3d ago

Elon Musk testifies xAI partly distilled OpenAI models

1955.9K0614859.3K

Elon Musk testified in the OpenAI lawsuit that xAI partly distilled OpenAI models to train its AI systems. Responding to OpenAI counsel, Musk confirmed xAI used distillation techniques on OpenAI technology and incorporates it in development. He noted AI companies generally distill from one another. The admission emerged during ongoing litigation between Musk and OpenAI, corroborated by journalists including New York Times' Mike Isaac.

Original post

MB#28@MILES_BRUNDAGE @MIKEISAAC

rat king 🐀@MIKEISAAC

Musk also affirms xAI has used OpenAI technology to improve xAI. wowie woo

4:04 PM · Apr 30, 2026

AI 1000 · 13 actions

QUOTEC🤗#111@CLEMENTDELANGUE @MIKEISAACWhat people call "distillation" is a super common practice (you use other models to benchmark your model, to evaluate your inputs or to add a little bit to your datasets) that in my opinion should be covered by fair use (just like using public data is), especially when the resulting models are open-source as it benefits all and helps break out monopolies that are strongly forming in AI!
What people call "distillation" is a super common practice (you use other models to benchmark your model, to evaluate your inputs or to add a little bit to your datasets) that in my opinion should be covered by fair use (just like using public data is), especially when the resulting models are open-source as it benefits all and helps break out monopolies that are strongly forming in AI!
Quotingrat king 🐀@MIKEISAAC
it is quite significant that Musk admitted on the stand that xAI is distilling OpenAI models to train xAI, and that it is using OpenAI's technology to build xAI!
QUOTEC🤗#111@CLEMENTDELANGUE @MTSLIVEWhat people call "distillation" is a super common practice (you use other models to benchmark your model, to evaluate your inputs or to add a little bit to your datasets) that in my opinion should be covered by fair use (just like using public data is), especially when the resulting models are open-source as it benefits all and helps break out monopolies that are strongly forming in AI
What people call "distillation" is a super common practice (you use other models to benchmark your model, to evaluate your inputs or to add a little bit to your datasets) that in my opinion should be covered by fair use (just like using public data is), especially when the resulting models are open-source as it benefits all and helps break out monopolies that are strongly forming in AI
QuotingMTS@MTSLIVE
LIVE TRIAL UPDATE: OpenAI's counsel asked Musk whether xAI has ever "distilled" technology from OpenAI. Musk: "Generally AI companies distill other AI companies." "Is that a yes?" Savitt asked. Musk: "Partly."
QUOTEC🤗#111@CLEMENTDELANGUE @DZHANG50I think the expression is “pulling the ladder”! All labs trained their models by distilling (at the very least distilling the web) which allowed them to become the fastest growing businesses in the history of humanity and now that they have armies of lawyers and lobbyists, they are trying to prevent others from doing the same thing.
I think the expression is “pulling the ladder”! All labs trained their models by distilling (at the very least distilling the web) which allowed them to become the fastest growing businesses in the history of humanity and now that they have armies of lawyers and lobbyists, they are trying to prevent others from doing the same thing.
QuotingDan Zhang @ ICLR@DZHANG50
does anyone else remember how Jacob Devlin went to the media about how Google was supposedly distilling from ShareGPT (it was not) and the resulting blowback? How times have changed 😅 https://futurism.com/the-byte/google-denies-bard-openai https://twitter.com/MTSlive/status/2049886679876632724
QUOTENL#190@NATOLAMBERT @MTSLIVEDistillation is largely an industry standard and not just something done by Chinese labs targeting OpenAI/Anthropic. Many American companies also distill Chinese (open) models.
Distillation is largely an industry standard and not just something done by Chinese labs targeting OpenAI/Anthropic. Many American companies also distill Chinese (open) models.
QuotingMTS@MTSLIVE
LIVE TRIAL UPDATE: OpenAI's counsel asked Musk whether xAI has ever "distilled" technology from OpenAI. Musk: "Generally AI companies distill other AI companies." "Is that a yes?" Savitt asked. Musk: "Partly."
QUOTEKP#265@KELSEYTUOC @MIKEISAACI think OpenAI did basically betray its commitments to its early investors but it's so absurd to have Musk being the one suing when he 1) didn't seem to care until it affected him personally and 2) is now running a fully for-profit competitor.
I think OpenAI did basically betray its commitments to its early investors but it's so absurd to have Musk being the one suing when he 1) didn't seem to care until it affected him personally and 2) is now running a fully for-profit competitor.
Quotingrat king 🐀@MIKEISAAC
it is quite significant that Musk admitted on the stand that xAI is distilling OpenAI models to train xAI, and that it is using OpenAI's technology to build xAI!
QUOTET(#586@TEORTAXESTEX @MTSLIVEBro forgot to distill the Chinese distilling ways first 😭
QUOTEAM#595@ANDREWMAYNE @MIKEISAAC👀
QUOTEAM#595@ANDREWMAYNE Concerning https://twitter.com/techcrunch/status/2049913247638380796
QUOTEPW#971@PETERWILDEFORD @MIKEISAAC👀
REPLYMB#28@MILES_BRUNDAGE @MIKEISAAC@MikeIsaac Noted is a bit generous (vs. claimed) IMO - complex topic but I doubt the big three do what xAI did IIUC
@MikeIsaac Noted is a bit generous (vs. claimed) IMO - complex topic but I doubt the big three do what xAI did IIUC
Replying torat king 🐀@MIKEISAAC
i am getting a lot of questions about this so i just want to make clear that it was a brief exchange and OAI’s lawyer did not seem to want to pursue it further after musk affirmed the practice. (musk also noted that it was a common practice among the labs) https://twitter.com/mikeisaac/status/2049883094535356454
REPLYDW#232@DEANWBALL @NATOLAMBERT@natolambert @ClementDelangue I think a multibillion dollar company distilling probably crosses a line into being straight up wrong
@natolambert @ClementDelangue I think a multibillion dollar company distilling probably crosses a line into being straight up wrong
Replying toNathan Lambert@NATOLAMBERT
Distillation is largely an industry standard and not just something done by Chinese labs targeting OpenAI/Anthropic. Many American companies also distill Chinese (open) models. https://twitter.com/mtslive/status/2049886679876632724
REPOSTC🤗#111@CLEMENTDELANGUE @DZHANG50does anyone else remember how Jacob Devlin went to the media about how Google was supposedly distilling from ShareGPT (it was not) and the resulting blowback? How times have changed 😅 https://futurism.com/the-byte/google-denies-bard-openai https://twitter.com/MTSlive/status/2049886679876632724
REPOSTC🤗#111@CLEMENTDELANGUE @WONGMJANELast week: White House condemns Chinese AI labs for alleged distillation attack This week: Elon admits under oath that xAI distilled OpenAI models to train its models 🫣 https://x.com/wongmjane/status/2049963585112351089/photo/1 https://twitter.com/mikeisaac/status/2049883094535356454