Live data from Hacker News

AI’s Instagram Problem

deeplearning.ai

71–80 of 99 posts

Re: AI’s Instagram Problem

#71
post #63

Earlier quoted context omitted.

> Human creations are far better still than AI ones. I have seen many examples of AI-generated art where it's beyond my taste level to see how it could be improved on by human hands. And stuff that I genuinely appreciated for what it is, divorced from how it was made. I am reminded that very early on with Midjourney, before this wave of AI gen art was wider known, there was a story of someone winning an art competiti…

I hear you but: > I have seen many examples of AI-generated art where it's beyond my taste level to see how it could be improved on by human hands This is my overriding point, beyond matters of taste. Everyone who talks about this stuff can't seem to talk about its value without reference to the hypothetical human artist or writer or whoever who didn't make it. The awe of just the fact it was made by AI can't be sepa…

> Everyone who talks about this stuff can't seem to talk about its value without reference to the hypothetical human artist or writer or whoever who didn't make it.

Consider that it's because that's the interesting areas of conversation to be had. The overwhelming majority of art I see all day does not warrant comment, and so it goes with AI art that is of a quality to replace that everyday art. The AI art that I do share and comment on tends to be works that are hilariously broken in some bizarre way. And I agree that that's possibly a novelty that will pass, but that would be due in part to AI gen improving to the level where it's consistently indistinguishable from artist-made art.

> I just don't think we have the ability to separate these things.

Perhaps not entirely, but it seems to be an overreach to claim the impact is necessarily debilitating. With my very modest amount of formal arts training, I approach analysis piece by piece. Composition, color, shape. Line and form, light and shadow, anatomy, texture. Subject. Abstraction and realism, style and tone. While not objective, it's easy to clinically assess the subjective quality achieved with an individual element, up to one's limit of taste.

Further, when I am knowingly looking at AI art I'm not doing it from a place of enthusiasm, but criticism. I am aware of a set of flaws that current art gen models are prone to include, and I actively look for them. Each one found speaks to the limitations of the technology, and the patterns suggest challenges that may be un-resolvable with consistency.

So yes, when I say that I genuinely appreciate an AI-made work for itself, I am indeed confident in the truth of that.

> The painting, and all the others like it, can only speak one thing: themselves as novel works.

Your concern with "statements" suggests that the area of art you are talking about is narrow, indeed. To clarify, I speak of the wide world of works made for primarily aesthetic goals. Well no, that's not entirely true. For the purposes of this discussion I've been speaking of static visual arts. But that ranges from museum oil paintings to product labeling, from the stunning to the mundane. That's all preamble to this blunt point: being visually pleasing is sufficient to provide worth. AI gen can provide that.

I am sympathetic however to what may be your position: what meaning can be learned from a work if no mind was behind it to intend meaning? If your appreciation of art requires the belief that you are part of communication between souls, then clearly knowing a work is authored by AI would render that work worthless in your eyes. And I can understand that, or some version of it. It does not apply to me, however. In a death-of-the-author way I care much more about the meaning I can find myself from a work, considering the meaning intended as optional spice.

Lastly, no: I'm aware of what I like and do not, thanks.

Re: AI’s Instagram Problem

#72

This is correct, but there's also some meta comment to be made that research has an "instragram problem". What he's talking about isn't really even AI so much as deep learning: there are lots of other branches that now get virtually no play relative to deep learning. And then there are all sorts of adjacent CS, image processing (when was the last time you saw someone talk about wavelets), language analysis, and other…

Why does it matter? We can focus on othery approaches after deep learning stops yielding so much progress...

Re: AI’s Instagram Problem

#73

I've definitely been feeling deflated. Every angle I've come up with... some project has narrowly shipped before me. And they gain so much popularity so quickly that mindshare per project becomes a power law overnight. It's like the first person to release is at the top of the App Store, and there's no unseating them; the feedback loop is already too powerful. It very much feels like first to release wins. And I swea…

It is not enough to be the first one to ship. I shipped a Mac app that uses whisper to transcribe audio before anyone else, I even implemented a dictation algorithm on top of it. It was my first app, I am not an ios developer. However, after a while, someone with a better track record in app development came up and made a similar stuff. He knows how to build an audience, and he knows how to market an app. As a result, smashed out the app I developed.

