Legit question: if users of LLMs are supposed to “verify the results”—and that disclaimer justifies the results being inaccurate—in a world where LLMs eventually taint/replace all the results, how will these users verify the results? Through what service or method is the verification of accuracy expected to occur?

“And is the Operating System in the room with us now?”

My favourite thing about #uxn is that every time I look at it, despite how remarkable it is, all I can think about is building my own mini compute space. It’s good at what it does, but for me its power is in inspiring the notion that little compute spaces informed by a small perspective are possible, and should be more common.

It’s just like Morrowind in that I could never get very far in the story without starting to build mods and my own games.

It’s an uncommon metric—useful and inspiring.

I want to build a web spec subset that’s essentially 8-bit websites. I feel like the idea of responsive design (beyond dark mode) hasn’t really considered allowing for layers of interaction that might alter the overall interface in a way that’s not tied to viewport size, device, connection speed, or browser.

Like a “low poly mode” might work in a game engine? The advantage here might be a rendering mode that’s creatively inventive, but efficiently rendering a site in a limited resource context.