Mark Zuckerberg criticized AI companies like OpenAI and Anthropic for restricting access to their models and promoting alarmist narratives about AI risks. The Meta founder argued for more open distribution of powerful AI systems.
Mark Zuckerberg has launched a public critique of closed-model AI developers, positioning Meta's open-source approach as the superior path forward.
Speaking to the Financial Times, the Meta founder took aim at companies keeping their most advanced AI systems proprietary, characterizing their strategy as unnecessarily restrictive. He specifically called out the "discourse so filled with doom" that permeates discussions around AI safety and capability, suggesting such rhetoric obscures more productive conversations about the technology's development.
Zuckerberg's comments target major competitors including OpenAI, maker of ChatGPT, and Anthropic, both of which have emphasized careful deployment of their largest models. These companies argue that restricted access allows for better safety monitoring and reduces risks of misuse.
The Meta founder defended the principle of distillation—a technique that compresses larger AI models into smaller, more efficient versions that can run locally on devices. This approach, central to Meta's AI strategy, enables broader distribution while maintaining performance. Zuckerberg framed this as democratizing access to AI capabilities.
Meta has already released Llama, its open-source large language model, with subsequent versions gaining significant adoption among researchers and developers. The company positions open distribution as beneficial for innovation, competition, and preventing monopolistic control over advanced AI systems.
The debate reflects a fundamental division in AI development strategy: whether powerful systems should remain controlled by their creators or distributed openly. OpenAI and Anthropic emphasize safety and deliberate scaling. Meta argues that openness drives faster progress and prevents any single entity from dominating the field.
Zuckerberg's comments come as Meta continues investing heavily in AI infrastructure and competing directly with OpenAI and Google for dominance in generative AI markets. His critique of "doom-filled" discourse suggests Meta views safety concerns as overblown—a position that could invite scrutiny from AI ethics researchers and policymakers focused on responsible development.
Google has created a dedicated section in its Play Store for third-party app stores, with Aptoide becoming the first available to US Android users. The shift follows legal pressure from Epic Games.
Daring Fireball retracted its criticism of Apple's App Store rejection of the Dark Hours app, determining the rejection was justified. The developer subsequently conceded the decision was correct.
A generation of billionaires enriched by artificial intelligence developments are pledging to donate substantial portions of their wealth. The commitments raise questions about the reliability of such promises.
X is discontinuing its revenue-sharing program after September 7, replacing it with a new Original Content Rewards program launching September 8. The shift marks another iteration of creator monetization under Elon Musk's leadership.