14 Comments
User's avatar
David Perlmutter's avatar

I don't want to sound like a Luddite here, but I'd rather make sure that a talented human artist could afford to keep working (even if I couldn't afford their rates) than generate instant art for my fiction via AI. And fortunately I have become an acquaintance of a few through social media....

Machines have always meant to be created as labor-saving devices, but the particular labor related to the arts and the humanities does not require them to succeed.

Andrew Smith's avatar

David, you don’t sound like a Luddite to me! I love that you’ve found some meaningful connections with artists. I think we can reconsider what patronage means, too- it’s not just in the realm of the ultra-wealthy like in the days of the Medici; this is more in our hands today. Many things need constant rethinking right now.

David Perlmutter's avatar

Humpty Dumpty is going to fall off the wall soon, and then, hopefully, the AI business will resolve into something more fiscally and ethically responsible. Just look at what happened to the Internet as a whole after the dot-com bubble burst…

Daniel Nest's avatar

We went from "enshittification" to "ensloppification" pretty quickly, eh?

Andrew Smith's avatar

One man’s slop is another man’s doodiepoop, as the old saying goes.

Andrew Sniderman 🕷️'s avatar

I'll make a prediction, cuz if I'm wrong so what?

genAI novelty will wear off and specialized systems like medical, design, code, etc where the value is evident will persist.

and - most importantly - ML will become more and more plumbing that make tech better and we don't see it

Andrew Smith's avatar

I think you’re 100% right about the plumbing. I don’t really have enough up-to-date knowledge to comment on whether gen AI or specialized systems will win, but it sounds very plausible. It’s even more likely if gen AI reaches any sort of plateau for a minute. I’m still waiting for that to happen, though.

Bill C. White's avatar

AI as a black box bothered us as researchers. We need to know what's happening for interpretation. So we wrote a paper on it:

https://scholar.google.com/citations?view_op=view_citation&hl=en&user=mb2m9ZEAAAAJ&citation_for_view=mb2m9ZEAAAAJ:u-x6o8ySG0sC

Andrew Smith's avatar

Nice, Bill! Also: holy crap at 772 citations.

Bill C. White's avatar

My best is a Lancet paper on lung cancer with a bunch of high level folks at Vanderbilt.

It used my AI software (genetic algorithm) to detect signal peaks in noisy mass spectrometry signals to see if those classify disease status. Some tough work, but it has done well, and it is HIGH IMPACT. Sometimes I wonder if these things ever actually affect the real world, but The Lancet can make that happen, like Nature and Science journals, the biggies. Seems like a dream now.

Andrew Smith's avatar

You've made the world better, dude. It's all cumulative and you can't always see the impact because of 2nd order effects and so on, but this is exactly how progress happens, man. That is extraordinary.

Bill C. White's avatar

In research I had to learn to be patient and not expect too much at first. Sometimes the delay is decades!

Rudy Fischmann's avatar

I kinda think we’re still pretty early in how we relate and use AI. This is like the honeymoon period where too many are just far too in love with the power and speed of AI. I mean, the potential is great, but there’s some excessive fawning on the quality when much of it is objectively meh. To me, the magic that will be developed more will be the raw speed and leveling of the field when it comes to access of doing big things. When it comes to high end quality and creativity? It’s lacking. Hopefully the opportunitists (I don’t wanna say scammers) and big industry don’t bury us in mediocrity before it plays out properly.