“In 10 years, computers will be doing this a million times faster.” The head of Nvidia does not believe that there is a need to invest trillions of dollars in the production of chips for AI::Despite the fact that Nvidia is now almost the main beneficiary of the growing interest in AI, the head of the company, Jensen Huang, does not believe that
Oh well. The world is going to burn anyway. Fuck this shit hole we call earth
Honestly as someone who has watched the once-fanciful prefixes “giga” and “tera” enter common parlance, and saw kilobytes of RAM turn to gigabytes, it’s really hard for me to think what he’s saying is impossible.
Nobody is saying it won’t happen eventually. But a million times within the next decade, i.e. 4x better every year for 10 years?
This generation isn’t better than last generation by even close to that. Nevermind doing 4x for 10 years straight.
Even if he is accurate, specialist hardware will outperform generic hardware at what it is specialized for.
I remember a story sometime in the 00s about PCs finally getting to the point where they were as fast as one of the WWII code breaking computers (or something like that). It wasn’t because we backtracked in computer speeds after WWII, but because even that ancient hardware was able to get good performance when it was purpose-built, but it couldn’t do anything else and likely would have required a lot of work to adjust to a different kind of cypher scheme, if it could be adapted at all.
So GP compute might be a million times faster in a decade, but specialist AI chips might be a million times faster than that.
A hardware neural net might be able to eliminate memory latency by giving each neuron fast resisters to handle all their memory needs. If it doesn’t need to change connections, each connection could be hard wired. A GPU wouldn’t have a chance at keeping up no matter how wide that memory bus gets or how many channels it gets split into. It might even use way less power (though with the elimination of memory latency, it could go fast enough to use way more, too).
So a Cerebras wafer will be 10^6 faster for the same computation as now, for the same price, in just 10 years? Not after Moore scaling ended many years ago and neural hardware architecture has matured. You can sure go analog, but that’s not the same computation. And that’s the end of the line, not without true 3d integration.
It depends what you call AI.
True artificial intelligence likely requires quantum computing because there’s some quantum stuff happening our brains and probably the smartest living human (Sir Roger Penrose) thinks that’s where the secret to consciousness is hiding after spending the last couple decades investigating that after helping Hawking finish up Einstein’s work
If you just mean a chat bot that can pass the Turing test, then yeah we can just wait a decade instead of developing special tech for AI.
I mean, if we really develop artificial intelligence before we understand our own consciousness, we’re probably fucked anyways.
It’d be like somehow inventing a nuclear bomb before understanding what radiation was. We’d have no idea what we’re creating or what the consequences of flipping the switch would be.
Can we stop with this “not real AI” meme… it’s a painfully dull response at this point, why does the goal post have legs? Just because Penrose thinks quantum mumbo jumbo is needed doesn’t mean he is right, machine learning is completely outside his field of expertise.
Mate, I was using chatbots on AIM 24 years ago…
It wasn’t AI then, it’s not AI now.
The only reason to get super excited about current chatbots, is if you think they came out of nowhere and not something after decades of slow progression. There’s no reason to expect there to be a sudden huge jump to actual AI unless you don’t know the history.
People aren’t changing definitions on you…
Well, some people are, it’s just the ones telling you chatbots are AI.
They’re just lying to generate hype to get investor money. You’re a bystander that fell for it.
I completely agree on the idiotic consensus around the no-true-AI meme.
The goal post is practically mounted on wheels they’re having to move it so fast. Machine learning and complexity seems to be enough.
I think that ChatGPT represents a “deep blue” moment for AI. Finally, something fairly generally, that is at least some what competitive with humans. Hell chat gpt can probably play chess better than the average human too.
But what we’re waiting for is the “alpha go” moment of AI. The moment when the unconquerable is toppled. I expect it to happen in 2-3 years. I think we’ve got almost all we need from a theoretical side, and that the rest will be engineering.
I expect AI to be largely independent, to have agency indistinguishable from a humans, but to be better, faster and broader in its scope than most humans in their ability. It will still get beaten by the best of the best humans. It will still make weird, sideways mistakes that don’t seem like obvious mistakes to make to humans. But it will be generally better than most humans at most tasks.
