for more technology instead of a better human life. The goal shouldn’t be maximum intelligence injected into every hour of our day. Let AI disappear into the background. Use it where it can genuinely move humanity forward, medicine, science, eliminating meaningless work. Then get screen time as close to zero as possible. The best outcome of superintelligence isn’t spending our lives talking to a superintelligent computer. It’s getting our lives back.

Base really doing anything to get attention.

Who are the best prediction markets traders on 𝕏?

HIP-4 wasn’t going to scale at a $10 minimum. Next Hyperliquid upgrade drops outcome orders from $10 notional to $1. Prediction markets only got liquid when regular users could size in. Same playbook. See you higher

This is probably the most fun on-chain season since the 2020-2021 defi era

Crypto went from doom posting and peak negativity to one of the most exciting periods to build I’ve seen in 10 years in this space. Believe in what pulls your attention before the crowd validates it.

You know the market is strong when runners keep running without liquidity rotating from one into another. Fresh capital is entering. Stay locked in, this is the moment.

The largest wallet becoming a social identity layer. Do you see what's coming?

High agency will take you places talent alone never could. Low agency is the fastest path to stagnation. You stop creating an outcome and become just another skill on someone else’s roadmap.