Just now! The AI bigshot deleted the company profile in seconds, and then left in a lightning-fast 68 days? The person responded: Fake news! But behind it, an open-source vs. closed-source battle has already begun.

After waking up, Karpathy deleted the line about Anthropic from his personal bio.

Netizens went into an uproar. He joined Anthropic just 68 days ago—this move looks too much like a prelude to running away. Someone posted screenshots, saying internal消息 indicated he submitted his resignation application a few hours ago and politely thanked the company. His email is also alleged to have been deactivated. Yet Wikipedia still lists: “Anthropic (2026–present)”. Neither side has responded officially.

The timing is too coincidental. Two days ago, Huang Renxun posted on X a joint open letter supporting open-weight AI. 25 institutions including Nvidia, Microsoft, and Meta signed it. OpenAI and Google later also joined—only Anthropic was absent. Karpathy is a flagship figure in the open-source world; netizens immediately started speculating: is he ashamed for his employer, and in a fit of pique deleted his bio?

Even more bizarre: an Anthropic employee personally jumped in to fire back, saying, “Why doesn’t Nvidia open-source CUDA? Why doesn’t Microsoft open-source Windows?” Wu Enda delivered a pointed shot: “Closed-source is your right, but you can’t both keep it closed and stop others from opening it up.” The comment section overwhelmingly berated that employee.

Karpathy’s 68 days at Anthropic have not been smooth. When it was announced on May 19, the entire internet was shaken—he was an OpenAI founding member and Tesla’s AI lead, yet he went to a rival company; the post racked up 27.7 million views. The mission was to build a new team and use Claude to accelerate pretraining research. After Claude 5 was released, it was frequently banned, with permissions being turned on and off. On June 23, Claude Tag was launched; he praised it as the “third major redesign” for large-model interactions, from the web chat box to personal agents to team collaboration. But then netizens began questioning: he used to teach people to handcraft models from scratch, so why is he now endorsing a closed-source product?

“Back then you taught us to build from scratch, and now you’re asking us to blindly trust the vendor?” This line stung many people. Karpathy has shifted from an educator to a company man. Huang Renxun’s open letter sparked the contradiction—Anthropic didn’t sign, Karpathy didn’t take a stance, yet he quietly deleted his bio.

Of course, this is just a social media bio edit. But the protagonist is Karpathy—he spans research, industry, open source, and education, and is a thought leader in the AI era. Everyone’s expectations of him go far beyond those of ordinary employees. So every evaluation, every silence, and even a line less in the bio will be magnified endlessly.

While everyone was busy watching the drama, Karpathy himself appeared: “Fake news like this is apparently circulating on Twitter. I haven’t resigned. If I really were to resign, I’d definitely post a 10-paragraph essay.”

The farce then temporarily subsided. But the core issue remains unchanged: the game between open source and closed source, and the conflict between Karpathy’s personal brand and a company stance—this kind of drama won’t disappear.

Follow me: Get more real-time analysis and insights into the crypto market! $BTC $ETH $SOL

#直通IPO第二期JerseyMikes # Longxing listed today,成交90.1 billion #ETH回到1900美元