5 ms·
It is hard to figure out when exponential growth will end while you are in the middle of it. Linus said something relevant to this in a recent interview on Sla
by sangnoir 11y ago
It is hard to figure out when exponential growth will end while you are in the middle of it.
Linus said something relevant to this in a recent interview on Slashdot[1], answering a question on dangerous AI. To paraphrase him: people are crazy to think that exponential growth lasts forever. As impressive as it may be (at the time), it's only the beginning of an S-curve.
I've extracted a small part of his answer below. I think the whole interview is worth reading (it's a general interview, not AI-specific)
... I'd expect just more of (and much fancier) rather targeted AI, rather than anything human-like at all. Language recognition, pattern recognition, things like that. I just don't see the situation where you suddenly have some existential crisis because your dishwasher is starting to discuss Sartre with you.
The whole "Singularity" kind of event? Yeah, it's science fiction, and not very good SciFi at that, in my opinion. Unending exponential growth? What drugs are those people on? I mean, really..
It's like Moore's law - yeah, it's very impressive when something can (almost) be plotted on an exponential curve for a long time. Very impressive indeed when it's over many decades. But it's _still_ just the beginning of the "S curve". Anybody who thinks any different is just deluding themselves. There are no unending exponentials
1. http://linux.slashdot.org/story/15/06/30/0058243/interviews-linus-torvalds-answers-your-question http://linux.slashdot.org/story/15/06/30/0058243/interviews-...
- tim333 11y agoI don't think singularity stuff relies on exponential growth lasting for ever. Only for AI performance to significantly exceed human. And I'm not sure there's anything magical about the human brain that will stop tech achieving similar performance before it runs into barriers.
- daveloyall 11y agoRight. I'll attempt to bolster your argument, if I may... We know that human and animal brains are made out of simple components. Probably the same components. The difference between animals and humans seems to be in how those components are arranged and how many of them there are. We've already demonstrated digital components that approximate individual neurons, and we've already demonstrated wiring groups of them up to perform useful work. I think that if we find ourselves at an inflection point in history, it is because there might exist some relatively simply arrangement of digital neurons which results in a machine which can learn some sort of rudimentary self awareness[1] while also possessing a computer's ability to do math, store data, and recall data, communicate on the internet, etc. 1. Spend year of days with a newborn if you don't already know that humans have to learn it, too.
- TeMPOraL 11y agoAnother thing is that we've figured out a medium that can compute faster than human and animal brains. In theory, a brain mimicked in silicon should run faster than wet one, and that itself coupled with some optimized access (i.e. not eyes OCRing text) to information would be enough for an entity to be smarter than humans. Of course one shouldn't say that a silicon brain would be better than a wet one - by doing computing the way we do it, we're sacrificing a lot. Biological systems are more resilient, can repair and replicate themselves without a need for complex processing industry we have (to be honest, the ecosystem is sort of an industry, but it's already well developed and much better integrated than the human-made). But there are things we could trade off for increased intelligence when designing our own minds.
- rhaps0dy 11y agoSingularity stuff just needs an AI that is human-level and can improve itself. Then again, the self-improving feedback loop might stop for reasons unknown now.