Once upon a time … Before Willy Wonka came to Lumpaland, hundreds of fierce dragons preyed on ten thousand Oompa Loompas and other wildlife. A thousand Oopmas tried in their free time to oppose dragons, and a hundred had full time jobs at the Dragon Institute.
I find that a lot of people think that academia just rushes in and immediately analyzes any interesting new idea, but of course that's not how it works at all. I.e., people ask me, "What do mainstream AI scientists think of your ideas?" and I have to explain that if I wanted to actually get commentary, I'd have to spend the next several years pushing my ideas in journals and even then it probably wouldn't work.
Great parable. Eliezer could learn a lot about fiction writing from you. I read the parable as an extended answer to Michael Vassar's gripe some weeks back that you ignore his posts/ideas and the posts/ideas of other non-academics on some of these topics. There's no easy answer, but as a third party in a world with both experts and amateurs, I think my rational goal is that experts should engage the brief writings of amateurs to the extent that it will overall maximize the elimination existential thread dragons from Lumpaland. No more, no less. Any more engagement isn't just charity towards amateurs, it's also a reduction of my personal persistence odds. Any less engagement, also a reduction. That's where things begin and end from my perspective.
Okay, so with Robin's addition, here is how I understand the parable. Killing dragons is like making significant progress on scientific problems. Amateurs usually just pontificate, but occasionally they do make real progress. In the parable, dragon-killing was mostly observed by professionals. That is like, recognizing when an amateur has made a genuinely valuable contribution is usually only possible by professionals. And then the professionals may have self-serving biases to downplay the amateur's role or even steal credit.
The bottom line question here, also alluded to by Eliezer's recent postings, is whether we can do better than science as we know it? And the meta-question, how can society know in general if an institution like science, or for that matter the Dragon Institute, is the best way that things could be done?
Science is making genuine progress. I don't think there is much question of this, although obviously it is more true in some fields than others. Would there be more progress if scientists spent more time listening to amateurs and responding to their ideas? Would there be more progress if scientific opinion were more open to new ideas and less wedded to the dominant paradigm? Or would this lead to an ineffectual expenditure of wasted effort chasing down numerous blind alleys?
See my added to the post.
My question is, what are the most difficult parts of dispatching dragons? Is it the ideas the instituters come up with? Or is it other supporting technologies developed in the market at large? Can the average Oompa identify the best dragon-slaying investments, or are they just impressed by the big and important-sounding ideas from the institutes?
IMO, the things that come out of formal academic and research institutions are often vastly over-valued. The "big ideas" which come from institutional think-tanks are often the "easy" part of whatever problem is trying to be solved, while the supporting technologies needed to make the idea viable are often vastly more complex. For example, prediction markets are very simple if you compare them to the tools needed to support them on the Internet (try counting the layers of abstraction between some XML market data and the prediction market itself, on a remote server). In my experience the real bottlenecks to progress are in the devil's details, which aren't obviously identifiable to laymen.
But then might the institutes play an important role? People listen to them. They are impressive. When public opinion itself is a barrier to progress (such as prediction markets under United States law), mightn't the institutions have a valuable role to play?
Sounds like the amateurs have more fun. And why must we always be-doing-something? And why would we ever want to get rid of all the dragons? So in a sense the do-nothings are having more fun and are being more useful.
I don't understand the parable, so I'm just going to take it respond in its own terms.
Seems like the bottom line is, are Instituters successful in getting rid of dragons? Do they have a good track record? Were there more dragons in the past, and fewer today? If they are doing well, I (as an Oompa Loompa) would be inclined to forgive the various inefficiencies.
As far as biases, the main issue IMO is whether the larger society is correctly perceiving the contribution of do-something amateurs. If such amateurs are rare and it is hard to distinguish the ones who are likely to be successful, then again any bias against them is relatively unimportant. However if the Institute's downplaying of the role of successful amateurs is significantly slowing down progress against dragons, then society's bias is imposing real social costs. I suppose the hard part is knowing which is the case.
Instituters were impressive ... Others wanted to read about, sleep with, and study under them.
This is a parable about planet earth?
I think the point is: real-world results are more important than publishing journal articles since (remember "Lost Purposes"?) the latter are supposed to be merely a means to the former.
To give a concrete example, let's say I claimed that we don't currently understand how humans recognize letters and numbers, and my proof is how no one has written a program that beat OB's captcha. Then, let's say someone responded by pointing to the numerous journal articles in the cognitivie sciences attacking the problem of character and object recognition. In such a case, I would be correct: the articles do not substitute for the real-world failure of computers beat the captcha.
Robin_Hanson's point, then, is that academia trivializes the real-world accomplishments, even as that is supposed to be their ultimate goal, and proof, wherever they happen, of the academics' inadequacy.
Very neat parable, but what is it an indictment of? Universities being risk averse in research and gating out independent researchers? Bloggers and old media? Big social problems as dragons and beltway think tanks as institutes? Curiously, I cannot guess what your 'future of humanity' institute would count as (despite 'institute' in its name)...insider or outsider? :)...you've got Ray Kurzweil, but given the stuff you write about (q. mech to behavioral economics and AI), and your use of the 'singularity' concept which isn't a mainstream dragon, I'd say you guys are a definite gray-zone candidate.
But then, Santa Fe managed to work itself from the periphery to a fairly legitimized position in complexity...
Just heard Dean Kamen speak at the IRI btw. Now there's a true amateur dragon-slayer.