I really appreciate the message in this article. I think another way to partition the world is into zero-sum and non-zero sum situations. In non-zero sum situations it's more important to be correct than novel, but in zero-sum situations it's important to be both correct and novel.
One example I like to think about is the following: imagine having two boxes (A and B), one on each side of a football field. Exactly one of the two boxes contains a million dollars: box A contains the prize with 99% probability and box B contains the prize with 1% probability.
Suppose you're standing at the center of the field together with n other people and are told to race to get the million dollars (first person who finds it wins). You have to decide whether to run for box A or box B since they're in opposite directions. If n = 0, it's clearly in your best interest to go for box A, but as n grows, the probability of you getting to box A first drops since you're competing with other people, meaning that at some point you ought to actually go for box B.
This is a toy example of where it's not sufficient to be right, but it's also important to be novel. If each person was awarded its contents (rather than just the first person to get there), there'd be no rush and you ought to go for box A. In other words: in zero-sum situations you need to be both right and novel, but in non-zero sum situations it's sufficient to just be right.
I think the example of academia in the article is not quite right for two reasons. As pointed out elsewhere, the incentives in academia are actually quite twisted which means that there's actually more incentive to just be verbose than to be novel, but additionally I don't think it's really zero-sum. It's true that if you're working in a crowded space, it starts approaching zero-sum and then becomes a game of novelty. However, if you're not in a crowded space (e.g. interdisciplinary stuff) you can actually be quite successful by discovering rather mundane things (you might say that this is just being novel on a different axis, which I think is reasonably accurate, but the point made in the article was that the discoveries have to be surprising).
Yes and no. If I wanted to start a social network today, it would not be sufficient to just clone Facebook because the space is already so saturated (it's zero-sum). I would have to have something very novel. If, however, I wanted to start a social network back in the early 2000s, I think it would have been sufficient to clone Facebook because the space was not yet saturated. This is not to say that it would have worked out, of course. Most likely I would have been crushed by Zuckerberg.
When a new technology emerges, there will be a lot of "obvious" ideas that are ready to be "plucked", but as that technology matures, these obvious ideas become saturated, effectively turning the space into a zero-sum game. For example, in the early days of the app store / mobile revolution, I think there were a lot of reasonably straight-forward ideas out there ready to be executed on, e.g. Uber. I'm not saying that Uber was necessarily obvious to everyone, but I don't think that the idea of summoning a taxi from your phone would have been met with comments like "that's crazy, nobody would ever want that." However, if you wanted to come up with a startup around mobile today, you'd be better be very novel because pretty much everything obvious is already taken.
Looking at trends today like autonomous cars, I think there are plenty of reasonably obvious ideas that we all sort of expect will be part of the first generation of applications (shipping, robo-taxies, last-mile delivery). That's not to say that they're easy to implement or by any means are guaranteed to succeed, but I don't think that they'll necessarily be super controversial.
One example I like to think about is the following: imagine having two boxes (A and B), one on each side of a football field. Exactly one of the two boxes contains a million dollars: box A contains the prize with 99% probability and box B contains the prize with 1% probability.
Suppose you're standing at the center of the field together with n other people and are told to race to get the million dollars (first person who finds it wins). You have to decide whether to run for box A or box B since they're in opposite directions. If n = 0, it's clearly in your best interest to go for box A, but as n grows, the probability of you getting to box A first drops since you're competing with other people, meaning that at some point you ought to actually go for box B.
This is a toy example of where it's not sufficient to be right, but it's also important to be novel. If each person was awarded its contents (rather than just the first person to get there), there'd be no rush and you ought to go for box A. In other words: in zero-sum situations you need to be both right and novel, but in non-zero sum situations it's sufficient to just be right.
I think the example of academia in the article is not quite right for two reasons. As pointed out elsewhere, the incentives in academia are actually quite twisted which means that there's actually more incentive to just be verbose than to be novel, but additionally I don't think it's really zero-sum. It's true that if you're working in a crowded space, it starts approaching zero-sum and then becomes a game of novelty. However, if you're not in a crowded space (e.g. interdisciplinary stuff) you can actually be quite successful by discovering rather mundane things (you might say that this is just being novel on a different axis, which I think is reasonably accurate, but the point made in the article was that the discoveries have to be surprising).