Andrej Karpathy says X is so toxic he hangs out on Anthropic Slack now, Elon Musk responds

Andrej Karpathy, the former OpenAI co-founder and Tesla AI director who recently joined Anthropic, says X has become so toxic he prefers to hang out on his new company's – Anthropic – Slack channel now. Here is what Elon Musk had to say about this.

In Short

• Andrej Karpathy says X has become more toxic than ever
• He blames recommendation algorithm for rewarding negativity
• Elon Musk agrees to him and says the algorithm needs a complete change

Twitter aka X is increasingly being criticised, in part for allegedly becoming toxic. Multiple user complaints suggest people are not happy about the X algorithm. The latest to join the chorus is Andrej Karpathy, the former OpenAI co-founder and Tesla AI director who recently joined Anthropic. In a recent post, Karpathy said X has become more toxic than ever before, blaming the platform's recommendation algorithm for encouraging hostile behaviour. Surprisingly, X owner Elon Musk couldn’t agree more.

In a recent thread on X discussing Anthropic's newly introduced Claude Tag workflow, Karpathy, who joined Anthropic earlier this year, argued that the new system represents a major shift in how people interact with AI. According to him, many users misunderstood his comments and dismissed the idea without reading beyond the headline.

In his original posts, Karpathy argued that Claude Tag is not simply another chatbot feature or a Slack bot. Instead, he described it as a new way of working with AI, where the assistant is integrated across tools, memory, compute environments and workflows, making it feel less like chatting with a bot and more like collaborating with a teammate.

However, his post quickly sparked debate on X. While some users agreed that AI is evolving beyond standalone chatbots into workplace assistants that can work across different tools, others mocked the comparison and questioned whether the feature was as amazing as Karpathy claimed.

Responding to the criticism, Karpathy said many people had reacted without reading the full explanation. "This is correct, I think a number of people on the tl (timeline) didn't read past the title and made inferences and comparisons that are just wrong," he wrote, adding that Claude Tag should not be viewed as "some crappy Slack bot."

The discussion soon shifted from AI to the state of conversations on X itself.

In a follow-up post, Karpathy wrote: "I've been on Twitter for almost 2 decades now so I can say with confidence that it has never been this toxic and Reddit-like. I think the alg (algorithm) actively encourages it and people get RL'd by it. Main reason I've been coming here less and posting less."

By "RL'd", Karpathy was referring to reinforcement learning, suggesting that users are effectively rewarded by X's recommendation algorithm for posting outrage, criticism and inflammatory content because such posts receive greater engagement and visibility.

Musk responded directly to the post with a brief reply, writing: "We need a complete overhaul of the algorithm."

Musk did not elaborate on the changes he wants to make to X's recommendation system, but his comment echoed his long-standing position that the platform's algorithms should evolve to improve the quality of conversations. Since acquiring Twitter in 2022 and rebranding it as X, Musk has repeatedly spoken about making the platform's recommendation systems more transparent and improving the overall user experience.

- Ends