"The shit version of chatgpt is free"

Fixed that for you. And it will get shittier as investment dries up and the pressure to monetize increases.

ChatGPT Free uses full 4o since half a year. Rate limited of course but the limits are relatvitely high for free users.

And then downgrades to the shit version right?

Yes, they show a notification when you reach the limit and then it falls back to GPT-4o-mini.

4o-mini is more than capable of doing grade school homework.

It's the other way around, no model is capable (or should be put in charge of) doing grade school homework. So the pro version is less shit at it.

The fact that we use the LLM should always be preceded with, "I know I shouldn't be doing this but..."

We get it, you don't like LLMs, but nothing of what you mentioned pertains to the point I made initially, which is that locking down one's own account does not prevent someone else from making their own.

I do like LLMs, but I do no think they are appropriate for children. And the shit version much less so.

And if you are a parent or a big brother, you absolutely have control over what online services a child uses.

What an insane world we would live in otherwise.

Yes, I agree with you on the first point, but on the second, it's not always so easy to lock down devices, kids will always find a way around them if they want to, in my experience.

I mean, that's just shifting goal posts at this point; no one said anything about only using the free version. And anyway, Altman said they're losing money on the paid versions too, so I expect OpenAI to go under sometime in the next few years. They'll be like Xerox or Kodak, pioneering a novel innovation but ultimately failing to other bigger companies.