Anthropic has weird vibes

Summary

The video critiques Anthropic's approach to developer relations and transparency, highlighting frustrations with restrictive practices, DMCA takedowns, and poor communication. The speaker compares Anthropic to OpenAI, praising the latter for better developer support despite their models' impressive capabilities. While acknowledging Anthropic’s technical achievements, the video stresses the problematic nature of their company culture and developer experience, arguing that their closed-off policies hinder collaboration and innovation. The speaker also reflects on the challenges faced when working with Anthropic's infrastructure, painting a picture of a company that is difficult to engage with despite its technological potential.

Takeaways

• 😀 The speaker expresses frustration with Anthropic's lack of transparency and support for developers.
• 😀 Anthropic's decision to DMCA open-source code has been criticized, and the speaker points out that this has happened multiple times.
• 😀 Despite the technical excellence of Anthropic’s models, their corporate practices make it difficult for developers to work with them.
• 😀 OpenAI is highlighted as a more developer-friendly company compared to Anthropic, with the speaker having positive experiences working with OpenAI.
• 😀 Anthropic has been accused of creating a toxic work culture that discourages collaboration and innovation.
• 😀 There is a suggestion that Anthropic’s recent innovations might not be enough to maintain its edge, as other models are catching up.
• 😀 The speaker believes that Anthropic should open-source more of their code and drop prices to repair their PR image.
• 😀 There is an emphasis on the importance of the developer experience, suggesting that no matter how advanced a model is, it won’t succeed without a strong developer community.
• 😀 The speaker contrasts their ease of communication with OpenAI, who actively reaches out, versus Anthropic’s poor responsiveness and difficulty in getting rate limit increases.
• 😀 The speaker admits that despite the good work done by some of the people at Anthropic, their corporate behavior has led to a general negative perception of the company.

Q & A

What is the main frustration of the speaker with Anthropic?

The speaker's main frustration with Anthropic lies in the company's lack of transparency, poor communication with developers, and restrictive practices, especially regarding the DMCA takedowns of code. The speaker also highlights difficulty in getting rate limits increased and working with Anthropic's infrastructure compared to other companies like OpenAI.

What was the issue with the code that Anthropic had taken down?

The code in question was taken down from npm, but the speaker's colleague, Dave, was able to retrieve the source code through the source map URL. The code itself was using the Ink animation system, and the concerns over 'secret sauce' in the system prompt turned out to be unfounded.

What does the speaker suggest would improve Anthropic's approach?

The speaker suggests that Anthropic should open-source their code, make their models available for open use, reduce prices, and focus on repairing their public image. These steps, according to the speaker, would help regain the trust of developers and users.

How does the speaker compare Anthropic with OpenAI?

The speaker contrasts Anthropic's communication and developer relations with OpenAI's approach, stating that OpenAI has been much easier to work with. They mention that OpenAI has been proactive in supporting developers, including offering early access and working through issues like rate limit increases without much trouble, unlike Anthropic.

What specific actions by Anthropic have caused tension with the speaker?

Actions like frequent DMCA takedowns, difficulty in getting rate limits increased, and the overall lack of transparency have contributed to the speaker's negative view of Anthropic. The speaker also expresses frustration over how difficult it is to get information from Anthropic compared to other companies.

What is the significance of the DMCA takedowns mentioned in the transcript?

The DMCA takedowns are significant because they illustrate Anthropic's tendency to remove code related to its models and restrict access, which has frustrated developers and led to accusations of the company being overly controlling. These actions are seen as damaging to the open-source community and innovation.

What is the speaker’s overall opinion of Anthropic despite their technical achievements?

Despite Anthropic's impressive research and technological advancements, the speaker feels the company's management practices are problematic. The speaker believes Anthropic's restrictive policies and poor developer relations overshadow their technical achievements and make them one of the worst companies to work with in the AI space.

How does the speaker feel about their interactions with Anthropic as a developer?

The speaker describes their experience working with Anthropic as miserable, particularly when it came to getting rate limits increased and dealing with the company's lack of responsiveness. This negative experience contrasts with the positive interactions they've had with companies like OpenAI.

What does the speaker believe is the key to Anthropic's success in the AI market?

The speaker acknowledges that Anthropic's success has been due to developers liking their models and building things around them. However, the speaker notes that this advantage is rapidly diminishing as other companies, like OpenAI, are improving and innovating in ways that surpass Anthropic.

What is the speaker’s final message regarding Anthropic?

The speaker's final message is one of disappointment, emphasizing that despite the impressive work being done by some smart individuals at Anthropic, the company’s overall practices and approach make them a difficult and frustrating partner for developers. The speaker expresses hope that they won’t lose access to Anthropic's models but anticipates potential backlash for voicing these concerns.