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

Launch HN: Eggnog (YC W24) – AI videos with consistent characters

news.ycombinator.com

Post reply on HN
Print view View original

11–20 of 59 posts

Re: Launch HN: Eggnog (YC W24) – AI videos with consistent characters

#12

Post by samplank2 »

nextworddev wrote: post #6

This is an important problem. Just curious though - how do you plan on competing against OpenAI etc? Or will you guys just compete on developer experience and integrations, and wrap around popular OSS / OpenAI models?

We're not going to compete on having the best model -- at least that's the plan. We'll wrap around open source models or whatever we can license. We will compete on having the best library of assets (characters, sets, sounds etc) for putting videos together. We like this plan because 1. most people don't like a blank canvas where they have to create everything from scratch, they'd rather be able to assemble things together 2. people like participating in trends, and they will be able to riff on a popular character super easily.

Re: Launch HN: Eggnog (YC W24) – AI videos with consistent characters

#13

Post by samplank2 »

carlio wrote: post #5

It takes too long to iterate on a character design. For more explanation: I've been playing around with stable diffusion on my laptop recently; I have a gtx 4070 with 8GB dedicated VRAM so it's not nothing. The main problem I have is that it takes a lot of iteration on a prompt to get what I want, at lower resolution and sampling steps, before I know that I'll get roughly what I want. I tried making a character in Eg…

Thanks, this is helpful feedback. We're definitely frustrated with how long it takes to load a character. We'll see what we can do to give a better sense of what the character will look like before the training job kicks off. We should be able to show some intermediate results.

Re: Launch HN: Eggnog (YC W24) – AI videos with consistent characters

#16

Post by millgrove »

sandis wrote: post #7

Prompt: A film scene of Young Woman 3 doing backflips https://streamable.com/8hm4vu

Love it. One of the things we're focused on is getting better anatomical movement out of existing models. These models are great with animations like waves, cars, trains, flames, etc., but they struggle with people outside of basic movements. We're optimistic about future models, and we still think there are interesting, funny, exciting stories you can tell with what's available today!

Re: Launch HN: Eggnog (YC W24) – AI videos with consistent characters

#17

Post by nextworddev »

samplank2 wrote: post #12
nextworddev wrote: post #6

This is an important problem. Just curious though - how do you plan on competing against OpenAI etc? Or will you guys just compete on developer experience and integrations, and wrap around popular OSS / OpenAI models?

We're not going to compete on having the best model -- at least that's the plan. We'll wrap around open source models or whatever we can license. We will compete on having the best library of assets (characters, sets, sounds etc) for putting videos together. We like this plan because 1. most people don't like a blank canvas where they have to create everything from scratch, they'd rather be able to assemble things to…

I see, I guess the Canva playbook could work here.

Re: Launch HN: Eggnog (YC W24) – AI videos with consistent characters

#20

Post by frankdenbow »

I use stable diffusion and LoRAs frequently so I'm wondering what you hope to offer beyond that? I would love to see something that develops 3d models from these generations, is that the area you're looking to explore? More thoughts here: https://www.youtube.com/watch?v=18eIfWb0wug
Post reply on HN