Hacker News Reader

Comment by taylorfinley | Hacker News Reader

37 points·taylorfinley··7 comments

It was probably a bug or unintended, I opted out a while back and nothing happened. I suspect anyone using it at work is opting out.

I wouldn't be surprised if OpenAI people are scouring hackernews.

I still remember that one time someone posted that GPT3 was leaking its own source code[1] and within the hour not only was that post flagged, but ChatGPT was down for immediate maintenance[2] and it no longer behaved in that way.

I'm not saying that it was actually leaking source code, but it was an extraordinary coincidence.

[1] https://news.ycombinator.com/item?id=33875231

[2] https://news.ycombinator.com/item?id=33875545

I think you must be right, I was paranoid I'd been excluded from the future, but it was likely just a glitch.

I haven't signed up (yet) precisely because I resent it being developed on the backs of the public while being a windfall for them - and that I need to provide a verifiable identity in order to participate.

Will I eventually give in? Possible. I've been meaning to learn how to set it up on my home network.

> I've been meaning to learn how to set it up on my home network.

Well I don't think you can get the weights to ChatGPT itself (or the training dataset and compute needed to reproduce them). Unless I've misunderstood what you're planning to set up?

Well, llama or one of the other models at least.

So you put down on paper that you're against the AI machine.... that's bold, dude.

Reply on news.ycombinator.com