AI/LLMs Where is the "Ideas Guy Revolution" of Creativity and Productivity?

[doctorbadwolf]

Because the types of barriers AI removes are not the ones preventing an author from writing infinite jest...

If there is a huge increase in how many people with good or great ideas can execute those ideas, there should be an increase in good or great works being made.

There isn't any such increase.

If you want to posit that it is simply because the people claiming that ai is allowing them to make things they couldnt otherwise and is thus a good thing just...don't have the creativity or interesting ideas, we probably agree. I don't buy the claim in the first place.

But the op is just asking that question. Why no massive increase in new creative works that are actually interesting and creative?

No one is claiming that ai will write great novels, they do claim often that people who could potentially have the next great novel in their head are being allowed to write novels when they couldnt before. (A claim i think is malarkey)

 

[The Firebird]

If there is a huge increase in how many people with good or great ideas can execute those ideas, there should be an increase in good or great works being made.

There isn't any such increase.

There has been an increase in production, but not for things like novels. Programming, various bureaucratic tasks...the stuff ai is actually good at. I gave several examples above.

No one is claiming that ai will write great novels, they do claim often that people who could potentially have the next great novel in their head are being allowed to write novels when they couldnt before. (A claim i think is malarkey)

I've not encountered this claim. Probably there is someone out there who believes it. But phrasing it as such seems to avoid the core, and the strongest parts, of the pro AI position.

 

[doctorbadwolf]

There has been an increase in production, but not for things like novels. Programming, various bureaucratic tasks...the stuff ai is actually good at. I gave several examples above.

A reasonable argument can be made thst the only thing llms are reliably good at is faking conversation or content. Every other task is debatable at best.

I've not encountered this claim. Probably there is someone out there who believes it. But phrasing it as such seems to avoid the core, and the strongest parts, of the pro AI position.

I find that...extremely surprising. As i said upthread, it comes up in literally every debate i have ever seen about the merits of ai vs the ethical concerns around it.

 

[The Firebird]

A reasonable argument can be made thst the only thing llms are reliably good at is faking conversation or content. Every other task is debatable at best.

That was true a year ago. But it is no longer a reasonable argument. Cybersecurity, for example, has demonstrated impact.

I find that...extremely surprising. As i said upthread, it comes up in literally every debate i have ever seen about the merits of ai vs the ethical concerns around it.

You see people specifically saying AI is going to write great novels? Then you are seeing different discussions than I am; and I'd suggest that the pro AI people holding that position do not have a good understanding of AI or its capabilities.

 

[SableWyvern]

A reasonable argument can be made thst the only thing llms are reliably good at is faking conversation or content. Every other task is debatable at best.

I'm not sure I've encountered anyone in the IT sphere who denies that AIs are great at writing code. They certainly still need oversight, and a senior developer needs to check and refine (and there are big questions about how you create new senior developers when the AI replaces the juniors), so anyone saying they will completely replace skilled developers is probably wrong, but anyone saying they're not good tools for assisting in coding is almost certainly in deep denial.

 

[Tea Cozy]

I keep seeing talk that AI augments people's creativity and makes game creation and creative work more accessible to a broad audience.

We've had heavily subsidized OpenAI and Google models for a while now. Companies are basically giving away this marvelous and extremely expensive technology away for free to capture market share. It's been a while. These models can crank out thousands of pages per minute. Hundreds of picture perfect 4K art that put whoever WotC commissions for original art to shame. AI creators had the technological capacity to crank out 10^12 copies of Infinite Jest at this point.

Yet, where are those magnum opii? Where is the big AI american novel? Where is the Claude coded indie video game to break sales records? Where is the next big AI graphic novel? Where is the OSR iteration to fix AD&D for good?

Where are they?

In the US, works created only by text prompt are not eligible for copyright, so I imagine that has a lot to do with it. There’s also the fact that a lot of the tech is not quite ready for prime time (soon but not yet), but anyway….

Hybrid human/AI work can be eligible for US copyright. Last year, CBS’ Gayle King interviewed a young woman who took her poems and made them into songs that found fans.

Earlier that year, CBS’ Dave Malkoff did a segment with Rick Beato where they used Claude and Suno to make a song. Both of them were unnerved by how quickly they were able to churn out a serviceable pop song. “You can see the moment Beato’s heart breaks” one internet comment said. The song really isn’t bad.

The difference is that their song can’t be copyrighted. CBS just uploaded it to YouTube for anyone to do whatever with. They ended up doing a follow-up on what people did.

A reasonable argument can be made thst the only thing llms are reliably good at is faking conversation or content. Every other task is debatable at best.

Its usefulness in coding shifted recently, and AI’s applications in high-level mathematics have been in the news lately.

The AI Revolution in Math Has Arrived | Quanta Magazine

AI is being used to prove new results at a rapid pace. Mathematicians think this is just the beginning.

www.quantamagazine.org

 

[Retros_x]

Well, I conceded at both the very start and the finish of my post that these sorts of claims are going to be seen from people trying to sell you AI and AI services. That's not general "talk" though, that's "sales pitches".

Yes, but it OP implied exactly that, that was my point. AI did not got sold directly to the users, but via the buzz/hype. The general hype "talk" was the sale pitch.

 

[doctorbadwolf]

That was true a year ago. But it is no longer a reasonable argument. Cybersecurity, for example, has demonstrated impact.

We will see. It is a fundementally and unavoidably extremely flawed technology.

You see people specifically saying AI is going to write great novels? Then you are seeing different discussions than I am; and I'd suggest that the pro AI people holding that position do not have a good understanding of AI or its capabilities.

We literally just went over this. No one is saying that the claim is that ai will write great novels.

Do you actually not read what people write before replying?

 

[Umbran]

To build on that (that is, not arguing with you) I keep seeing two claims:

1. Everything AI does is crap and it always will be.
2. Oh no! AI is going to take all our jobs!

I don't see how both of those things can be true, unless most people don't care about quality, which I don't believe.

Um... "most people" don't determine if you get a job.

Corporate America is filled with executives who care more about the bottom line than quality.

Our collected experience of "enbleepification" comes from that basic fact.

 

[The Firebird]

We literally just went over this. No one is saying that the claim is that ai will write great novels.

Do you actually not read what people write before replying?

You've lost me, and this is now more heated than I have any interest in.