I have also appreciated this as an exercise in paying attention to small details in art. There's a lot of beauty in tiny details people would typically miss, but which make for a nice card back.

I'm not sure LLMs will soon able to solve the long tail of cases needing expert human invention any time soon, but I suspect with the right formal verification they'll rarely need the expert in the loop.

A leap forward for productivity, but not so great for the relevance of my skills.

A big part of went wrong, for me, was Meta's software.

Software jank is many times worse in VR and the software + headset discomfort rewired my brain into not enjoying the medium.

Before that I had spent dozens of hours in the DK2 and HoloLens and mostly enjoyed the experience.

Pre the metaverse wave I was building out a VR product, a 3D modeling tool, banking on the idea that the killer feature of VR would be 'worlds' crafted by users.

It became difficult to stay motivated working on a product in a medium that was very uncomfortable for me.

I'm leery of the term. It evokes this idea that a designer is just 'discovering' what the player wanted all along, but I think most of the time designers are creating new experiences the players never quite knew they wanted.

Asking "Does this fulfill the player's fantasy?" feels limiting.

I've thought about something like this in a cyberpunk setting where most of what you do is figure out how to navigate esoteric public transit, traverse neighborhoods, gather info, and figure out daily life.

Something that scares me is how so many more people want to pick and choose "facts" that feel the best. So many people just say "no [some alternate reality] is correct" despite the evidence, and it seems AI generated video may greater enable that.

For creative disciplines this means it will become important to have a public persona, and to help people understand the personalities and lives behind a work.

Maybe it's too complex, but yes it'd be more automatic.

I think it could get to the point where if just a few people with some credibility block someone it'd immediately take effect for most other reasonable people.

It's probably like a "recommended follows" algorithm but in reverse.

A thought: On BlueSky it's public who blocks who.

I think something useful would be an ability to say "I trust this person" and that defines a sort of graph of who trusts who.

Based on networks of trust if someone credible blocks an account it could instantly percolate to thousands.