AI is still a fad.
It certainly lets untrained people create stuff they probably should not
Why OpenAI's $157B valuation misreads AI's future (Oct 2024)
61–70 of 138 posts
Re: Why OpenAI's $157B valuation misreads AI's future (Oct 2024)
#62For about a month now I've been paying $20-$30/day to delegate the bulk of my coding to Sonnet. The agentic loop thats trained into it is just simply not matched by another other model. I can't admit to myself there's any open question as to if there is any long-term value. I expect within 2 years, this will seem like a non-controversial idea, and it won't bring in a ton of assumptions about the speaker. I have inves…
What tools are you using to delegate?
Re: Why OpenAI's $157B valuation misreads AI's future (Oct 2024)
#63Earlier quoted context omitted.
K, just keep this conversation in mind in a decade or two when you realize that your input is the product, not what you got out of it.
Oh yeah—tell that to my AI pipelines on my local compute.
Local pipelines are great and all for now but there’s practically no way those will be able to keep up with server based models long-term. If those are useful for you today, that’s great.
Re: Why OpenAI's $157B valuation misreads AI's future (Oct 2024)
#64We’ve just learned that it’s possible to do AI on less compute (deepseek). if OpenAI doesn’t scale and that’s the problem then I’d argue that in the long run, if you believe in their ability to do research, then the news this week is a very bullish sign. IMO the equivalent of moores law for AI (both on software and hardware development) is baked into the price, which doesn’t make the valuation all too crazy.
> We’ve just learned that it’s possible to do AI on less compute (deepseek).
There's a huge motte and bailey thing with DeepSeek conversation, where the bailey is "It only took $5.5 million!*" (* for exactly one training run for one of several models, at dirt-cheap per-hour spot prices for H100s) and the motte is all sorts of stuff.
Truth is one run for one model took 2048 GPUs fulltime for 2 months, and my experience with FAANG ML, that means it took 6 months part-time and another 1.5-2.5 runs went absolutely nowhere.
Re: Why OpenAI's $157B valuation misreads AI's future (Oct 2024)
#65We’ve just learned that it’s possible to do AI on less compute (deepseek). if OpenAI doesn’t scale and that’s the problem then I’d argue that in the long run, if you believe in their ability to do research, then the news this week is a very bullish sign. IMO the equivalent of moores law for AI (both on software and hardware development) is baked into the price, which doesn’t make the valuation all too crazy.
Does anyone know how Deepseek does it yet?
Re: Why OpenAI's $157B valuation misreads AI's future (Oct 2024)
#66A pretty good read that succinctly picks apart the realities of current AI businesses. Easily something I’d reference as a “primer” to someone that is more business-minded than technically-minded. One point I’ll agree on is his final one: that the true big players haven’t even been founded yet. Right now, the AI hype seems to still revolve around the dream of replacing humans with machines and still magically making…
Aren't there already a ton of startups doing finetunes for their local niche? Many aren't even "AI" companies - it's pretty easy to slap a finetune together if you enough data.
If you mean developing a model from scratch just for your niche - the bitter lesson is that scale is everything and that a finetune from an internet-scale model will outperform you easily.
Re: Why OpenAI's $157B valuation misreads AI's future (Oct 2024)
#67For about a month now I've been paying $20-$30/day to delegate the bulk of my coding to Sonnet. The agentic loop thats trained into it is just simply not matched by another other model. I can't admit to myself there's any open question as to if there is any long-term value. I expect within 2 years, this will seem like a non-controversial idea, and it won't bring in a ton of assumptions about the speaker. I have inves…
What tools are you using to delegate?
My own app, it's called Telosnex.
Unfortunately, the current available version doesn't have the agent stuff yet.
Hopefully in a week, realistically two.
I had the existing client app I've released-but-not-released-out-loud. Couple days before Christmas, for fun, I spent a couple hours wiring up the Anthropic Model Context Protocol filesystem server example. Within an hour it was clear this was special and I needed to get it out ASAP. Stunning stuff in action.
Re: Why OpenAI's $157B valuation misreads AI's future (Oct 2024)
#68Earlier quoted context omitted.
Oh yeah—tell that to my AI pipelines on my local compute.
Local pipelines are great and all for now but there’s practically no way those will be able to keep up with server based models long-term. If those are useful for you today, that’s great.
Right and those "server based models" can be deployed by me or my team, on clusters we own.
Re: Why OpenAI's $157B valuation misreads AI's future (Oct 2024)
#69For about a month now I've been paying $20-$30/day to delegate the bulk of my coding to Sonnet. The agentic loop thats trained into it is just simply not matched by another other model. I can't admit to myself there's any open question as to if there is any long-term value. I expect within 2 years, this will seem like a non-controversial idea, and it won't bring in a ton of assumptions about the speaker. I have inves…
What sort of coding? Are you maintaining any large projects?
Re: Why OpenAI's $157B valuation misreads AI's future (Oct 2024)
#70For about a month now I've been paying $20-$30/day to delegate the bulk of my coding to Sonnet. The agentic loop thats trained into it is just simply not matched by another other model. I can't admit to myself there's any open question as to if there is any long-term value. I expect within 2 years, this will seem like a non-controversial idea, and it won't bring in a ton of assumptions about the speaker. I have inves…
$20 to $30 a day? How?? I have been using Sonnet every day for a month and have spent a total of... $7 on OpenRouter (7/30 = 0.23 a day on average).
Granted, I ask it very specific questions that generate short answers (many of which are incorrect, btw), but still, it's difficult to imagine what kind of tasks, done by a single person, would generate such amounts?