this post was submitted on 28 Jun 2025
226 points (91.9% liked)
Fuck AI
3341 readers
1413 users here now
"We did it, Patrick! We made a technological breakthrough!"
A place for all those who loathe AI to discuss things, post articles, and ridicule the AI hype. Proud supporter of working people. And proud booer of SXSW 2024.
founded 1 year ago
MODERATORS
you are viewing a single comment's thread
view the rest of the comments
view the rest of the comments
I have a question...
LLMs and generative AI are really only good at a subset of things, and we use them outside those things too often. But one thing they're extremely good at is identifying similarities in text.
I don't think AI as a technology is a problem, simply how we're using it, much like a knife.
If there was an art search engine that indexed every piece of human-made art that could be found and allowed you to search for it in natural language, would you use it? This would be different from a search engine in that natural language allows for more clarity of context and emphasis.
A lot of the joy of viewing and experiencing art, for me, comes from growth. Your tastes grow as you explore art, manually. Your horizons expand as you see art, including art that you don't like, and it can lead you to developing a real sense of your own taste and identity.
To show this even in microcosm: I've got dozens of books containing all sorts of drawn or painted pieces, or photographs, but one of my favourites is The Art of Heraldry. Flicking through the pages you'll see hundreds of different heraldic crests and coats of arms, crowns, all sorts. There's stuff in there I think is thoroughly ugly, but that explored an idea I'd never thought of, or showed a new way to include an element that I'd never seen before.
If I could sit down and type exactly what I wanted into an AI and have it return the exact thing I'm looking for, why would my tastes ever evolve? I would miss out on so much human effort and creativity. I wouldn't grow.
That is a damn good point.
It's tricky because sometimes you need to find something specific for good reason, and other times, exploring would do you good.
In any case, very helpful perspective. Maybe it's good I'd never have the time to build something like this anyway.
I wonder... this is slightly off topic, but fun to think out loud. In Korea, search super sucks. Most content is shared via blogs, and back links aren't a thing. All bloggers, and even the blogging platform, actively discourage any kind of copying. In the West, we solved this with sharing and back links, but that hasn't caught on here. But Koreans seem to (to me, at least) almost enjoy drudging through blog after blog trying to find what they're looking for. I always thought it's because they felt they had no alternative, but now I wonder if they have your perspective in mind.
In what context would that ever be useful?
Advertising
Maybe you want to buy some art. Maybe you saw some art but can't remember where it came from. Maybe you have a specific feeling and you'd like to see some art to go along with it. Maybe there's a comic with some joke that you want to find and share.
Basically, it fulfills the same role as a search engine.
The reason I propose it for art is because most people may have trouble describing art, and most art isn't indexed by its contents. An LLM can help with both of these problems.
It could be useful in other domains, too, but the required information isn't as readily available. I'll give two contrasting examples:
I imagine at least some people would find it useful. I don't know if it could be profitable with ads, but... I just like having thoughts like these. Have the knowledge but not the time, energy, or money to build something like this :/
Like all other LLMs, I would use it if I lacked the skill to use a better, more accurate, less halucination-prone search engine.
There's a great place for LLMs - as training wheels for non-experts.
But not everyone needs to use a carefully engineered command line interface, but those who do so get better results.
For about 80% of all problems, a halucination-prone human language interface will get the job done. This is lovely and opens up exciting opportunities for non-experts.
The problem is when people believe the lie that the human language interface to the computer gets better results for an expert. It still doesn't. It never will.
A rock that we taught to think will always be most powerful when operated directly using special "rock-that-we-taught-to-think" codes. Computers are still computers, and they are operated most powerfully and accurately by using computer languages. No abstraction has ever changed this.
Usually that doesn't matter anymore. But sometimes it still does.
And abstractions are still great. They make computers useful to more people.
Most domains already have specialized human computer interface languages. Experts in those domains already know those languages. Experts may find LLMs convenient, but LLMs do not meaningfully increase the capabilities of an expert.
But LLMs can help the rest of us, and the interface you purpose would be cool and useful, to the rest of us.