Karpathy slams X's toxicity algorithm, Musk vows complete overhaul
OpenAI co-founder says X is more toxic than ever after 20 years on the platform.
Andrej Karpathy, renowned AI researcher and co-founder of OpenAI who recently joined Anthropic, publicly criticized X's algorithm for fostering toxicity after his thread about Anthropic's new Claude Tag feature devolved into mockery. Karpathy noted that after nearly 20 years on the platform, it has 'never been this toxic and Reddit-like,' attributing the decline to an algorithm that 'RL'd' users into posting outrage—a reference to reinforcement learning driving engagement. He stated that he now posts less and prefers working in Anthropic's Slack.
Elon Musk, owner of X and Karpathy's former boss at Tesla, responded with a terse agreement: 'We need a complete overhaul of the algorithm.' Musk offered no specifics on timeline or metrics, but the promise aligns with his long-standing goal of making X's recommendation systems more transparent. Since acquiring Twitter in 2022 and rebranding it as X, Musk has repeatedly pledged to improve feed quality. Whether this exchange will lead to actual changes or become another viral moment the algorithm rewards remains uncertain.
- Karpathy said X's algorithm uses reinforcement learning to promote outrage, making the platform more toxic than ever in 20 years.
- Musk agreed to a complete algorithm overhaul but provided no timeline or concrete details.
- The controversy started when critics mocked Karpathy's explanation of Anthropic's Claude Tag, a persistent AI workforce tool for Slack.
Why It Matters
If Musk follows through, it could reshape how X curates content and reduce toxicity for 500M+ users.