So I don't think being the first one is that much important, it is not also important to take care clean coding etc. Someone with better marketing skills probably would ruin every work you did.

Re: AI’s Instagram Problem

#75
The article is correct, but it comes from someone that is already on the AI field. LLMs will peak and pass to become mature, that's high-frequency signal. But there's also a low frequency component, a more fundamental signal if you like, that tells us that AI is definitely out of the academia --it's been for some years, it's only reaching general public now. It cannot be ignored.

Re: AI’s Instagram Problem

#76
This article is NOT entirely correct. From personal perspective to improve yourself and doing your own art knowing that no one will ever see it, and that you doing your art only for your own sake, then it is true, the other people art does not matter.

But it is completely opposite if your livelihood depends on it, your income or your business, all trades depends on money, and there is limited amount human attention out there equal to number of people online multiplied by number of minutes spent.

So, there is uneven spread, and the same reason why some artist thrive and many starve. Same for writers, and many other arts and sports. If you take tennis as example, you will see that only maybe 100 or so tennis players earn amount of money that is good for the comfortable life style. Others are struggling to get by.

Same in any arts, or creative work ... only very small will actually manage to live out of their art ...

Re: AI’s Instagram Problem

#77
post #73

I've definitely been feeling deflated. Every angle I've come up with... some project has narrowly shipped before me. And they gain so much popularity so quickly that mindshare per project becomes a power law overnight. It's like the first person to release is at the top of the App Store, and there's no unseating them; the feedback loop is already too powerful. It very much feels like first to release wins. And I swea…

It is not enough to be the first one to ship. I shipped a Mac app that uses whisper to transcribe audio before anyone else, I even implemented a dictation algorithm on top of it. It was my first app, I am not an ios developer. However, after a while, someone with a better track record in app development came up and made a similar stuff. He knows how to build an audience, and he knows how to market an app. As a result…

Did you build macwhisper?

Re: AI’s Instagram Problem

#78

I've definitely been feeling deflated. Every angle I've come up with... some project has narrowly shipped before me. And they gain so much popularity so quickly that mindshare per project becomes a power law overnight. It's like the first person to release is at the top of the App Store, and there's no unseating them; the feedback loop is already too powerful. It very much feels like first to release wins. And I swea…

I may be out of the loop, but I have seen very little apps with new AI tech get any traction. Or, I've seen them, tried for a few minutes, and then went on with my day. Most of these are just flash in a pan; a cool tech demo or some initial hype from an over-promising landing page, but not many viable businesses. Just someone downloading a model, fine-tuning it on some dataset and trying to pawn it off.

And if some project "narrowly shipped before you", how much value or moat is there really if we're talking multiple projects and they're so quick&easy to do? Sorry to be a bit harsh, I just don't get feeling deflated by others making the same MVP as you do if it's just a natural idea on top of some existing tech. Your focus on code quality is also moot, it's delivering value that counts, not the tech stack beneath.

Re: AI’s Instagram Problem

#79
post #34
post #12

Idk about AI having an Instagram problem, but I know Instagram has an AI problem. So many fake accounts with AI as actors and they send you chat messages trying to pretend to be real people. They even react to comments and can discern good from bad comments. At first it was interesting. Now it's just annoying. Then the Instagram algorithm, if you comment on coffee ads that you drink tea, you'll get tea ads. It will s…

IDK. Seems like usual social media that it’s all about who you follow.

No, I follow athletes related to cycling, swimming, skiing etc. What content does Instagram promote me? Female models dressed in cycling clothes with their zippers down. Maybe because their algorithm says that it's the "most trending cycle-tagged content at the moment". But it's not related to the professional cyclists I'm following at all. Same goes for almost all content/tags: there will be some half-nude version of it that gets promoted.

Re: AI’s Instagram Problem

#80

The article is correct, but it comes from someone that is already on the AI field. LLMs will peak and pass to become mature, that's high-frequency signal. But there's also a low frequency component, a more fundamental signal if you like, that tells us that AI is definitely out of the academia --it's been for some years, it's only reaching general public now. It cannot be ignored.

It has already reached the public a while ago, it just wasn't announced or marketed so heavily. Google for example has been using BERT to improve search results since 2019. If you did slightly more sophisticated english language queries back then, you've already been benefiting from LLMs directly.
Post reply on HN