171–180 of 218 posts
I don't see how nvidia come out of this stronger their huge customers will be able to produce ASICs hat will be faster and cheaper to operate than their GPUs jensen has to be the luckiest man in the world, first crypto, now "AI"
> The first web cycle burned through dot-com exuberance and left behind Google, Amazon, eBay, and PayPal: the hardy survivors of Web 1.0. The next cycle, driven by social and mobile, burned again in 2008–2009, clearing the underbrush for Facebook, Airbnb, Uber, and the offspring of Y Combinator. Both fires followed the same pattern: excessive growth, sudden correction, then renaissance. I note your AI missed the cryp…
I'm excited for the AI wildfire to come and engulf these AI-written thinkpieces. At this point I'd prefer a set of bullet points over having to sift through more "it's not X (emdash) it's Y" pestilence.
Even if bubble burst will be massive, the slop factories will not stop with using it, because it's one of the use cases LLMs are just good at
I think this article is nearing the truth in the future of AI. I think the avoidance of claiming it is a bubble is a good sign, but saying it's an AI wildfire is still hyperbole. The idea that inference will drive compute demand is not what I experience because inference is a much easier problem than training. The training of an AI (LLM) is especially demanding and if and when we complete that, inference will be a pi…
Earlier quoted context omitted.
That is a very optimistic scaling assumption - and would almost certainly require substantial in space infrastructure (large scale lunar and asteroid mining and refining, lunar mass drivers, at least solid core nuclear drives) before you can even thin building all the necessary radiators and structural mass.
Starlink (collectively) already has 10 times the solar power of ISS, with present-day launch capabilities. If Starship works out (not guaranteed) launch cost should drop to ~$100/kg, which would enable very large constellations. Musk is planning for 1 megaton/year of satellites, each with 100kW, yielding about 100GW per year.[1] He thinks they can do that in 4 years, but adjusting for Elon-time, it's probably no less…
So about 2.5 MW of solar potential ? That's indeed quite impressive, but for serious compute a lot of energy needs cooling will eat into that.
Earlier quoted context omitted.
California gold rush is practically the textbook example of a mania. Gold, tulips, real estate, rail, witch hunts, satanic panic.. We shouldn't be worried because "this time it is different". "AI" is none of those things. It is "totally different this time". "AI" is going to do all the work for us, we are all going to get UBI then at some point as it grows the "ASI" is going to either figure out how to grant us immor…
> "AI" is going to do all the work for us, we are all going to get UBI then at some point as it grows the "ASI" I don't see those companies promising that. They are bragging about being able to replace jobs for millions people and their CEOs are also simultaneously taking pretty dismissive attitude toward lesser people. If is more of "we will replace millions of people of eff you, billionaire class will become more p…
If energy is indeed the limiting factor here, then maybe the companies building space-based compute (in which energy scales linearly) will remain after the wildfire. The key is for them to build before the money runs out--I'm not sure they will have enough time.
Earlier quoted context omitted.
If they are profitable on inference currently, they will be able to recover. People in SWE will continue paying for current SOTA models even if better models are not achieved.
Alternatively, latecomers will make use of newly cheaply available compute (from the firesales of failing companies) to produce models that match their quality, while having to invest only a fraction of what the first wave had to, allowing them to push the price below the cost floor of the first wave and making them go under.
See also: the big deepseek smear campaign.
Earlier quoted context omitted.
A million different people: I've used AI in X way and it helped me. You: No one knows anything AI helps with. Yeah, okay, if you ignore everything every user says then it is indeed a mystery.
How much astroturfing is happening online? These companies certainly have the funds to do so on a wide scale The only thing I trust about these right now is my own experience
Earlier quoted context omitted.
That's a cool dream, but my question is: is it happening? Out of the things you listed the only ones that seem plausible are translation team and data entry team, though even there, I'd want humans to deslop the output.
You are not capable of telling the difference between human translated and AI translated communication.