Learning ML, even just as a tool user rather than programmer, seems to be an extremely trial-and-error high-energy grind đź‘€
Conversation
Replying to
I’m not really trying much. I’m watching people posting about what they’re doing.
2
3
Show replies
Replying to
Having to explain my hyperparameter selections when I write ML papers for class is hilarious. “What was your thought process, why did you try this combination” bro I just threw shit at the wall and hoped it would stick
1
6
Replying to
Part of it is that there are so many tools (and there will always be more) and not as much material that explains the overall process.
Replying to
Yes, but there's a snowball effect. each part that you learn increases your ability to learn the others. and the more people that get into it, the more layers of tooling are developed and shared
6







