If hallucinations is your main issue, use it in a way that minimizes risk/exposure to it. Be creative! We invented a new tool and we gotta learn how to use it yet.

For hard logical problems, DeepSeek’s reasoning chains can help me break down the problem and formulate an attack angle when I’m stuck. It’s not perfect, and you gotta stay paranoid and verify for yourself (not much difference with other types of knowledge aqc.)

Not sure how he’s using it, but to me it’s a valuable first stop in knowledge domains I know less than avg. on. It’s a great inspiration tool helping me pick up some unfamiliar vocabulary/concepts that I can later google.

Pt. 2

Source?

F

ol! Had to watch the video now. Honestly expected something worse than a summarization machine. But yeah… now I guess I can add anxiety by not knowing if the summary I just read is actually accurate.

Also, what’s the package?

What would you say is your main issue with its code quality looking back at it now?