4 ms·
The method used seems like it would work as well on bigger brains. The amount of data may mean we have to wait for Moore's Law to keep improving things for a w
by BurningFrog 2y ago
The method used seems like it would work as well on bigger brains.
The amount of data may mean we have to wait for Moore's Law to keep improving things for a while though.
- tsimionescu 2y agoThe method used required 3 million manual human corrections. Even if Moore's Law actually still meant anything for compute power, this is still many orders of magnitude from scaling to a human brain.
- ninetyninenine 2y agoMoores law ended.
- ben_w 2y agoDepends which of the many similar but subtly different things with that name was meant. In this context, what matters is "how many operations can I get done for a dollar?", and that's still very much improving very fast, albeit not quite as fast as before.
- ninetyninenine 2y agoIt applied to transistor density and it’s over. Its completely and utterly true and it’s agreed upon by experts. https://cap.csail.mit.edu/death-moores-law-what-it-means-and-what-might-fill-gap-going-forward#:~:text=THE%20REALITY%3A%20MOORE'S%20LAW%20IS%20OVER&text=Although%20miniaturization%20is%20still%20happening,two%20years%20has%20been%20broken https://cap.csail.mit.edu/death-moores-law-what-it-means-and.... I’m not making this stuff up.
- ben_w 2y agoThe original formulation was "The complexity for minimum component costs has increased at a rate of roughly a factor of two per year", which stopped being true almost immediately and very soon got mixed up with "performance doubles every 18 months". Dennard scaling is long dead, as is the clock frequency race; but features are still slowly getting smaller (your own link says so), as is energy consumption per operation. The latter, J/op, is the critical issue for big data centres. Brains are obviously better than transistors at this, and IIRC by that measure transistors are still getting roughly twice as good every 2.6 years.
- ninetyninenine 2y ago> but features are still slowly getting smaller (your own link says so) From the link: "Although miniaturization is still happening, the Moore’s Law standard of doubling the components on a semiconductor chip every two years has been broken" I'm not saying things aren't getting smaller. I'm saying moores law is broken.
- ben_w 2y ago> I'm not saying things aren't getting smaller. I'm saying moores law is broken. And I'm saying that Moore's original statement was already broken by 1975. And that the whole phrase means loads of different things that Moore never actually said, and the one of those which matters here is still true.
- ninetyninenine 2y ago> If you ask MIT Professor Charles Leiserson, Moore’s Law has been over since at least 2016. That’s from the link. I largely agree. Colloquially it’s over after 2016. Any other interpretation is too pedantic imo.
- ben_w 2y agoHow is anything other than joules per operation relevant in the context of this thread? That's the only variant that matters here. (Well, dollars per operation, but assuming energy costs are fixed…)
- ninetyninenine 2y agoThe context is improvement in general.