5 ms·
What a bizarre logical leap this early paragraph makes to try to explain why AI is beyond the capabilities of mere humans: "AI can be thought of as a search pr
by jameshart 11y ago
What a bizarre logical leap this early paragraph makes to try to explain why AI is beyond the capabilities of mere humans:
"AI can be thought of as a search problem over an effectively infinite, high-dimensional landscape of possible programs. Nature solved this search problem by brute force, effectively performing a huge computation involving trillions of evolving agents of varying information processing capability in a complex environment (the Earth). It took billions of years to go from the first tiny DNA replicators to Homo Sapiens. ... we have little idea of how to find the tiny subset of all possible programs running on this hardware that would exhibit intelligent behavior."
That's... a bizarre argument. You might as well argue that since heavier-than-air-flight is a search problem over an effectively infinite, high-dimensional landscape of possible machines - and that since it took evolution billions of years to produce birds, we have little chance of stumbling upon a working design for a wing.
Of course evolution's 'brute force' breadth-first hillclimb takes billions of years to find solutions to problems. It's undirected and unintelligent. Engineers don't have to perform undirected unintelligent searches across the infinite space of possible solutions to problems, they can think and plan and learn. I see no reason to see AI as a particularly different class of scientific endeavor, more complex than rocket science or nuclear engineering or biochemistry, such that ordinary, unenhanced human brains can't hope to comprehend it sufficiently to design machines that are capable of intelligence. This sounds a little like an 'if man were meant to fly, god would've given us wings' argument.
- jessriedel 11y agoHuh? He's not arguing that "AI is beyond the capabilities of mere humans". (Where did you get that from?) He's arguing that it's a hard problem, that the presence of lots of brains can lead people to mistakenly conclude it's easy, and that it will be easier with smarter humans. His argument would have worked just as well concerning heavier-than-air-flight, meaning that just because we see birds everywhere doesn't mean it's easy. It's conceivable that un-aided humans are too dumb to crack intelligence from the ground up (and therefore would have to brute-force the solution through simulation), but Hsu isn't even arguing for that.
- jameshart 11y agoHmm. I very strongly got the opposite impression. "Today, we need geniuses like von Neumann and Turing more than ever before. That’s because we may already be running into the genetic limits of intelligence.... AI research also pushes even very bright humans to their limits.... The detailed inner workings of a complex machine intelligence (or of a biological brain) may turn out to be incomprehensible to our human minds—or at least the human minds of today". (emphasis mine) The entire thesis seems to be we will need artificial cognitive enhancements to be able to comprehend AI sufficiently to create it. It really sounds like he thinks we are too dumb to figure this out, that we are just scrabbling around in the dark trying solutions at random.
- jessriedel 11y agoEven if it's true that humans are too dumb to crack AI from first principles -- which as I said is certainly conceivable, and which Hsu alludes to in the quote you've selected -- he's not basing that on the infinite-dimensional space description in the second paragraph. The point of the second paragraph was just to caution against concluding it's an easy problem based on the abundance of intelligence around us.
- jameshart 11y agoAnd my point is that that's not a very good argument. We don't have to use random walk algorithms to look for our solutions to the problem. If someone shows you a list of sorted numbers and tells you they were sorted over a period of millions of years using bogosort, that may be true but it is not a good argument against the existence of efficient sorting algorithms.
- darkmighty 11y agoEvolution is not exactly a random walk, it's a lot better. But it's plausible that creating an AI more "intelligent" than humans (whatever that means) is beyond our current intellectual capabilities, although I seriously doubt it. He's forgetting two other scaling modes applicable to human intelligence: accumulating knowledge and collaborating. By accumulating knowledge we can build upon the work of the generations before us and take essentially unlimited time to crack a problem. Collaboration allows not only parallel processing of problems but also interacting and exchanging inspiration towards the solution of a single problem; both which scale quite well with a lot of intelligent people. Note that there are almost trivial arguments showing some problems are beyond the means of our finite brain (viz unbounded algorithmic complexity or the halting problem).
- giardini 11y ago"You might as well argue that since heavier-than-air-flight is a search problem over an effectively infinite, high-dimensional landscape of possible machines - and that since it took evolution billions of years to produce birds, we have little chance of stumbling upon a working design for a wing." Tired analogies between AI and flight are dead-ends. Until we have identified AI-side components of the analogy that corresponds to "air", "wing", "lift" &c., the analogy is empty and unproductive. IOW these analogies neither get us off the ground nor do they take us anywhere.
- JoeAltmaier 11y agoAlso, flight took very little to evolve - it happened separately for birds, mammals, lizards, insects. In less than billions of years.
- olalonde 11y agoAmazingly, the eye also evolved 50 to 100 times independently (https://en.wikipedia.org/wiki/Evolution_of_the_eye https://en.wikipedia.org/wiki/Evolution_of_the_eye).
- jameshart 11y agoAs for the subject of this discussion, you could have quite the debate about how many times intelligence has evolved independently.
- jameshart 11y agoI'm not making an analogy between AI and flight, I'm refuting an analogy between 'evolving' and 'inventing' solutions using flight as a counterexample.
- giardini 11y agoWhatever. Unfortunately, in doing so you have used the "AI is analogous to heavier-than-air-flight" meme. Please discard it, it's an albatross to discussion. Or to be more precise, it is a lead balloon - it never gets off the ground; it's a red herring. IOW you've chosen a poor and annoyingly wrongheaded counterexample. Now try and digest that.
- andreyf 11y agoAlso, the somewhat more troubling genetic explanation for individual's achievements (e.g. von Neumann), based on anecdotes from "some guy who knew him" and "some guy who works at MIT". Genetics surely has a part to play in intelligence, but to say that it will turn out to be the determining factor in leu of resources or environment seems to me a bit of an extraordinary claim. I love and agree with the premise that AI augmented people will be enormously more productive than traditionally intelligent people, but I think that access to educational and child rearing resources will define the difference, not genetics.
- decisiveness 11y agoThe engineering of flight is something that's achievable through simple understanding of physics laws and principles. Birds did not engineer flight, we did, and we were only able to engineer a crude version of it. The natural abilities of flight in birds are still superior in many ways to engineered flight. So, how can we assume from this that we'll be able to build a version of ourselves that's completely more intelligent? Humans engineering functional variants of inferior species' physical capabilities is undeniably incomparable to humans engineering a better, more intelligent version of ourselves. Maybe, a beneficial albeit simpler version can be engineered, but one that makes us obsolete? It's hard for me to conclude that's not a huge leap in assumption.