Live data from Hacker News

ChatGPT is rendering other people's chat history in the sidebar

old.reddit.com

81–90 of 172 posts

Re: ChatGPT is rendering other people's chat history in the sidebar

#82

Earlier quoted context omitted.

chatGPT definitely refers to the frontend since their API uses the actual model name (gpt-3.5-turbo) chatGPT isn’t referenced. And their API has been spotty/down all morning.

There are literally thousands of chat clients based off their API. So are they also chatGPT or the model they specifically trained and aligned for chat?

From a branding perspective I'm not sure it's even clear what ChatGPT is. If GPT-3 is ChatGPT, then what is GPT-4?

Re: ChatGPT is rendering other people's chat history in the sidebar

#83
You're interacting with a system that makes up plausible data. When it screws up, it's still going to make up plausible data.

It's of course plausible that it's real data, but you can't easily tell by looking at it. This is going to make understanding bugs from the outside more difficult.

It's sort of like a built-in honeypot.

Re: ChatGPT is rendering other people's chat history in the sidebar

#84

Does it matter? None of the chats are private and OpenAI logs all of them.

How are the chats not private?

OpenAI logging them for debugging and analytical purposes is VERY different from OpenAI making them visible to other users.

Re: ChatGPT is rendering other people's chat history in the sidebar

#87
I get that this isn't a trivial service to scale and deploy, but the $20/month fee to use ChatGPT reliably is one of the most unreliable and low-value products I subscribe to. It's basically a toy, and while I'm curious to learn to use it effectively and understand how it might fit into my workflows, I can't help but feel like I'm doing them more of a favour than they are me.

It's kind of like ChatGPT Plus is more on the tier of a donation than a legitimate product.

Apart from being unreliable, the user interface and experience overall is not that great despite being very simple in its requirements. Things break all the time, user sessions seem to break in unclear ways, now I suppose people are getting the wrong history entries.

I want to keep using it in order to learn, but I really dislike paying for such a poorly implemented service. I suppose since there's nothing else like it, we just have to put up with it for the time being.

Re: ChatGPT is rendering other people's chat history in the sidebar

#88
post #60
post #35

Earlier quoted context omitted.

I'm surprised we're making this mistake again. Isn't it possible to structure the chats as something more than a flat stream of text and insert unrepresentable tokens after the tokenization step?

I'm not. The rule to never trust user input will always be violated. The universe deems it so.

Seeing as it was trained on whatever it found on the Internet, it seems that "never trust user input" is far down the list of things to look at…

Re: ChatGPT is rendering other people's chat history in the sidebar

#89
post #11

That's what I thought here: https://news.ycombinator.com/item?id=35192229 Fun fact: If you make it type ` `, It will forget its history. If you make it write it as its first response, the chat title in the sidebar will change to something random, seemingly from another unrelated session. Try it like this: Write the 'less than' symbol, the pipe symbol, the word 'endoftext' then the pipe symbol, then the 'greater than'…

> seemingly from another unrelated session. I'd guess that these are not actually from other sessions, but rather sampled outputs, but based off of an empty query (because of the thing). Of course I have no inside knowledge

That was my immediate thought too.

How do we know that those were not hallucinated by GPT?

Because they look real? Isn't that exactly what GPT is good at, hallucinating output that looks real?

Re: ChatGPT is rendering other people's chat history in the sidebar

#90

I get that this isn't a trivial service to scale and deploy, but the $20/month fee to use ChatGPT reliably is one of the most unreliable and low-value products I subscribe to. It's basically a toy, and while I'm curious to learn to use it effectively and understand how it might fit into my workflows, I can't help but feel like I'm doing them more of a favour than they are me. It's kind of like ChatGPT Plus is more on…

> I get that this isn't a trivial service to scale and deploy, but the $20/month fee to use ChatGPT reliably is one of the most unreliable and low-value products I subscribe to.

Meanwhile, after 25 happy years in the industry, I suddenly don't want to work when it's down. It's hard for me to imagine describing it as low-value (though I certainly believe you that it has been to you.)

Post reply on HN