5 ms·
The thing is, the models do work. They add value to me each and every day, to a degree almost no other tech has done before. But that doesn't take away the fac
by TrackerFF 1y ago
The thing is, the models do work. They add value to me each and every day, to a degree almost no other tech has done before.
But that doesn't take away the fact that this is extremely expensive stuff (not for me, but for the companies pushing the envelope), far too expensive. And it is really taking its toll on other resources, like electricity.
- nitwit005 1y agoA bubble doesn't imply the product is bad. We all need homes, but there have definitely been housing bubbles.
- infecto 1y agoI think the thing that most of the folks miss is exactly that. AI is adding value and it’s here to stay. Absolutely the OpenAI story is looking concerning but I am not sure it is this doomsday AI winter.
- 369548684892826 1y agoIf revenue doesn't catch up with the cost of developing and running these huge LLMs then the only way to avoid an AI winter is to find a way to make them way cheaper to develop.
- infecto 1y agoInference costs are largely break even and profitable. Now yes there is absolutely questions on the training side. You could turn off the switch today though and most of your leading models could keep the lights on.
- algorithmmonkey 1y agoI keep seeing everyone guessing what the margin is on inference. You say that it's largely break even. We have this person in the thread claim 80% margin (https://news.ycombinator.com/item?id=45462442 https://news.ycombinator.com/item?id=45462442). These numbers are complete make believe.
- infecto 1y agoWhy would they be make believe? I am generalizing across multiple companies and simplifying it as much as possible as we won’t know all the cost buckets but we do know at current costs the bare inference cost of keeping the machines running is being covered.
- emilecantin 1y agoBoth can be true at the same time. Similar to the early days of the Internet, the dot-com bubble eventually popped, but the Internet (and dot-coms, for that matter) didn't go away. What people are saying is that this mad race to throw cash at anything that has "AI" in it will eventually stop, and what will remain are the useful things.
- infecto 1y agoNo the tone is generally that some mythic AI winter is going to happen because of current valuations and that AI is simply the current crypto grift.
- emilecantin 1y agoYeah, I saw "AI winter" mentioned elsewhere in the thread... IMO there is a real qualitative difference between AI and crypto in terms of the durable impact it's going to have on the world. Does that mean I've bought into the AI hype? Maybe. But I think the signs are there.
- dragonwriter 1y agoAI winters are a recurring phenomenon, not a myth, and, like the dotcom bust, involve a collapsing hype bubble, reductions in focussed speculative investment in the field, but the technologies that were big during the preceding hype cycle continuing to be important, and develop, though in the case of AI winters often they stop being thought of as AI and just get referred to with a name for the specific technology (often a different one than the main one they were known by in the hype cycle, e.g. “expert systems” from th blast hype cycle are largely “business rules engines” now.)
- tim333 1y agoGoogling, there seem to have been two AI winters, the first (late 1970s - early 1980s) when people first figured AI was overhyped, and the second (late 1980s - early 1990s) with the collapse of expert systems. I don't think we are about to get one now - more like AI spring leading to AGI summer.
- rossdavidh 1y agoNo question that neural networks can add value, as has been true for a long time. The question is, how much? The current expenditures indicate that the answer is, "trillions of dollars worth". If you call it "neural networks" or "large language models" or "machine learning", this would sound absurd. Only by calling it "AI" (with the implication that it will achieve general intelligence) does this sound, for most people, possibly correct.
- SoftTalker 1y agoBut are you paying for them commensurate with the value they create? That's the problem. Tons of money being invested/borrowed on technology with not nearly enough revenue to justify it. Same thing that happened in the late 1990s.
- vachina 1y agoAdd what value though, enable you to write more slop to power more slop that requires more slop?