Emad Mostaque resigned as CEO of Stability AI

Post by BhavdeepSethi » Sat, Mar 23, 2024, 7:40 AM UTC

alfiedotwtf wrote: post #168

manquer wrote: post #88

Earlier quoted context omitted.

For a founder maybe , definitely not for employees . AI startups need not an insignificant amount of startup capital , you cannot just spend weekends to build like you would a saas app . Model training is expensive so only wealthy individuals can even consider this route Companies like that have no oversight or control mechanisms when management inevitably goes down crazy paths, also without external valuations optio…

Sometimes you need to say fuck the money, I’ve already got enough, and I just want to do what I enjoy. It may not be an ideal model for HN but damn not everything in life is about grinding, P/E ratios, and vesting schedules

Post by renewiltord » Sat, Mar 23, 2024, 7:45 AM UTC

llm_trw wrote: post #138

AnthonyMouse wrote: post #135

Earlier quoted context omitted.

The released image generation models run on consumer GPUs. Even the big LLMs will run on a $3500 Mac with reasonable performance, and the CPU of a dirt cheap machine if you don't care about it being slow, which is sometimes important and sometimes isn't. Also, things like this are in the works: https://news.ycombinator.com/item?id=39794864 Which will put the system RAM of the new 24-channel PC servers in range of the…

The `big' AI models are trillion parameter models. The medium sized models like GPT3 and Grok are 185b and 314b respectively. There is no way for _anyone_ to run these on a sub $50k machine in 2024, and even if you can the token generation speed on CPU is under 0.1 tokens per second.

Post by ramraj07 » Sat, Mar 23, 2024, 7:48 AM UTC

benreesman wrote: post #167

ImprobableTruth wrote: post #21

Really sad to see. Emad pivoting to posting crypto nonsense on twitter made me think the writing is on the wall for Stability, but I still didn't expect it so soon. I expect they'll pivot to closed models and then fade away.

I mean, is AI a less sketchy space in 2024 than crypto/blockchain in 2024? Two or three years ago sure, I guess, but today? The drama around OpenAI is well documented, there are multiple lawsuits and an SEC investigation at least in embryo, Karpathy bounced and Ilya's harder to spot than Kate Middleton (edit: please see below edit in regards to this tasteless quip). NVIDIA is pushing the Dutch East India Company by s…

Post by HlessClaudesman » Sat, Mar 23, 2024, 7:50 AM UTC

famouswaffles wrote: post #45

Satam wrote: post #18

The underlying reason is that current AI businesses and models are failing to capture any significant economic value. We're going to get there, but it will take some more work. It won't be decades, but a few more years would be helpful

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'…

Post by fauigerzigerk » Sat, Mar 23, 2024, 7:51 AM UTC

AnthonyMouse wrote: post #144

llm_trw wrote: post #138

Earlier quoted context omitted.

The `big' AI models are trillion parameter models. The medium sized models like GPT3 and Grok are 185b and 314b respectively. There is no way for _anyone_ to run these on a sub $50k machine in 2024, and even if you can the token generation speed on CPU is under 0.1 tokens per second.

You can get registered DDR4 for ~$1/GB. A trillion parameter model in FP16 would need ~2TB. Servers that support that much are actually cheap (~$200), the main cost would be the ~$2000 in memory itself. That is going to be dog slow but you can certainly do it if you want to and it doesn't cost $50,000.