Random Thread

[alice]

@malloc: what do you think all those unemployed actors are going to be doing once AI takes over? Sitting around grumbling about the good old days?

Presumably they'll have to find another line of work, undoubtedly less appealing with lower pay and considerable competition from everyone else losing their jobs to AI.

And not, perhaps, getting together in large numbers and and protesting? Or making movies themselves, with proper people, to show how much better it is?

[zompist]

And not, perhaps, getting together in large numbers and and protesting? Or making movies themselves, with proper people, to show how much better it is?

Now now, one movie made using AI by a small UK firm obviously shows that AI has completely won and no other movies exist; also that the Ruling Class has determined that human movies must be eliminated.

psst don't tell malloc that the Care Bears are not actually played by bears

[Man in Space]

psst don't tell malloc that the Care Bears are not actually played by bears

…does this mean Mr. Ed didn’t really speak?

[zompist]

psst don't tell malloc that the Care Bears are not actually played by bears

…does this mean Mr. Ed didn’t really speak?

The Enrichment Center reminds you that a horse cannot talk. In the event that it does talk the Enrichment Center asks you to ignore its advice.

[malloc]

Now now, one movie made using AI by a small UK firm obviously shows that AI has completely won and no other movies exist; also that the Ruling Class has determined that human movies must be eliminated.

All technological innovations work this way, though. The first airplanes could barely fly and someone living in 1905 might well have dismissed them as gimmicks with no practical application. One decade later and airplanes were dog-fighting over the battlefields of the Great War. Obviously that doesn't guarantee that AI will follow the same trajectory but it does show how barely functional prototypes can evolve into genuinely useful and even revolutionary technologies. Merely five years ago, the very notion of AI actors was pure science fiction with no plausible mechanism for making it work. Today it has reached the prototype stage with an AI generated actress landing film gigs. Another five years and AI actors might well become commonplace.

[rotting bones]

Many inventions also fail. They keep trying until something sticks. What's unique about this attempt is the level of investment in a product when no path to profitability is known.

[rotting bones]

One thing LLMs can't do is humor. I recently found a nice German example. There is this little joke:

Treffen sich zwei Jäger im Wald. Beide tot.
'Two huntsmen meet/hit each other in the forest. Both dead.'

This works because of the dual meaning of treffen. An LLM asked to reproduce this joke produced:

Begegnen sich zwei Jäger im Wald. Beide tot.

This version doesn't work because begegnen can only mean 'to meet', not 'to hit'. There is nothing funny about this.

I think this is more a case of LLMs being worse in German than English. They are even worse in Bengali than German. My guess is that they have less training data to generalize from. In English, their sense of humor isn't amazing, but it's not always terrible either.

On the other hand, an LLM doesn't really understand anything. It's the furthest a parrot could go if it had brains the size of data centers while still not knowing what it's talking about.

[malloc]

Many inventions also fail. They keep trying until something sticks. What's unique about this attempt is the level of investment in a product when no path to profitability is known.

The path to profitability seems quite obvious in this case. Imagine making a blockbuster movie without the expense of A-list actors. You would save millions by not having to pay salaries and residuals and so forth.

[rotting bones]

Many inventions also fail. They keep trying until something sticks. What's unique about this attempt is the level of investment in a product when no path to profitability is known.

The path to profitability seems quite obvious in this case. Imagine making a blockbuster movie without the expense of A-list actors. You would save millions by not having to pay salaries and residuals and so forth.

Why would the AI count as an "A-list" actor?

[zompist]

Many inventions also fail. They keep trying until something sticks. What's unique about this attempt is the level of investment in a product when no path to profitability is known.

The path to profitability seems quite obvious in this case. Imagine making a blockbuster movie without the expense of A-list actors. You would save millions by not having to pay salaries and residuals and so forth.

You could save those same millions by paying a no-name actor instead. Why do you think they don't?

[Man in Space]

I weep for the day that someone accuses the Marx Brothers’ Duck Soup is accused of being AI. Half the fun in the last act of the movie is the changing hats every scene.

