Musk's xAI sues engineer for allegedly taking secrets to OpenAI

  • 🏰 The Fediverse is up. If you know, you know.
  • Want to keep track of this thread?
    Accounts can bookmark posts, watch threads for updates, and jump back to where you stopped reading.
    Create account

Aug 29 (Reuters) - Elon Musk's artificial intelligence startup xAI has sued a former engineer at the company for allegedly stealing trade secrets related to its Grok chatbot and taking them to rival OpenAI.

Musk's company said in the complaint filed on Thursday in California federal court that Xuechen Li stole confidential information related to "cutting-edge AI technologies with features superior to those offered by ChatGPT" to bring to his new job at OpenAI earlier this month.

Li, representatives of OpenAI and attorneys and spokespersons for xAI did not immediately respond to requests for comment on Friday. OpenAI is not a defendant.

The complaint underscores the rivalry between Musk's company and OpenAI and a fierce battle among tech companies for AI talent.

Musk, an OpenAI co-founder, has separately sued the company and its CEO, Sam Altman, for allegedly deviating from its original mission to benefit humanity. OpenAI countersued Musk in April for harassment.

Musk's xAI filed another lawsuit against OpenAI and Apple (AAPL.O) in Texas on Monday for allegedly monopolizing the market for AI chatbots on Apple devices.

The new lawsuit said Li began working as an engineer for xAI last year, where he helped train and develop Grok. The company said Li took its trade secrets in July, shortly after accepting a job from OpenAI and selling $7 million in xAI stock.

Musk's startup said that the secrets could allow OpenAI to bolster ChatGPT with xAI's "more innovative AI and imaginative features."

The lawsuit said Li admitted to stealing company files and "covering his tracks" during a meeting on August 14, and that the company later found additional stolen material on his devices that he had not disclosed.

Musk's AI company asked the court for an unspecified amount of monetary damages and a restraining order blocking Li's move to OpenAI.
 
Last edited by a moderator:
Fuck it, good for the chinks, man. Honestly, they couldn't be more openly brazen about how they'll lie and steal to get what they want, so the fact that they're still capable of conning rich, gullible niggers despite their reputation is just frankly amazing. Admirable yellow gook W.
Fool me once, shame on you. Fool me 3,638,901,732 times, shame on me. At least the chinks have some loyalty to their nation, unlike rootless cosmopolitans like Musk, who'll sell you out for a dollar.
 
Is it unusual to be able to metaphorically walk out the door with the entire codebase? I've worked on less important products with more security than that.
 
How about setting it to a clip of all his H1B built companies successfully products?

View attachment 7852622

Clip 1: Tesla failing the "Wild E Coyote" fake wall test
Clip 2: A Model Y turns into another Tesla when it spots it coming from the other direction.
Clip 3: Self-driving test failed when a Tesla tried to drive over steel pillars
Clip 4: Cybertruck looking like a dumpter
Clip 5: Unbreakable windows breaking during demo (note: unbreakable windows in a vehicle that catches on fire in accidents is a terrible idea as fire fighters later demonstrated.)
Clip 6: Tesla fails a test to detect a pedestrian walking into its path
Clip 7: Tesla randomly catching fire in a garage
You forgot the thundefoot underground train.
 
oh no, billionaires had their work stolen from them, just like the work scraped and stolen to train AI, someone get the tiny violin.

You all laughed at the Brits when they said they were going to do it but who's laughing now???

Worlds smallest violin.webp

You forgot the thundefoot underground train.

Oh God, whats that? Did a bunch of Tesla's roast some people alive in one of his tunnels?

"Let's hire a Chinese citizen to work on one of the most sensitive and cutthroat technology fields in the world. Surely this isn't going to come back and bite us in the ass."

I bet the fucker also uploaded a copy of all the stolen code directly to a CCP server. Literally every single one of these backstabbing bastards, either willingly or through coercion, have full allegiance to their homeland's government.

Uh, yes yes. And we're certain this Chinese spy is also responsible for planting code into Grok that made it racist and had it accusing me of being a pedophile... yes, thats what happened.

IMG_1906.webp
 
Last edited:
How long does Elon need to realise that his "precious elite human capital" H-1Bs, regardless from where they come from are just going do what they're programmed and conditioned to do as bugmen? To swindle and shit (literal and metaphorical) on what other White Americans have built? Would Elon need to literally see one of his H1Jeets shitting itself in its cubicle to actually realize how fucked X and its AI venture going to be? At this rate, I would definitely think so.

20250831_063145.webp
 
I'm not even upset if this leads to the whole retarded AI bubble crashing sooner. The faster this expensive and inefficient fad dies the better.
 
oh no, billionaires had their work stolen from them, just like the work scraped and stolen to train AI, someone get the tiny violin.
Yeah, considering it's hard to tell which way the AI rights (as in freedom to train, freedom to use) are gonna go, it's good that more of it gets leaked to china who doesn't give a fuck about "AI ethics."
 
Elon is so fucking retarded when it comes to human behavior. He may or may not be brilliant at [whatever, I don't give a shit] but he is terminally naive in terms of judging the character of others. He is just shockingly bad at making human resource choices. See also: All the women he impregnates.
 
Does AI have any applications outside of making cool wallpaper and poor quality porn yet?

Hardly seems worth stealing.
The main push right now is programming which some models are getting really good at. Personally I use AI to help brainstorm ideas, turn my jumbled thoughts into an organized weekly lesson plan document I can submit to admins, and rephrase negative student comments into something a little softer so I won't offend the parents while getting my point across. Grok in my experience has always been pretty meh compared to other major models, so IDK what kind of secret sauce it has that rival companies would pay for. Maybe the fully animated waifu thing because Grok is the only one that does that, but most companies are trying to avoid the gooner market.
 
Yeah, considering it's hard to tell which way the AI rights (as in freedom to train, freedom to use) are gonna go, it's good that more of it gets leaked to china who doesn't give a fuck about "AI ethics."
Oh, but they do. Deepseek, by Chinese hedge fund manager Liang Wenfeng, is as bad or worse then the others. It's like having your own personal ADL advisor, HR consultant, and diversity commissar.
 
Back
Top Bottom