Full Summary
This Thursday morning, Google DeepMind has launched DiffusionGemma, a new text-generation model that reaches up to 1,000 tokens per second on Nvidia GPU hardware. Both Yellow.com and AIBase confirm this model runs four times faster than previous Gemma models, using a diffusion-based architecture to generate entire text blocks simultaneously. The Register also notes this innovation could significantly change how AI generates text. Meanwhile, Anthropic has released Claude Fable 5, its most advanced AI model, which Digital Health News and The Economic Times confirm is the first publicly available model in its new Mythos-class family. This AI powerhouse boasts advanced coding abilities, performing tasks like a 50-million-line code migration in one day and designing drugs ten times faster than humans, as detailed by The Economic Times and pharmaphorum. However, it comes with high compute costs and significant token consumption. Digital Health News and The New Indian Express explain that Fable 5 includes strong safeguards, with requests in sensitive areas like cybersecurity or biology redirected to an older model, Claude Opus 4.8. Anthropic plans to alert users when this rerouting occurs. In other AI news, Apple has unveiled a redesigned Siri AI, integrated into upcoming iOS, iPadOS, and macOS versions. Let's Data Science and SSBCrack report that this new Siri references on-screen content and uses personal context across apps, built on Google's Gemini foundation model. Apple clarifies they use their own proprietary models refined with outputs from Google's Gemini. This surge in AI capabilities means your future interactions with technology, from your phone to enterprise software, will be significantly faster, more integrated, and potentially more costly for businesses.