Victor's principles for making programming cognitively accessible (seeing state, time, flow) remain relevant with LLMs because they address how humans fundamentally understand systems, not just how we write code.

Bret's life's work is all about direct manipulation and building mental models. LLMs are a layer between humans and our work; they make our tools less direct. My guess is that his advice about LLMs would be to not use them.

Yes, his website shows it. Read his latest update and the footnote on the home page: https://worrydream.com/