Sad in a way on the same day this is posted https://news.ycombinator.com/item?id=44241549 (which for me is super inspiring). Mikeal would've done so much with the stronger and stronger AI out there (though perhaps questioning it of course in his own way). His cancer diaries was just one example of how he couldn't help always being a leader in what he did.
Hot take: based on how fast models are accelerating and replacing large parts of the development process (my two cents in https://x.com/TomHarada1/status/1926193211678023953), I think more and more you want to work backwards from a world where AI does 90% of things. Script kiddies : prompt engineering :: current engineering :: future of engineering. Either path makes sense - going deep in research or development. One is a kernel and one is the rest of the egg.
I think it’s more like AI empowers and 3x’s the creatives that learn to use it. In all fields where highly intelligent auto-complete is useful it replaces 2/3rds of who you need to hire. The key is to learn to use the tools. As it has always been with new tools like computers etc.
It's not just a new tool. It replaces the person using the tool, and we're all being conditioned into believing whatever AI churns out, which will never be anything more than average, is good enough.
That’s why those “creatives” who used llms to write their movie won 3x as many Oscars, and the ones who used them to write their tv show won 3x as many Emmys, and the ones who used them to write their music won 3x as many Grammys!
Why not call it Ardukid (if it's resurrected)? I.e., evolve from Gameboy and not bring in gender from the very beginning where it's irrelevant (for the next generation). Imagine if you're a young girl interested in tech. It's those little paper cuts that lead to a world with gender imbalances for no very good reason.
Indeed, you only have to simulate what one person perceives and where they explore, etc. One person's conscious experience in theory should be able to be fully simulated with very little information relative to the complexity of matter etc. You just have to simulate the senses and model the universe around a person.
And why not do that as an experiment? If science/experimentation is a useful thing - why not have the main reality and lots of individual experiments?
A curious observation. I was opposed to this article based on the title. Esp when I clicked on it and then saw it was the same author as "But what if I really want a faster horse?"
Because I assumed I knew what the argument was. I assumed this was another material ui doesn't look great -- while not appreciating the nuance of Google. Just as I assumed "what if I really want a faster horse" was actually about some anti-AI thing that just didn't appreciate how game-changing AI models are.
I pigeonholed the author based on my pattern matching of similar titles.
Then I clicked around https://rakhim.exotext.com/. Always curious about clean design. So clicked around on https://exotext.com/. Clicked on a few more blog posts of Rakim again and thought maybe they were ok. I was still a bit in pattern matching mode (stereotyping perhaps) assuming the author was like other people on Bluesky etc. Perhaps reactionary etc.
Then somehow in my clicking I saw a thumbnail of the author and I was like - oh that guy looks like me.
It's messed up it took me to this point to get there - but at least I persisted in trying to understand where they were coming from. At that point though I started to click around more and more and actually read the articles and I realized I agreed with all of them. Part of my appreciation was that Rakim had created exotext.com etc.
Just a cautionary tale to not pattern match prematurely. Premature optimization...
Though building in guardrails to prevent premature optimization is an important hack. E.g., faces really do matter in terms of slowing people down and taking them more seriously I think. And just building in anti-premature-optimization "tread" or friction for lack of a better word. E.g., avoiding click bait titles that people might pattern match on. It's not the author's fault but it might be more successful that way. Cause - really insightful blogposts. I feel like a fool for dismissing them at first. And yet I hope they get wider reach via perhaps subverting the ways we pattern match (not that I or other people should but I think based on other comments and how people pattern match on me - yet another tech bro, etc. - it exists...)
Author here. I appreciate your honesty; this is indeed a curious observation. I, too, have found myself patter-matching personalities by titles of their works, design styles of their blogs, etc. I sometimes would even feel some sort of resistance when I started to discover that my patter-matching was off the mark.
reply