I think you’re missing that “every” K-12 student is using ChatGPT for all their work right now. Yes, the state of education is in peril (with ChatGPT actually being pretty far down the pareto chart), but the generation coming up after us is absolutely growing up using LLMs the way we used calculators, and using it for everything.

We may not see universal adoption in people who are currently >30 but I think we will in the generations that are <25 now.

But are they using it for anything else? Even GPT-2 was "good enough" for high school writing assignments. The fact that this power user class of young people hasn't found any other use cases for LLMs despite vast improvements in LLM capability doesn't reflect well on the merits of LLMs as consumer software.

I think you may be having a failure of imagination. That generation will be creating additional tooling to use this wherever possible, and taking advantage of any interoperability they or their peers can hack onto any interface. Our generation will be hesitant but they will not be, and many of the upcoming generation will have a deep understanding of where LLM’s have appropriate vs. awkward application.

> The fact that this power user class of young people hasn't found any other use cases for LLMs

Why are you assuming they haven't? High school writing assignments and homework are the majority of the problems teenagers face daily, but they're also having fun with it, and why wouldn't they try it on new problems as they come along?

Of course students are using ChatGPT. It helps them to write assignments.

But it isn't translating into better across the board test results and at least in Australia we would be able to tell because we have yearly standardised testing.

And so schools are looking at it as more of a form of cheating and simply moving back to in-person, hand-written tests.