Hacker News Board Hacker News, browsed like a classic message board
Live data from Hacker News

Make-A-Video: AI system that generates videos from text

makeavideo.studio

Post reply on HN
Print view View original

391–399 of 399 posts

Re: Make-A-Video: AI system that generates videos from text

#391

Post by yazaddaruvala »

sdenton4 wrote: post #353
maxov wrote: post #179

Earlier quoted context omitted.

I love this! But after trying it a few times I got this result :). So fascinating. https://thismoviedoesnotexist.org/movie/the-terminator Brings up the age-old question of how much the learning in these models is just memorization. Though in cases like these it’s hard to tell.

Given Hollywood's proclivity to remake everything on a 20-year cycle, it seems completely appropriate to get a 2023 Terminator reboot in a 1920's style.

It also seems completely appropriate for The Terminator to be written, directed, and acted by an AI.

Re: Make-A-Video: AI system that generates videos from text

#392

Post by firstcommentyo »

pfalke wrote: post #199
maxov wrote: post #179

Earlier quoted context omitted.

I love this! But after trying it a few times I got this result :). So fascinating. https://thismoviedoesnotexist.org/movie/the-terminator Brings up the age-old question of how much the learning in these models is just memorization. Though in cases like these it’s hard to tell.

Here’s another Terminator one: https://thismoviedoesnotexist.org/movie/terminator-war-of-th...

Common, this is genius. Plot reads: in 2154 a soldier from the future is send back in time to the present.

Why send a soldier from the present to the past when you can also have a soldier from the future send to the past!

Re: Make-A-Video: AI system that generates videos from text

#393

Post by treis »

yreg wrote: post #350
treis wrote: post #294

Earlier quoted context omitted.

IMHO this particular avenue is a dead end. It's an extraordinarily impressive dead end but it's clear that there's no real understanding here. Look at this video of the knight riding a horse: > https://makeavideo.studio/assets/A_knight_riding_on_a_horse_... The horse's face is all wrong The gait is wrong The interface with the ground & hooves is wrong The knight's upper body doesn't match with the lower and they're n…

Of course there "is no understanding here", but yet it's not all wrong. Somehow it did move the horse's legs roughly correctly (using the proper joints and all), somehow the cape is moving roughly as it should through the air and the knight's body absorbs the force of stomping on the ground… It doesn't seem that the fundamental inability to understand what is going on in the scene is stopping models of this kind to e…

I guess it comes down to how much wiggle room is in "roughly". If you watch the video closely the horse briefly gets a 5th leg when the front left one moves the first time. And yes the legs and joints are sort of right but they don't match up with the direction the horse is moving and wouldn't work in the real world.

It's superficially close but when you look at details they're all slightly off. To wit:

> the knight's body absorbs the force of stomping on the ground…

But the Knight doesn't have any way to see through the helmet.

Re: Make-A-Video: AI system that generates videos from text

#394

Post by agentwiggles »

extragood wrote: post #382
slhomme wrote: post #121

As an owner of a Video Production studio, this kind of tech is blowing my mind and makes me equally excited and scared. I can see how we could incorporate such tools in our workflows, and at the same time I'm worried it'll be used to spam the internet with thousands and thousands of souless generated videos, making it even harder to look through the noise. A fun related experiment, I thought it was fun to see what ki…

Sleeping Beauty actually looks really good: https://thismoviedoesnotexist.org/movie/sleeping-beauty Edit: Worst (best?) tone clash: https://thismoviedoesnotexist.org/movie/stalked-by-a-friend

That second one is seriously hysterical

Re: Make-A-Video: AI system that generates videos from text

#395

Post by kurxig »

soperj wrote: post #190
maxov wrote: post #179

Earlier quoted context omitted.

I love this! But after trying it a few times I got this result :). So fascinating. https://thismoviedoesnotexist.org/movie/the-terminator Brings up the age-old question of how much the learning in these models is just memorization. Though in cases like these it’s hard to tell.

Yeah, just got: https://thismoviedoesnotexist.org/movie/the-legend-of-zelda-... It's crazy that it just made up those names...

Since the generated output is so close to the training data, the model is probably overfitted and trained with too few data...

Re: Make-A-Video: AI system that generates videos from text

#396

Post by yreg »

treis wrote: post #393
yreg wrote: post #350

Earlier quoted context omitted.

Of course there "is no understanding here", but yet it's not all wrong. Somehow it did move the horse's legs roughly correctly (using the proper joints and all), somehow the cape is moving roughly as it should through the air and the knight's body absorbs the force of stomping on the ground… It doesn't seem that the fundamental inability to understand what is going on in the scene is stopping models of this kind to e…

I guess it comes down to how much wiggle room is in "roughly". If you watch the video closely the horse briefly gets a 5th leg when the front left one moves the first time. And yes the legs and joints are sort of right but they don't match up with the direction the horse is moving and wouldn't work in the real world. It's superficially close but when you look at details they're all slightly off. To wit: > the knight'…

I would say that the wiggle room is smaller then I would expect it to be. Don't you agree?

And I was surprised about how small the wiggle room was the first time I interacted with GPT in text or saw the first images from DALL-E, since I too expected them to be (severly) limited by not understanding what's actually respresented in the input/output.

With new versions the wiggle room shrinks further. So I guess the question is, whether it will be able to shrink enough to be satisfactory. We will see…

Re: Make-A-Video: AI system that generates videos from text

#398

Post by Nevermark »

trention wrote: post #322
Nevermark wrote: post #312

Earlier quoted context omitted.

Porn companies, Facebook, Uber, ... these are just the latest companies, in a long line, for whom huge lawsuits and fortune-sized settlements are simply a line item on the expense sheet.

Are you somehow under the impression that, if a government decides to impose a fine that will financially kill the company once and for all, it is incapable of it? Incidentally, Uber has been banned in the entire country where I live for 8+ years now.

Of course. And that kind of action would make corporations sit up.

But what a government can do, and what its politicians will do are very different things. The political backlash for going nuclear on corporations would be massive and sustained. Not even good acting corporations want that kind of thing happening.

Regarding, Uber and your country: your point makes my point stronger.

A whole country is now off limits to Uber, but on the whole, across all the countries, their aggressive behaviors has been a win for them anyway. Just another cost of business.

Re: Make-A-Video: AI system that generates videos from text

#399

Post by cutups »

slhomme wrote: post #121

As an owner of a Video Production studio, this kind of tech is blowing my mind and makes me equally excited and scared. I can see how we could incorporate such tools in our workflows, and at the same time I'm worried it'll be used to spam the internet with thousands and thousands of souless generated videos, making it even harder to look through the noise. A fun related experiment, I thought it was fun to see what ki…

I was thinking about something like this site, but also taking some randomized existing plot outlines to generate specific stills from each part of the plot. Might require isolating character archetypes too?

Great work with this as is!

Post reply on HN