If you are using 5.6 Sol, you have to be aware of the it's notion to delete something and start something from scratch. What are the guardrails doing then?
- I tried Kimi K3, mostly for design. It feels overfitted to “radical” design, as if everything has to be a little psychedelic to avoid looking familiar. But underneath that, the usual AI-design tells are still there: tiny eyebrow labels above every heading, too much italics,
- Since software engineering (architecture, backend, etc) is mostly solved, I feel this is what the next biggest thing frontier models are going to compete on aggressively, design and taste. And the best thing is that design and taste is every evolving because good design and tasteBig news: Kimi-K3 by @Kimi_Moonshot is now #1 in the Frontend Code Arena with 1679 pts, surpassing Claude Fable 5. This is a 17-place jump from Kimi-k2.6 (#18 -> #1). In Frontend, Kimi-K3 ranked #1 in 6 of 7 domains: Brand & Marketing, Reference-Based Design, Data & Analytics,
- So cool.New Anthropic research: A global workspace in language models. Of everything happening in your brain right now, only a tiny fraction is consciously accessible—thoughts you can describe, hold in mind, and reason with. We found a strikingly similar divide inside Claude.
- Everyone's either fable-maxxing their features today or fable-maxxing the plan for features to give to Opus 4.8 or gpt-5.5 from tomorrow.left a few fable agents running over lunch while i went for a walk and got some vitamin d. came back to a load of completed work. can’t believe we’ve only got a few hours left of this. openai better release 5.6 soon so anthropic is forced to keep fable around.





