My three cores for developing myself: Mental Model of Tech, Judgement, Taste
My three cores for developing myself: Mental Model of Tech, Judgement, and Taste.
1. Mental Model of Tech
I've used AI to code since 2022 — for data analysis, ML, and full-stack development. As I moved through many different projects, I felt like I was losing my sense of control over the code. As the volume of code and project work grew, I often had no choice but to code directly with AI (we still test, of course), but my tech stack also became fragmented and messy, because AI would sometimes keep giving me different opinions. So I think a mental model of the tech matters — for example, for the backend, if I'm using Python I can use FastAPI, and if Node.js I can use Express.js — to keep my sense of control over the code.
2. Judgement
For judgement, I don't want to talk about how to judge other people or things — I'm thinking about judging my own judgement. For example: does the logic still apply to future cases? Keep looking at my own logic. And don't just look from one view; instead, look from different perspectives — it can be horizontal, it can be vertical — to broaden my horizons. For example, when I'm trying to figure out the future of AI harnesses, I should think about how LLMs will develop, the capabilities of the models, and how that impacts the direction of AI harnesses...
3. Taste
For taste, I don't know yet. Maybe it will come along with the first two.
← Back to all posts