4 ms·

What we are seeing with students is that it's harder to get them to do work if they are demotivated, but that there's so many more tools for them to learn quick

What we are seeing with students is that it's harder to get them to do work if they are demotivated, but that there's so many more tools for them to learn quickly and effectively when they are. Yes, LLMs can be used to cheat on an assignment, but they can also be used to get instant feedback and reasonably good advice when the teacher might bring judgement. This has always been the issue of the internet: It's good at giving us what we want. It's just that many times, what we want is really bad for us. The same tool that finds friends that share a hobby is the same whether the hobby is building gundams or participating in conspiracy thinking. So what I expect we'll see is outcome divergence. For some people it's a great boon. For others, the worst thing we could have done for them. What made someone successful in the 1980s might be very different in the 2030s

aleph_minus_one 2y ago
> but they can also be used to get instant feedback and reasonably good advice when the teacher might bring judgement. Honestly, the "advice" that I got from basically all of these chatterbots when I presented them some provocative, polarizing, "very different from ordinary thinking" political statements was far more "bossy" and "lecturing" than the judgements that basically every teacher would give. It was rather easy to get some AI models to actually stop willing to continue the communication with me.
- viraptor 2y ago
  Without context of what you mean, it's hard to tell how "different" your take is. I mean, are we taking "pineapple on a pizza", "sex with kids is OK actually", or something else in between? For many "very different from ordinary thinking" ideas, failing to engage/continue is not a bad result at all.