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

Emad Mostaque resigned as CEO of Stability AI

stability.ai

Post reply on HN
Print view View original

241–250 of 429 posts

Re: Emad Mostaque resigned as CEO of Stability AI

#241

Post by michaelt »

llm_trw wrote: post #115
famouswaffles wrote: post #45

Earlier quoted context omitted.

Midjourney is the most popular discord channel by far with 19.5M+ members, $200M in revenue in 2023 with 0 external investments and only 40 employees. The problem has nothing to do with commercializing image gen AI and all to do with Emad/Stability having seemingly 0 sensible business plans. Seriously this seemed to be the plan: Step 1: Release SD for free Step 2: ??? Step 3: Profit The vast majority of users couldn'…

>The vast majority of users couldn't be bothered to take the steps necessary to get it running locally so I don't even think the open sourcing philosophy would have been a serious hurdle to wider commercial adoption. The more I think about the AI space the more I realize that open sourcing large models is pointless now. Until you can reasonably buy a rig to run the model there is simply no point in doing this. It's n…

Huge models are the best type to open source.

You get all the benefits of academics and open source folks pushing your model forward, and a vastly improved hiring pool.

But it doesn't stop you launching a commercial offering, because 99.99% of the world's population doesn't have 48GB+ of VRAM.

Re: Emad Mostaque resigned as CEO of Stability AI

#242

Post by khazhoux »

seanhunter wrote: post #207
khazhoux wrote: post #180

Earlier quoted context omitted.

As a counter-counter-point that gets rarely discussed on HN, VCs aren't taking as much of the pie as people think. In a 2-founder, 4-engineer company, it wouldn't be unusual to have equity be roughly: 20% investors 70% founders 2-3% employees (1% emp1, 1% emp2, 0.5% emp3, 0.25% emp4) 7% for future employees before next funding round

This is not a fair comparison because you are not taking into account liquidation preferences. Those investors don't have the same class of equity as everyone else. That doesn't matter in the case of lights out success but it matters a great deal in many other scenarios.

Sure. My point was that most employees think that VCs take 80+%, and especially the first few employees usually have no idea just how little equity they have compared to the founders.

Re: Emad Mostaque resigned as CEO of Stability AI

#244

Post by sigmoid10 »

mirekrusin wrote: post #230
sigmoid10 wrote: post #215

Earlier quoted context omitted.

That's wrong. What you described is data parallelism and it would indeed be very tricky to e.g. sync gradients across machines. But this is not the only method of training neural nets (transformers or any other kind) in parallel. If we'd like to train, say, a human brain complexity level model with 10^15 parameters, we'd need a model parallelism approach anyways. It introduces a bit of complexity since you need to ma…

Lack of data parallelism is implied by computation that is performed. You gradient descend on your state. Each step needs to work on up to date state otherwise you're computing gradient descend from state that doesn't exist anymore and your computed gradient descent delta is nonsensical if applied to the most recent state (it was calculated on old one, direction that your computation calculated is now wrong). You als…

Your gradient descent is an operation on a directed acyclic graph. The graph itself is stateless. You can do parts of the graph without needing to have access to the entire graph, particularly for transformers. In fact this is already done today for training and inference of large models. The transfer bottleneck is for currently used model sizes and architectures. There's nothing to stop you from building a model so complex that compute itself becomes the bottleneck rather than data transfer. Except its ultimate usability of course, as I already mentioned.

Re: Emad Mostaque resigned as CEO of Stability AI

#245

Post by polygamous_bat »

kalkr wrote: post #219
vishnugupta wrote: post #104

Earlier quoted context omitted.

"Decentralised" AI probably means something to do with crypto. Like mint this ocin to get access to that model? That's my guess.

decentralization != coinshit

https://en.wikipedia.org/wiki/Emad_Mostaque

Read his Wikipedia page and tell me he doesn’t sound like your run of the mill crypto scammer.

> He claims that he holds B.A. and M.A. degrees in mathematics and computer science from the University of Oxford.[7][8] However, according to him, he did not attend his graduation ceremony to receive his degrees, and therefore, he does not technically possess a BA or an MA.[7]

Re: Emad Mostaque resigned as CEO of Stability AI

#247

Post by FeepingCreature »

aerhardt wrote: post #187
Closi wrote: post #169

Earlier quoted context omitted.

I’ve come to the opposite conclusion personally - AI model inference requires burst compute, which particularly suits cloud deployment (for these sort of applications). And while AIs may become more compute-efficient in some respects, the tasks we ask AIs to do will grow larger and more complex. Sure you might get a good image locally but what about when the market moves to video? Sure chat GPT might give good respon…

Ok but yesterday I was on a plane coding and I wouldn’t have minded having GPT4 as it is today available to me.

Thanks to Starlink, planes should have good internet soon.

Re: Emad Mostaque resigned as CEO of Stability AI

#248

Post by ben_w »

brandall10 wrote: post #227
ben_w wrote: post #208

Earlier quoted context omitted.

Why would those celebs pay Stability any significant money for this, given they can get it for a one off payment of at most a few hundred dollars salary/opportunity cost by paying an intern to gather the images and feed it into the existing free tools for training a LoRA?

I think in this case the celebs are getting paid for using their likeness.

That sounds like the "lose money on every sale" philosophy of the first dot-com bubble, only without even the "but make it up in volume" second half.

Re: Emad Mostaque resigned as CEO of Stability AI

#249

Post by FeepingCreature »

ImprobableTruth wrote: post #72
ctrw wrote: post #32

Earlier quoted context omitted.

Think seti at home. Instead of wasting all the compute on bitcoin we pretrain fully open models which can run on people's hardware. A 120b ternary model is the most interesting thing in the world. No one can train one now because you need a billion dollar super computer.

Transmission speeds aren't fast enough for this, unless you crank up the batch size ridiculously high.

LoRA training/merging basically is "crank up the batch size ridiculously high" in a nutshell, right? What actually breaks when you do that?

Re: Emad Mostaque resigned as CEO of Stability AI

#250

Post by FeepingCreature »

stale2002 wrote: post #162
ShamelessC wrote: post #141

Earlier quoted context omitted.

> socials groups like e/acc Good riddance.

Hey, let me know when the doomers have built anything that has mattered because they choose to, and not because they are forced to my market forces. At least the open source AI people have code that you can use, freely without restriction. The doomers, on the other hand, don't do anything but try and fail to prevent other people from releasing useful stuff. But, in some sense I should be thanking the doomers because…

"Let me know when the people who think AI will destroy the world actually build cool AI toys for me to play with."

Yeah uh.

Post reply on HN