Re:OpenAI, the Partition Principle, and Mathematics
by asey on Friday October 09, 2026 @12:07PM
(#50022579)
This is complete malarkey - OpenAI is _not_ "straight-up channeling ideas from user input"
Re:OpenAI, the Partition Principle, and Mathematics
by jsw97 on Friday October 09, 2026 @12:09PM
(#50022592)
Yeah that is my strong baseline prior. But it's colliding with what Ma reports. How do you reconcile those things?
Re:OpenAI, the Partition Principle, and Mathematics
by Robotbeat on Friday October 09, 2026 @12:31PM
(#50022954)
Impossible to know without either more information about the actual occurrence or a deep understanding of the person making the claims. Just as OpenAI could easily be acting improperly, researchers who (understandably) feel deeply attacked by their problem getting solved right from under them might not be the most unbiased, either.
Re:OpenAI, the Partition Principle, and Mathematics
by cameldrv on Friday October 09, 2026 @01:46PM
(#50024165)
AFAIK OpenAI has not directly responded to the claims. If they ignore it, I don’t see what choice you have except to assume the worst.
Re:OpenAI, the Partition Principle, and Mathematics
by jazzyjackson on Sunday October 11, 2026 @10:30PM
(#50050609)
They’re inhaling the plain text conversations of millions of people, why wouldn’t they have summarizers and anomaly detection sitting on top of conversation logs. What assurances does OAI provide that they are not reading your messages for any purpose? Don’t they “review conversations for safety” ? How would they do that if they couldn’t just query everything you ever said to the chatbot. If you can see it so can they.
But this is even worse, because there is no way that OpenAI "trained" on this data between September 8, 2026, the date Chenglong Ma uploaded the paper to ChatGPT; and October 6, 2026, the date that OpenAI released a paper with "identical proof strategy and specific choices of notation" (Ma). That's one month, that's not the timescale for model training.
So this implies _not_ that OpenAI is training on user input, in the conventional sense of adjusting weights; but rather that they are *straight-up channeling ideas from user input*, and with a very short lag. You would think there would be about a million controls to prevent this.
This is next-level alarming. I would be very interested in knowing whether Chenlong activated the privacy (do not train, etc) options in ChatGPT, and any other details of their setup (which plan, etc). Also, note that "do not train" might be, in a lawyerly sense, considered by OpenAI to be strictly about weights, and not covering "we hoover up your results and regurgitate them".