If I had to take an honest guess? Theoretical discovery will behave "inefficiently" when it requires a breadth-first (or at least, breadth-focused) search through the idea space before you can find things that "fit together". Only once you have a bunch of things which "fit together" can you look at the shape of the "hole in idea-space" they all border, dive to the bottom of that lake, and bring up an entirely new idea which links them or unifies them.
So:
1) Mostly agreed, as described above.
2) As described above. My reasoning is sociological: our current reward system for researchers optimizes our research process for depth rather than breadth. Looking where others are paid not to look would usually be a decent way to find things others haven't seen.
3) I don't know more than the bare minimum about decision theory, so I can't say.
Now, as to analysis of an intelligence explosion taking such a long time, I have an Opinion (beware the capital letter): there may not be an intelligence explosion. Current research into AI indicates that coming up with formal models of utility-optimal agents in unknown active environments is the easy part, making them conscious and (stably) self-modifying is the "next part" currently under research, and scaling them down to fit inside the real universe is the hard part.
Schmidhuber (and when it comes to UAI I definitely root for the Schmidhubristic team ;-)) has claimed that Goedel Machine self-rewrites would dramatically speed up a mere AIXI paired with a mere HSearch until they became effective within the real world, but that's putting his faith in the Goedel Machine's proof searcher rather than in his (and his student's) own proofs that their algorithms genuinely are optimal and the problems they're tackling genuinely do have these nasty time-and-space bounds. If the first team to build an AI has to either keep it in an exponentially-small environment (existing AIXI models) or wait for astronomical periods of time for even the first self-rewrite, then the human race will die of an asteroid strike long before Friend Clippy can take us out.
This is the same nasty problem that AI has faced since the GOFAI days: easy to find an algorithm that locates the optimal move by searching the entire solution space, hard as hell to prove it will take any usefully small period of time to do so.
Previously: Why Neglect Big Topics.
Why was there no serious philosophical discussion of normative uncertainty until 1989, given that all the necessary ideas and tools were present at the time of Jeremy Bentham?
Why did no professional philosopher analyze I.J. Good’s important “intelligence explosion” thesis (from 19591) until 2010?
Why was reflectively consistent probabilistic metamathematics not described until 2013, given that the ideas it builds on go back at least to the 1940s?
Why did it take until 2003 for professional philosophers to begin updating causal decision theory for the age of causal Bayes nets, and until 2013 to formulate a reliabilist metatheory of rationality?
By analogy to financial market efficiency, I like to say that “theoretical discovery is fairly inefficient.” That is: there are often large, unnecessary delays in theoretical discovery.
This shouldn’t surprise us. For one thing, there aren’t necessarily large personal rewards for making theoretical progress. But it does mean that those who do care about certain kinds of theoretical progress shouldn’t necessarily think that progress will be hard. There is often low-hanging fruit to be plucked by investigators who know where to look.
Where should we look for low-hanging fruit? I’d guess that theoretical progress may be relatively easy where:
These guesses make sense of the abundant low-hanging fruit in much of MIRI’s theoretical research, with the glaring exception of decision theory. Our September decision theory workshop revealed plenty of low-hanging fruit, but why should that be? Decision theory is widely applied in multi-agent systems, and in philosophy it’s clear that visible progress in decision theory is one way to “make a name” for oneself and advance one’s career. Tons of quality-adjusted researcher hours have been devoted to the problem. Yes, new theoretical advances (e.g. causal Bayes nets and program equilibrium) open up promising new angles of attack, but they don’t seem necessary to much of the low-hanging fruit discovered thus far. And progress in decision theory is definitely not valuable only to those with unusual views. What gives?
Anyway, three questions:
1 Good (1959) is the earliest statement of the intelligence explosion: “Once a machine is designed that is good enough… it can be put to work designing an even better machine. At this point an ”explosion“ will clearly occur; all the problems of science and technology will be handed over to machines and it will no longer be necessary for people to work. Whether this will lead to a Utopia or to the extermination of the human race will depend on how the problem is handled by the machines. The important thing will be to give them the aim of serving human beings.” The term itself, “intelligence explosion,” originates with Good (1965). Technically, artist and philosopher Stefan Themerson wrote a "philosophical analysis" of Good's intelligence explosion thesis called Special Branch, published in 1972, but by "philosophical analysis" I have in mind a more analytic, argumentative kind of philosophical analysis than is found in Themerson's literary Special Branch. ↩