Claude Fable 5

Key takeaways

Anthropic's Claude Fable 5 is a safe, guardrailed AI model aiming to prevent malicious use, with a cautious rollout to trusted organizations.
OpenAI has filed for an IPO, seeking to expand its capital while continuing to build its AI infrastructure and navigate competitive pressures.
Microsoft is actively addressing security vulnerabilities by removing compromised repositories on GitHub, highlighting ongoing cybersecurity challenges.
Apple's AI strategy appears focused on integrating advanced models into the iPhone, aiming to dominate consumer AI through intuitive, privacy-preserving features.
Major tech companies are investing heavily in AI, but also facing issues like security breaches and the need for responsible, safe deployment.

Summary

Anthropic's New AI Models: Claude Fable 5 and Mythos 5

Anthropic has launched Claude Fable 5, a Mythos-class AI model designed with safety features that prevent it from being used for cyberattacks. This model, along with Claude Mythos 5, targets trusted organizations and aims to provide advanced capabilities with built-in guardrails. The release has been cautious, partly due to concerns about potential misuse, and the models are being rolled out selectively, with protection mechanisms rerouting many sensitive queries to older, safer versions. These models boast significant improvements, including better performance in tasks like app creation and coding, and are priced at $10 per million input tokens and $50 per million output tokens, which is relatively high. Anthropic is working with the U.S. government and other research entities, indicating a focus on safe and responsible AI deployment.

OpenAI's Private IPO Filing and Industry Status

OpenAI has confidentially filed for an IPO but has not yet announced an exact timeline, citing strategic reasons for staying private. The company aims to raise capital through an IPO with assistance from banks like Goldman Sachs and Morgan Stanley. Despite the looming IPO, OpenAI continues to raise significant funds, with a recent valuation of around $852 billion. They are also planning employee share sales and are closely monitoring their expenditures, especially as they build infrastructure to support their expanding AI initiatives. Competition is heating up with Anthropic's recent funding and validation, and SpaceX's filings indicate a booming AI investment landscape.

Microsoft GitHub Repos Breach and Security Concerns

Microsoft temporarily disabled over 70 repositories on GitHub after hackers injected credential-stealing malware into these open-source projects, many related to Azure and development tools. The breach demonstrates ongoing vulnerabilities in the software supply chain, prompting emergency removal and investigation. Microsoft is communicating with affected customers and assessing the scope of the impact, highlighting the importance of cybersecurity in the rapidly evolving AI ecosystem.

Apple's AI Strategy and Consumer Impact

MG Siegler and Ben Thompson discuss Apple's recent AI developments, emphasizing that despite some perceptions of being behind in AI, Apple is poised to lead in consumer AI adoption through user-friendly, integrated features. Apple's focus on the iPhone as the central device for AI experiences—especially with the new Gemini-based foundation models—could enable widespread adoption, with Siri potentially becoming the primary AI assistant for most users. The discussion suggests that Apple’s approach leverages its product dominance, superior product instincts, and privacy-centric design to win at the consumer level, possibly outperforming tech giants like Google and Microsoft in real-world applications.

Show notes

Anthropic released Claude Fable 5, a guardrailed Mythos-class model, to the public and Mythos 5 to trusted partners. OpenAI confidentially filed for an IPO. Hackers injected credential-stealing malware into 70+ Microsoft GitHub repos, and Apple details its new Gemini-based foundation models.

Anthropic releases Claude Fable 5, a "safe" Mythos-class model it says can't be used for cyberattacks, to the public, and Claude Mythos 5 to trusted orgs (Wired)

OpenAI confidentially files for an IPO, says it has "not decided on timing yet", as "there are things we want to do that are likely easier as a private company" (CNBC)

Microsoft disabled 70+ of its repos on GitHub, including Azure-related tools like azure-functions-host, after hackers added credential-stealing malware to them (TechCrunch)

MG Siegler: after being left for dead in AI, Apple is set to win at the consumer level — the power of the default, superior product instincts, and no real competition (Spyglass)

Ben Thompson: the iPhone is the true core of Siri AI, and Apple is the only company positioned to work across apps with personal context — as long as it's not vaporware (Stratechery)