2023-03-24

Woke up at 7 AM today for a 7:30 meeting. This actually feels quite good. There's no running around to the next meeting, and I have time to actually think before getting to work. Will see if I can make this a regular thing.

I finished Masters of Doom yesterday. One big thing that stuck with me from the book: Carmack was focussed on the work while Romero became more and more distracted. Carmack had more success by the end. Another thing that you learn: Romero tried to hire a lot more and ship but it doesn't work that way. Carmack had a quote that said something like "There is an ideal number of people on a project. If there are more than that, productivity decreases. For id, the number is small". This is quite relevant to today's layoffs and such.

Next book I'm reading is The Making of Prince of Persia.

AI is coming for our jobs. I think a lot of people who churn code out without much sense of what problem the business actually solves are going to have a hard time. On the other hand, it's a great time to be a solo dev! You have interns and pair programmers available on demand!

This seems like something I should read and know about: https://react.dev/learn/you-might-not-need-an-effect

Uploading myself

Noting down an idea, need to flesh out: if text is the universal interface, writing as much as possible is a path to immortality. You're basically changing the data in your brain into a format that a computer can read and understand. (https://twitter.com/iliekcomputers/status/1639163648966004737)

Exploring this more, what if I fine-tuned GPT-4 on everything I'd ever written and then chatted with it. How would that work? Would it be good or bad?

Deep learning

Trying to solve this Patent matching kaggle problem using things I learned from Lesson 4, but struggling.

I need to spend a lot more time actually coding things before I can say with some confidence that I know a little deep learning.