[malloc]

You could save those same millions by paying a no-name actor instead. Why do you think they don't?

So you would say the problem is branding as much as technical capability. They can replicate the skills involved in acting but they can't create a persona that people would pay top-dollar to watch.

[Travis B.]

You could save those same millions by paying a no-name actor instead. Why do you think they don't?

So you would say the problem is branding as much as technical capability. They can replicate the skills involved in acting but they can't create a persona that people would pay top-dollar to watch.

And, of course, part of the whole reason for putting A-list actors in your movies is marketing ─ people go out and spend money to see A-list actors because they are personalities. Contrast this with an AI which is not going to get people to spend money to see them, except perhaps as a novelty to see just how lame it is for once.

[zompist]

You could save those same millions by paying a no-name actor instead. Why do you think they don't?

So you would say the problem is branding as much as technical capability. They can replicate the skills involved in acting but they can't create a persona that people would pay top-dollar to watch.

They can't do either.

[Ares Land]

Many inventions also fail. They keep trying until something sticks. What's unique about this attempt is the level of investment in a product when no path to profitability is known.

I think the AI industry, having seen three technological revolutions: personal computing, the Internet and cellphones; all three took over the world very quickly and were (relatively) easy money-making machines. They entirely expect to replicate the same success. There have been other attempts to produce The Next Big Thing (with VR, for instance).

[Raphael]

You could save those same millions by paying a no-name actor instead. Why do you think they don't?

So you would say the problem is branding as much as technical capability. They can replicate the skills involved in acting but they can't create a persona that people would pay top-dollar to watch.

And, of course, part of the whole reason for putting A-list actors in your movies is marketing ─ people go out and spend money to see A-list actors because they are personalities.

I guess from the point of view of movie makers, the ideal situation would be to hire a no name who is then catapulted into star-hood by the movie or show in question. But of course that doesn't happen that often, and you can't really plan it intentionally.

[zompist]

I guess from the point of view of movie makers, the ideal situation would be to hire a no name who is then catapulted into star-hood by the movie or show in question. But of course that doesn't happen that often, and you can't really plan it intentionally.

Well, you can, kind of. The old studio system was good at this: sign up a steady stream of stars, put them in a load of movies, and see who takes off.

In a sense, that's how directors still work. You find some promising foreign director and hand them a superhero movie. If they're good, they get more big movies; if not, they probably didn't do much damage.

[Travis B.]

Many inventions also fail. They keep trying until something sticks. What's unique about this attempt is the level of investment in a product when no path to profitability is known.

I think the AI industry, having seen three technological revolutions: personal computing, the Internet and cellphones; all three took over the world very quickly and were (relatively) easy money-making machines. They entirely expect to replicate the same success. There have been other attempts to produce The Next Big Thing (with VR, for instance).

Not everything has succeeded, though, like your example of VR, which has largely been a dud. Even AI, despite its 'success', has not been successful at actually making money. We will see how far investors will tolerate their money being burnt with no real return...

[Raphael]

We will see how far investors will tolerate their money being burnt with no real return...

"If the price of my AI investments stays high, it might not be so bad if I get no return on them. If the price falls, I'll lose what I paid for them. So, I'll do what I can to ensure the price stays high!"

[malloc]

So you would say the problem is branding as much as technical capability. They can replicate the skills involved in acting but they can't create a persona that people would pay top-dollar to watch.

They can't do either.

Then what is this project using? Clearly they must have some working model for simulating actors or they couldn't attempt something like this in the first place. Otherwise they would have to embarrass themselves by making a movie they know is rubbish.

More generally, one can hardly dispute that AI has made some astounding progress over the past few years. We have models that can churn out convincing potboilers at least, enough that professional writers are genuinely worried. You can reasonably argue that AI has not yet mastered true art, that it cannot replicate the innovative genius of the avant-garde. Indeed it seems difficult to advance beyond cliché and mindless entertainment without some form of consciousness and deliberate intent which AI currently lacks. Even so, that still means that novice arts still finding their voice and honestly anyone who isn't revolutionizing the field now face ever stiffening competition for commissions.