A recent Reddit post, titled "Dario Is Scared," suggests that Anthropic is reacting to the increasing popularity of open-source large language models (LLMs).

Adoption of Open Source Models

The post highlights how, on the OpenRouter platform, open-source models like Kimi K2.5 and Minimax M2.5 are surpassing the popularity of Anthropic's proprietary models, such as Claude. This shift is particularly evident in the context of AI agents, an area expected to experience strong future growth.

Criticism of Anthropic

The author of the post also criticizes Anthropic's practices, accusing it of using web crawlers that ignore the robots.txt file, training its models on copyrighted material, and attempting to limit the spread of open-source models. These actions, according to the author, are motivated by a desire to monopolize artificial intelligence and the power associated with it.