kohllision

5 ms·

1. 1.▲
   by kohllision 8d ago

   That should be the easiest part of the loop. Frontier models are good judges of human preference and taste and should be excellent at prioritizing proposed questions. The generation part is much harder.
2. 2.▲
   by kohllision 6mo ago

   I appreciate the sentiment, but this remark is far too common here. It does not address the content of the post and could apply to any submission from LessWrong. The author of the post has made none of those technical choices.
3. 3.▲
   by kohllision 4y ago

   > I think the direct impact of these competitions to hiring were minimal. I can confidently say that this is not the case in India. I have a countless list of people (including myself) who received emails from Google recruiters for advan
4. 4.▲
   by kohllision 5y ago

   This is comparable to the argument that programmers are learning how to eliminate compiler/linter warnings rather than learning how to program better. Immediate feedback is the best learning tool and one leads to the other. Of course,