Do you know if there are, or if there are plans for a “new” Turing test ?
Roger Penrose is a mathematician who made important contributions to theoretical physics in the 1960ies, for which he received a Nobel Prize. In later decades, he published speculative books on consciousness, quantum physics, and neurobiology. These ideas have been out there for about 30 years now but have not been able to convince scientists in general. Rather, they are generally considered implausible or outright contradicted by the evidence. Simply put: It’s wrong.
The idea that quantum physics plays a direct role in brain function is very much on the fringes of science.
No offense meant. I know these ideas are very important to many spiritual people, but I felt the casual reader should know that it is not important in science.
It requires 4X speed increase every year, production quality scale can’t provide even close to half of that, maybe 25%, then another 25% from design, and regarding increasing die sizes they are already close to the end. So the only way to get from 150% to 400% per year is by using multi chip designs, meaning they will have to use 2.5x more chips per year. so the multi chip package in 10 years will probably have to have almost 10,000 chips! All of them bleeding edge!!!
The H200 is estimated to cost $40K, the future 10 year chip will be more like $40 million. Or maybe more like impossible to achieve.
If chips = cpus, here, then I imagine that will hit a limit also (Amdahl’s law).
A chip is also called a die, it’s the piece cut out from the wafer, which is then packaged onto a chip package.
Since traditionally there were always 1 chip per chip package, the 2 words were used almost synonymously.
I this case it’s basically GPU chips, which AFAIK AMD has already figured out how to use in multi chip packages. Meaning one package contains multiple chips that work “almost” as well as a single chip of similar size.
The advantage of multichip packages are obvious, production costs are way lower because smaller dies causes lower percentage of flawed dies, and allows for better binning of higher end parts.
Additionally it allows designs of way more complex packages, than would be possible with monolithic chips. This is the reason AMD has been taking marketshare in server markets from Intel. Because Intel has not been able to match the multichip design AMD introduced with Epyc in 2016/17, which originally was 4 Ryzen chiplets/chips/dies packaged together as one big 32 core server chip. Where the biggest Intel could make was 28 cores.
But packaging almost 10000 GPU chips together is completely different, and I don’t think that will be relevant within 10 years.
Amdahls law however is part obvious and part bullshit. Everything your mind is able to do semi efficiently, can be multithreaded, it is very few things that can’t.
Amdahls law is basically irrelevant with regard to AI, as AI has a lot of patten recognition, and pattern recognition is perfect for multi threading.
Sorry I have doubts, because that would require a factor 4x increase every year for 10 years! 4x^10 = 1,048,576x
Considering they historically have had problems achieving just twice the speed per year, it does not seem likely.
Why does that make a difference? Compute for AI is build on the progress for compute first for GPU then for data center. They are similar in nature.
Yes they have exceeded 2x for AI for a while, but that has been achieved through exploding die size and cost, but even that won’t make a million times faster in 10 years possible, because they can’t increase die sizes any further.
Building an ASIC for purpose built computation is significantly faster than generic gpu compute cores. Like when ASICs were built for bitcoin mining/sha256 and a little 5 watt usb device could outperform the best GPUs
Yes, but usually we keep those 2 kinds of optimizations separate, only combining chip design and production process. Because if the software is optimized, the hardware isn’t really doing the same thing.
So yes AI speed may increase more than just the hardware, but for the most sophisticated systems, the tasks will be more complex, which may again slow the software down.
So I think they will never be able to achieve it even when considering software optimizations too. Just the latest Tesla cars boast about 4 times higher resolution cameras, that will require 4 times the processing power to process image recognition, which then will be more accurate, but relatively slower.
We are not where we want to be, and the systems of the future will clearly be more complex, and on the software are more likely to be slower than faster.
even software that does the same thing gets slower example: Microsoft Office, Amazon, the web in general, etc.
This isn’t necessarily about just hardware. Current ML architectures and inference engines are far from being at peak efficiency. Just last year we saw 20x speedups for llm inference on some hardware. “a million times” is obviously hyperpole though.
