[bko]

In reality they like LLMs because they're the highest user of them. Pew reported 64% of U.S. teens used AI chatbots, while a Harvard study found 51% of ages 14–22 had used generative AI at some point.

What you answer on a survey is meaningless. Look at their actions.

And no they're not being pressured to use LLMs, standards or expectations have not gone up dramatically.

https://www.gse.harvard.edu/ideas/usable-knowledge/24/09/stu...

[gcr]

Your last sentence is factually not true. Friends across five different companies report that LLM adoption is now a key metric in their performance review.

[thfuran]

It’s mostly not students working those sorts of jobs. But I know people taking architecture and design courses that are heavily pushing LLM use. Hopefully that doesn’t exist outside of SF, though.

[daveguy]

Yeah, "not being pressured to use them" conveniently sidesteps the fact they are RLHF trained to be as engaging (aka addictive) as possible.

[_bohm]

“Revealed preferences” are not the same as actual preferences. Treating them the same is what led us to the current situation we have with everyone addicted to social media and miserable. Also, expectations are not the only thing that could pressure someone to use these tools. If all your peers were using these tools and finishing their work in a fraction of the time it takes you, and getting the same or better grades, you would probably use them too.

[bko]

What you do says a lot more about you than what you answer in a survey.

The obvious difference between LLM adoption and social media, if you think about it for longer than 10 seconds, is that there is social pressure to use social media. Your friends are organizing and bonding on social media. None of this exists with LLMs. There is no social pressure to have an AI girlfriend, quite the opposite.

Also I preempted the "if all your peers were using these tools..." none of this applies to students. In fact i'm sure most teacher would prefer not getting AI slop. Standards have not increased.

Please read the entire comment next time before replying

[_bohm]

Social pressure from one's peers is not the only form of pressure. I don't think I'm the one who can credibly be accused of not reading the entire comment or not thinking for longer than 10 seconds in this exchange.

[davemp]

usage != positive opinion

I don’t like driving in traffic yet I do it pretty much every day. Why don’t I simply not drive?

[cryptopian]

Cars are a great example, because some parts of the world were so excited by the prospect of the automotive age that they bulldozed entire parts of their cities to make way for huge arterials and parking lots without looking closer at what they were throwing away.

[atwrk]

That is not a contradiction. Just look at social media use where you can observe the same.

People can hate on AI e.g. because they see it as a symbol of inequality and billionaires deciding important things over our heads and also actively use it.

[teeray]

> In reality they like LLMs because they're the highest user of them

It is possible to be a user of LLMs and to despise them.

[dominotw]

they are being pressured to use llm to do their homework.

teens are not using llm for fun.

> had used generative AI at some point

also this is bit of a ridiculous stat to claim "highest user"