That's a very interesting essay indeed. Some thoughts:
You see again and again, that it is more than one thing from a good person. Once in a while a person does only one thing in his whole life, and we'll talk about that later, but a lot of times there is repetition. I claim that luck will not cover everything.
I agree, but I'm reminded of the story of the mathematician on great generals. That's not necessarily the case here, but it is something to think about.
Most of you in this room probably have more than enough brains to do first-class work. But great work is something else than mere brains.
Indeed I'm reminded of Scott Alexander's comments on the Good Judgment Project and how although there is a correlation correct people aren't necessarily the smartest people.
Once you get your courage up and believe that you can do important problems, then you can.
Courage is a non-technical term which I hate. I think he means disagreeableness based on his wording which I'd agree that excessive agreeableness is probably more detrimental than excessive disagreeableness, but it seems like maximal disagreeableness would also be bad.
Very clearly they are not because people are often most productive when working conditions are bad. One of the better times of the Cambridge Physical Laboratories was when they had practically shacks - they did some of the best physics ever.
So less funding is better? (Yes, sentence fragment, get over it) That is an interesting idea. Most people, including a couple commenters on this post, would probably disagree with that. The drugmonkey blog spends many words complaining about funding issues. I'm wondering about possible confounds assuming this is a real phenomenon. Is it the lack of funding, the lack of respect, the isolation of having fewer researchers in one area (too many cooks problem)?
What Bode was saying was this: ``Knowledge and productivity are like compound interest.'' Given two people of approximately the same ability and one person who works ten percent more than the other, the latter will more than twice outproduce the former.
I'm not sure how someone could've figured this out. Maybe this is owing to my lack of experience; this is something you can only discover in your 50s or 60s after several decades of seeing people working on difficult problems. I'm not sure I trust this.
On this matter of drive Edison says, ``Genius is 99% perspiration and 1% inspiration.'' He may have been exaggerating, but the idea is that solid work, steadily applied, gets you surprisingly far.
I strongly doubt this is true. If the majority of possibilities are wrong, then improving search algorithms should have a bigger payoff than increasing productivity although there is a proper balance that needs to be found.
They believe the theory enough to go ahead; they doubt it enough to notice the errors and faults so they can step forward and create the new replacement theory. These seem like orthagonal ideas; not necessarily opposed ones as he suggests.
And after some more time I came in one day and said, ``If what you are doing is not important, and if you don't think it is going to lead to something important, why are you at Bell Labs working on it?'' I wasn't welcomed after that; I had to find somebody else to eat with! … The average scientist, so far as I can make out, spends almost all his time working on problems which he believes will not be important and he also doesn't believe that they will lead to important problems.
Oh! That's why you said I should read this.
I'm way off-task, so I'll come back to this after I finish some work. It is a fascinating read though. Thank you.
Courage is a non-technical term which I hate. I think he means disagreeableness based on his wording which I'd agree that excessive agreeableness is probably more detrimental than excessive disagreeableness, but it seems like maximal disagreeableness would also be bad.
Courage is an emotion. Emotions matter. Don't try to eliminate them for the equation just because you don't like them.
I've been thinking lately about what is the optimal way to organize scientific research both for individuals and for groups. My first idea: research should have a long-term goal. If you don't have a long-term goal, you will end up wasting a lot of time on useless pursuits. For instance, my rough thought process of the goal of economics is that it should be “how do we maximize the productive output of society and distribute this is in an equitable manner without preventing the individual from being unproductive if they so choose?”, the goal of political science should be “how do we maximize the government's abilities to provide the resources we want while allowing individuals the freedom to pursue their goals without constraint toward other individuals?”, and the goal of psychology should be “how do we maximize the ability of individuals to make the decisions they would choose if their understanding of the problems they encounter was perfect?” These are rough, as I said, but I think they go further than the way most researchers seem to think about such problems.
Political science seems to do the worst in this area in my opinion. Very little research seems to have anything to do with what causes governments to make correct decisions, and when they do research of this type, their evaluation of correct decision making often is based on a very poor metric such as corruption. I think this is a major contributor to why governments are so awful, and yet very few political scientists seem to have well-developed theories grounded in empirical research on ways to significantly improve the government. Yes, they have ideas on how to improve government, but they're frequently not grounded in robust scientific evidence.
Another area I've been considering is search parameters of moving through research topics. An assumption I have is that the overwhelming majority of possible theories are wrong such that only a minority of areas of research will result in something other than a null outcome. Another assumption is that correct theories are generally clustered. If you get a correct result in one place, there will be a lot more correct results in a related area than for any randomly chosen theory. There seems like two major methods for searching through the landscape of possibilities. One method is to choose an area where you have strong reason to believe there might be a cluster nearby that fits with your research goals and then randomly pick isolated areas of that research area until you get to a major breakthrough, then go through the various permutations of that breakthrough until you have a complete understanding of that particular cluster area of knowledge. Another method would be to take out large chunks of research possibilities, and to just throw the book at it basically. If you come back with nothing, then you can conclude that the entire section is empty. If you get a hit, you can then isolate the many subcomponents and figure out what exactly is going on. Technically I believe the chunking approach should be slightly faster than the random approach, but only by a slight amount unless the random approach is overly isolated. If the cluster of most important ideas are at 10 to the -10th power, and you isolate variables at 10 to the -100th power, then time will be wasted going back up to the correct level. You have to guess what level of isolation will result in the most important insights.
One mistake I think is to isolate variables, and then proceed through the universe of possibilities systematically one at a time. If you get a null result in one place, it's likely true that very similar research will also result in a null result. Another mistake I often see is researchers not bothering to isolate after they get a hit. You'll sometimes see thousands of studies on the exact same thing without any application of reductionism eg the finding that people who eat breakfast are generally healthier. Clinical and business researchers seem to most frequently make this mistake of forgetting reductionism.
I'm also thinking through what types of research are most critical, but haven't gotten too far in that vein yet. It seems like long-term research (40+ years until major breakthrough) should be centered around the singularity, but what about more immediate research?