Everything You Built Last Week Is Worthless

The EX Venture campus is a large room with high ceilings, whitewashed walls and windows onto the rice terraces. No standing desks. No craft coffee. No Ping-Pong table. Just laptop density that would embarrass a Silicon Valley office manager. The morning light comes in sideways at 6 AM and you can count the dust in it. The interns arrive in clusters. Some from KAIST in Seoul, where failure is a thing your parents do not discuss at dinner. Some from Berkeley, where failure gets rebranded and turned into a founder story. Some from Bocconi in Milan, where you have to convince your family you are not wasting a world-class degree. A few from Singapore, where practicality is religious dogma. A couple from Tokyo. There are 35 staff here and more than 200 interns through this campus in a year. The ones who have been here three weeks already know what is coming. The new ones can smell it in the air. Something like electricity. Something like dread. Something like the moment before the ground shifts. At 6:30 the circle forms. No agenda. No deck. Nobody sits down, because sitting down softens a message.

There is a new intern. Yuki. Top of her class at Tokyo University four months ago, then four months inside a proper Japanese company with a pension structure and a future everyone at home understands. She spent the weekend preparing how to present last week's deliverables in the best possible light. In most rooms that is the correct instinct. In this one it is the wrong instinct entirely. Everything you built last week, I tell the circle, is worthless. The tools underneath it are outdated. The assumptions underneath those are outdated. The market you were solving for on Tuesday is a different market today. The solution you arrived at on Thursday has a better version now, or it will by Friday. Yuki's eyes widen. I watch it happen in real time. She has been trained on continuity. Do good work this week, build on the good work you did last week, get graded, adjust, repeat. Progress as a line going up. Here it is not a line. It is a loop. The five people in this circle from KAIST are wondering what they will take home. The answer is not a skill they can put on a CV. It is a habit: build, throw it away, build again. Every other skill is tactical. This one is existential.

Every Monday we forget what we did. Then we answer three questions. If you can answer them honestly, you are positioned for what is coming. If you cannot, you are probably swimming against it. Where is your information coming from? Primary sources, actual transactions, actual measurements. Or consensus, which is always behind reality. Most people are working from a trend because the last person who repeated it sounded confident. That is not information. That is weather reporting from inside a closed room. Where is your spending going? Not what you wish it were doing. What it is actually doing. Subscriptions renewing on momentum. Credentials that expire before they pay back. A rent line that quietly contradicts the word independent. If your priority is independence and 80% of your income goes to housing and transport, you are working against the thing you say you want. What is your actual trajectory? Not what you hope it is. Income growth, skill demand, sector velocity, modelled honestly. If you are earning $55,000 and spending $52,000, the trajectory is negative and you are one bad month from the floor. The probability that any of those three answers flatters you is low. The probability that you are pricing the risk at exactly zero is close to 100%.

This is where I lose people every time. Most of them have one tool they trust. One model of the world. One answer to every question that matters. They know how it works, they have integrated it into their workflow, and they are locked in. They are confident. They are also probably wrong, because they are navigating a multi-dimensional problem with a single instrument. There are multiple frontiers. Claude. Kimi. DeepSeek. Grok. Each trained differently. Each failing differently. Precision in one. Speed and usable frameworks in another. A contrarian angle in a third that may be wrong and is still worth testing. So I ask a market-sizing question five different ways and line the answers up. The differences are where the signal lives. What survives all five models is real. What appears in one and dies in the next is an artifact of how I asked. Most people call that inefficient. Running the same analysis five times. What is actually inefficient is running it once, building for 18 months, and discovering that the assumption at the centre of it was an artifact. Eight minutes of multi-path thinking against 18 months in the wrong direction. You tell me which one costs more.

The only real failure is standing still and calling it strategy.

I tell them about Lehman, not the collapse, but the moment right before it. Smart people. Very smart people. Entire careers built inside a single assumption, rowing in the same direction inside the same boat. The probability of that building failing was not zero. Almost nobody in it priced it at zero. They priced it at 1%. Look at the leverage, the subprime exposure, the correlation, the concentration in mortgage-backed securities, and the honest number was closer to 15%. Not certain. Nowhere near negligible. A real, measurable, quantifiable risk that an entire organisation was blind to, because everyone inside it was subscribed to the same story. Talent is a fraction of the equation. Macro is the ocean and you are the swimmer. You can have the best stroke in the water and still lose, if the current is running the other way. In the 1990s the current ran toward New York. After 2008 it stopped running there. I have spent my life reading which way the water moves and standing in that direction. I have built companies across twelve countries doing exactly that. Not because I am smarter. Because I checked the water before I swam.

By the third Monday the rhythm is clear. Everything is worthless. Question it. Map the probabilities. Move in the best direction. Build. Test. Learn. Fail. Iterate. By Friday something has shipped. By Sunday you are psychologically ready to throw it away. At 9:30 PM Yuki asks whether what she built is good. I tell her I do not know. Does it work? Then it is good for now. Test it. See what breaks. Fix the breaks. By Wednesday you will know whether it is good, or whether it is good at one specific thing that nobody actually wants. She looks disappointed. That disappointment is data. Save it. Some people never make the transition. The reset feels like failure every single time and they leave by week six. The ones who stay reframe it. Everything is worthless means you are not married to what you already built. You are free to move in a new direction without it costing you an identity. The only real failure is standing still and calling it strategy.