AI's leaps forward owe a great deal to reinforcement learning. If there is something you want to improve with AI in the real world, it is worth looking at the world through the lens of reinforcement learning, starting by making your goals clear.
- Goals — Know what you want.
- Evaluation — Know what counts as doing well, or poorly, in pursuit of it.
- Time — Once the goal and the method of evaluation are clear, AI moves toward it. Weren't sprints in human agile organizations two weeks long? AI agents iterate so quickly that it can feel like a time warp.
- Prediction — Run these processes in parallel and, with a little exaggeration, you have a multiverse simulation.
- Putting it into practice — Today's decision-making is about taking the options that produce results closer to your goals and applying them in the real world.
Ultimately, the most important thing is knowing your goals—what you want. If the goal is ambitious, you also need the will to enjoy and work through the long, arduous process of achieving it. To know what you want, you have to understand the desires and needs beneath it, whether they stem from a sense of lack, a defense mechanism, vanity, or anything else. Nobody else can discover those for you. When you dig all the way down, understand yourself, and discover the zeitgeist—the desires of the world—you are ready to make the best use of AI. That is also the starting point for building a company.
At this point, when computing resources are finite, the way to use AI well is not to become familiar with tools—that is no longer a human's job—but to face yourself, observe the world, know what you want, and put your will into action.