Unfortunate that it ends this way, but still, thanks for taking the time to argue with me.

In grade school I agree, in high school though…. But even in grade school. These kids are already using it at home, wouldn’t it be good to teach them to use it properly rather than drinking everything it says as as fact?

I’m not sure why you don’t see the value though. If the market rewards such experience, why should it not be taught? Is that not the purpose of school? To prepare kids to work? Though honestly there should probably be mandatory personal finance courses before there are AI ones. That’s one area most people severely lack.

As for it being “dead simple to use” I guess that means you have no idea how AI works. Even people making them, the best in the field, still aren’t using LLMs to the best of their capacity. This is part of the current ongoing research. Of course if all you ask it for is how long to cook your pizza for, then yeah I guess it’s dead simple.

On *them, but I think you knew that.

What a dumb comment. How am I wasting time if it saves time? Adding a small 5% overhead script to be able to save 90% of the original task’s time is not wasting time. Besides, you’re already using pydantic and structured json output anyway so half the veritication is free.

That’s on you, when using AI you always have a verifier script afterwards, it’s basic common sense. And if a script can’t do it then human, if you’re saving 90% of the time instead of 100% it’s still fine.

Rimming AI? It’s reality my dude. None of what I said was false.

I swear people around here pretend to be so different, then their first solution to the AI dilemma is censorship rather than education.

“I don’t want people to know about X because I don’t like X, so we should ban X” what is this a right wing forum?

I’m not sure what your point is. Garbage in garbage out, that’s just the law of computers. That would br the pilot’s mistake.

Because as it turns out, it’s far from just a chatbot these days. Coding agents are but an example.

Kids need to learn about it because companies are using it more and more, and they prioritize hiring those who know how to use it. This will only increase as AI is proving itself to be more and more capable.

In this context, that means understanding how it works beyond the initial “chatbot”. That means knowing what agents are, what turns are, what tool calls are, what memory is and why the quality of the harness matters so much.

But it also includes knowing about it’s dangers, such as hallucinations and sycophancy and how to mitigate them (give it a way to prove its result for hallucinations and neutral questions for sycophancy).

Though I agree before high school kids probably don’t have the ability to judge its output yet. Hell most adults don’t.

That’s dumb. The “work” in maintaining this excel file was the prompting and overall agent logic. In reality you’d be using something like openrouter and the instant claude “dies” for whatever reason you swap to whatever the freshest AI is. This never becomes ancient, it keeps up with the times.

Probably because of the extra butter, because man you probably killed someone’s pan with that comment.

Saves input tokens and uses more output because of OCR, and output is priced way higher.

True, but I think the kind of people who send those monster PRs either already have AIs cosign it or they aren’t reading the policy anyway. There have also been a number of studies that show LLMs usually ignore this kind of instruction.

half-sunken ship

If you think the ship is ever sinking I’ve got bad news for you…

Because the idea is to judge the code by its quality, not by who or what wrote it.

Or so people have been saying for the past 3 years.

Probably cause it’s a dumb comment, you people see conspiracies everywhere.

And yet it’s LLMs doing these things, not only ML. People don’t cure cancer either, but the things that get close (eg. AAV viruses) are currently being done with LLMs.

More like because people are addicted to sugar.

I never said it’s a mental illness, it’s normal human behavior to get addicted to things.

Like many things, it’s only healthy in moderation. If you’re one of those reddit dudes janking it three times a day yet say they’re not addicted, it’s very likely an addiction.