I write code every night. Things I would once have given up on as reckless, months-long, or impossible to do alone, I now simply describe to an agent. I run it, throw it away if it does not work, and keep it if it does. The censor in my head that always used to stop me has grown quiet.
It is not that I suddenly became brave after starting to develop again for the first time in 15 years. Starting simply requires less courage now.
When I take out an idea I gave up on long ago, many of the reasons it would not work have already disappeared. The censor that stopped me back then was not wrong, either. The world he used as the basis for his judgment had simply changed. The world had changed before I did.
I think this change is opening the age of optimists.
Why now? Because the three barriers that stood in the way of making things have finally come down together.
The first is materials. For more than 30 years, code, documents, data, and open source have accumulated on the internet. Computing resources can now be rented when needed. The giant warehouse of everything humanity has made was already waiting for us.
The second is language. To work with machines, you no longer have to learn the machine’s language first. Now you can simply describe what you want in everyday language. It is not that people moved toward machines. Machines came toward people. There are far more people who can say what is inconvenient and what they want than people who know how to code.
The third is trial and error. Making things is ultimately a cycle of trying, getting things wrong, and fixing them. In the past, each attempt required specialized skills and a lot of time. Now, something you might have tried a few times over several weeks can be repeated dozens of times in a single night.
Even if a few barriers come down, the change is not great if the last one remains. Then, the moment that final barrier also falls, someone who was a consumer yesterday becomes a maker today. The world seems to change all at once not because the technology appeared overnight, but because changes that had been accumulating separately crossed a threshold together.
But there is a more important change. The cost of being wrong about optimism and pessimism has changed.
For a long time, pessimism was rational. If you started something saying, “This looks like it will work,” and turned out to be wrong, you lost money, people, and time. On the other hand, if you did not start because you thought, “This probably won’t work,” it was hard to know what you had missed, even if that judgment was wrong.
Wrong optimism came with a bill. Wrong pessimism did not. Pessimism looked free for a long time.
So we built systems that rewarded pessimism. Individuals put a censor in their heads, organizations piled up approval stages, and governments created regulations. Schools, too, valued people who did not make mistakes more highly than people who tried a lot. In a world where trying was expensive, this was a rational choice.
AI is changing this calculation. The cost falling fastest right now is not the cost of completing something, but the cost of finding out whether it is possible. A misconception that once lasted months becomes an experiment that takes hours. It is cheaper to build something small and check than to spend a long time reviewing it.
The price of wrong optimism is falling, while the price of wrong pessimism is rising. While I remain wrong-free, someone else gets ten things wrong and gains the clues needed for the next attempt. One person protects an existing judgment, while the other accumulates new information. By the time they make their next judgment, the worlds they know are already different.
Optimists do not get ahead because they believe more. They learn about reality faster because they try more and have been wrong more. Optimism is not the size of one’s belief. It is the size of one’s sample.
People often say that AI has made intelligence cheap. To me, there is a more important change. AI has made being wrong cheap. That changes the order of work, too. In the past, we thought, judged, and became certain before making something. Now we make what we thought of first, then judge after checking the result. Judgment moves from before execution to after it.
When the range of what is possible changes this quickly, experience also grows old faster. The phrase “I know because I’ve tried it” always has a hidden date. Something that did not work in 2018 and something that did not work last night with the latest agent are completely different pieces of information. The faster the world changes than experience does, the shorter the shelf life of experience about what does not work.
So the statement that something does not work needs a date attached to it.
When I think about Korea, this point troubles me in particular. The cost of individual attempts is falling quickly, while regulation, politics, and education still operate according to the logic of an era when the cost of pessimism was invisible.
In particular, positive regulation, where you can do only what is permitted, does not fit well with this kind of change. That approach rests on the belief that the state can list in advance everything that will be possible in the future. But in an age when new possibilities arise faster than institutions can, a list of permitted activities begins to grow outdated the moment it is completed.
The bigger problem is that the cost of wrong pessimism rarely comes to light. The reason a regulation was created remains in the documents, but the companies that never came into being because of that regulation do not remain in the record. If something is permitted and a problem arises, someone becomes responsible. On the other hand, even if an industry moves to another country because it was not permitted, there is rarely anyone held responsible. We demand accountability for permitting the wrong thing, but almost never for banning the wrong thing.
Education is similar. Outside school, the cost of being wrong is falling, but inside school, the price of a wrong answer remains high. Children are evaluated by how few mistakes they make on predetermined problems. But in the world they will live in, it will become more important to create their own questions, use multiple intelligences and tools, question the results, and try again.
Korea played the game of the last era extremely well. It achieved remarkable growth through its ability to quickly learn and standardize existing answers and reduce errors. That is precisely why it is difficult to change. What is harder to throw away than a failed system is a system that once succeeded enormously.
But the next game has no answer key. In a world with answers, a society that reduces errors is strong. In a world without answers, a society that expands the range of exploration is strong. Going forward, the total volume of attempts a society produces will matter as much as its success rate.
That does not mean everyone can set out with a light heart simply because the barrier to trying has fallen. I also meet many people who have become more powerless because of AI. The world is changing too quickly, other people are building apps in a day, and the skills I spent so long learning seem to become ordinary in an instant. When you keep watching that happen, you cannot help feeling left behind.
But when you look only at other people’s results, it is easy to miss one thing. The tools that produced those results are open to me, too. If AI has strengthened someone else’s abilities, it can strengthen mine. It can fill in knowledge I have forgotten, lend me skills I never learned, and let me test again ideas I gave up on long ago. I, too, had been away from development for 15 years, but I did not return after relearning everything I had missed during that time.
So the more powerless you feel, the less you need to try to catch up with the whole world. No one can keep up with all of those changes. One small inconvenience in your life, or one idea you have put off for a long time, is enough to begin. You only need to move one step farther than you were yesterday. If I only look at AI, I become smaller. If I use AI to start something, the range of what I can do grows.
I like this era. It feels as though an age has arrived when an optimistic person can even get statistics on their side. Optimists were often seen as people who pushed things a little too far. If they were wrong, they were called naive, and even when they were right, people said they were lucky. But in a world where the shelf life of the judgment “It won’t work” is getting shorter, leaving possibilities open is no longer naivety. It becomes wisdom and strategy for living.
Optimism is not a certainty you need to have before you begin. It comes from the process of trying for yourself, fixing things when you are wrong, and testing yesterday’s impossibilities again today. You do not begin because you are certain. You become a little more certain as you begin.
The censor in my head has not retired yet. But now, when he says something will not work, I ask one thing in return.
“When was that true?”
That is why I call this the age of optimists. Not because I believe everything will go well, but because an age has arrived when we can find out for ourselves what is possible and expand that range on our own.
