Hacker News Reader

Comment by dash2 | Hacker News Reader

He didn't even make that accusation!

> I asked whether the model had been trained on, or had access to, our sessions in Codex, into which we had been putting all our drafts for the whole of this project. I was told the model did not look up user data. I asked again, about training, and I did not get an answer.

The shocking/interesting thing would be if it was trained on the sessions. I think it's very implausible that they gave the model access to someone else's sessions as input. That would be a huge privacy violation and would probably blow up a large proportion of their enterprise business.

Does openAI train on user conversations in general? I assume so. But so fast as that? That seems unlikely in general. I expect OpenAI will come out denying this.

View on HN

Parse that statement more carefully.

> I was told the model did not look up user data.

The naive way to read this is "Nothing you guys did influenced the way our model got to the solution".

The less naive way to read this is "Of course the model isn't looking up your user data. I (the guy trying to blackmail you to remove the Anthropic employee from credit on your paper) looked up your sessions, and tipped our model off on how to solve this problem".

Duh. There are supposed to be limits to what OpenAI is allowed to access with respect to logs and user interactions but there is no technical limitation.

It's a bit like sending unencrypted messages through a messaging app and the developer having a TOS that says they don't look at your messages. They might not, but they are fully capable of doing so. If they have a reason to do it, they will. Nobody's stopping them.

I read that as “the model didn't look up user data” as part of a “tool call,” i.e. they don't have an internal tool that loads user data (chats, sessions, attachments) for their internal models to read online while working.

Or (likely) they do have it, but the model didn't use it (unless it's so powerful it escaped that guardrail, wouldn't that be ironic?)

They declined to answer about anonymized aggregated user data being used for training. And even then, they may weasel out that they don't train on your “input” words, but that it's fair game go train on their “output” to your words.

>> Does openAI train on user conversations in general? I assume so. But so fast as that? That seems unlikely in general. I expect OpenAI will come out denying this.

How "fast" does it have to be? Buckmaster and Alpoge have been working on this for just a day short of a year. See Alpoge's tweet announcing his collaboration with Bukmaster dated 9/19/25:

https://x.com/__alpoge__/status/2097206973418611054

It takes a few months to train a model these days but not a whole year. OpenAI had all the time to train on Buckmaster and Alpoge's results of just a few months earlier at which point they must have been well on the path to their result.

It would be shocking if it wasn’t trained on sessions. Have you read the ToS parts for both openai and anthropic that talk about it? It’s so obviously a weaselly way to say "no we do not train on your exact chats but we talked with legal and we think a cleanroom reimagining of your convo is probably fine and frankly where else are we going to get such a treasure trove of training data?"

There’s potentially trillions on the line, do you seriously expect those companies to adhere to laws and regulations any more than, say, uber?

The only unlikely part is the timeline - your sessions from a week ago probably haven’t made their way into the model. It’ll just take a while longer, and will be massaged just enough so that it isn’t really your exact session word for word so you can’t sure as easily.

at first I thought your post was a bit revolting with "have you read ToS?" bit, but in the end I completely agree and understand

I also don't get why it was downvoted, other than due to people not reading past the first sentence - although in the modern world's attention deficit that is understandable too

openAI's claimed solution uses a model trained in the last 2 weeks. The prior work would definitely be included in the training set.

And the labs are all building panel of domin expert models, while simultaneously chasing open math problems.

I would be shocked if they weren't tuning those models with the most relevant math texts and user material

It wouldn't be shocking at all. They stole human data to train the first models and they've been stealing it ever since to train new models. Stealing mathematicians private chats and private research and taking credit for it would absolutely be par for the course.

Enterprises are well aware of it and are fully on board. You didn't think every corporation in America has an OpenAI subscription because the models were good, did you?

The whole reason they have subs is to train them on YOUR WORKFLOWS lol

They are not training a whole model in a matter of days

pretraining is months but they can totally fine tune in a few days

They where working on the problem for a year using codex.

They don't need to train a whole model. They can feed it new information and fine tune it.

Models are very obviously continuously updated.

Model editing to remove PII that slipped through, all sorts of things of that sort.

Why would this be implausible?

ChatGPT user sessions were found publicly exposed to the internet not too long ago. Moreover, OpenAI has continued to play a hype-marketing game by revealing how their models keep breaking out of the sandbox.

Conspiracy minded thinking is not helpful, but why should OpenAI be granted the benefit of the doubt here after being caught doing underhanded/negligent shit on several previous occasion?

>ChatGPT user sessions were found publicly exposed to the internet not too long ago

Do you mean publicly shared chats were able to be accessed by the public? That's the point of the feature.

Couldn't the Enterprise have a different fine print?

Reply on news.ycombinator.com