News for hackers. Stuff that matters.
Re:Anthropic: "Applicants should not use AI assistants" by croes on Monday February 03, 2025 @04:17AM (#42916405)
But how likely is that?
Re:Anthropic: "Applicants should not use AI assistants" by janalsncm on Monday February 03, 2025 @04:22AM (#42916443)
It was the same way I think a lot of us used textbooks back in the day. Can’t figure out how to solve a problem, so look around for a similar setup in the chapter.

If AI is just a search over all information, this makes that process faster. I guess the downside is there was arguably something to be learned searching through the chapter as well.

Re:Anthropic: "Applicants should not use AI assistants" by weebull on Monday February 03, 2025 @04:27AM (#42916479)
Homework problems are normally geared to the text book that is being used for the class. They might take you through the same steps, developing the knowledge in the same order.

Using another source is probably going to mess you up.

Re:Anthropic: "Applicants should not use AI assistants" by its-summertime on Monday February 03, 2025 @04:54AM (#42916633)
> something to be learned searching through the chapter as well

Learning to mentally sort through and find links between concepts is probably the primary benefit of homework

Re:Anthropic: "Applicants should not use AI assistants" by TeMPOraL on Monday February 03, 2025 @05:09AM (#42916722)
Depends. Do they care about the problem? If so, they'll quickly hit diminishing returns on naive LLM use, and be forced to continue with primary sources.
Re:Anthropic: "Applicants should not use AI assistants" by Keyframe on Monday February 03, 2025 @04:21AM (#42916436)
doesn't sound much different than googling and finding a snippet that gets you unstuck. This might be a shortcut to the same thing.
Re:Anthropic: "Applicants should not use AI assistants" by Dylan16807 on Monday February 03, 2025 @04:26AM (#42916477)
But they asked how likely it is. My guess is it's a pretty small fraction of problems where you need to get unstuck.
Re:Anthropic: "Applicants should not use AI assistants" by Keyframe on Monday February 03, 2025 @04:34AM (#42916516)
fair enough