On Monday, Anthropic released Claude Fable 5, the first of a new class of AI models they call Mythos. I’d had my hands on it for barely two days when I recorded this, and that’s deliberate. Some weeks deserve a considered essay. This one deserved an honest first reaction.
Here’s what happened. I gave it one brief: review the full 10-month content plan for Path to AI Emergence, my senior leadership programme, against the published thinking of the 15 to 20 AI commentators I rate most highly. It worked autonomously for over 12 hours. It came back with section-by-section recommendations, eight working tools I can deploy to our learners, and supporting pedagogy across all ten months of the programme. I estimate that’s three to four weeks of my own work, and I say that as someone who used Claude Opus 4.8, the previous state of the art, every single day.
In the video I talk through what it did, where it still needs me, what it costs, and the question I can’t shake: once models like this are in everyday use, how does any leader’s hiring plan stay the same?
My honest take, 48 hours in: this is the biggest jump from one state-of-the-art model to the next that I’ve seen in the generative AI era. I don’t say that lightly.
Sources & citations
- Anthropic — Claude Fable 5 and Mythos 5 announcement (9 June 2026)
- Ethan Mollick — “What it feels like to work with Mythos” (One Useful Thing, 9 June 2026)
- Every — “Vibe Check: Fable 5 Is the Best Coding Model in the World” (Dan Shipper & Katie Parrott, 8 June 2026)
- Simon Willison — first impressions of Claude Fable 5 (9 June 2026)
