solving problems

When trying to solve a problem, you can either seek out the information required to solve your problem (and with that information draw your own conclusions), or you can seek out answers to similar problems and apply the best of those answers to your problem.

Which do you think is better?

After thinking about this a lot over some red wine and Conchiglie della Casa with K, I think that there is no doubt in my mind that the first option is infinitely superior to the second. Not only that, but I think that people who do the first will inexplicably succeed while people who do the second will inexplicably fail. The part that really interests me is how indistinguishable the two methods are--on their surface they seem to be almost exactly the same thing, and talking with people I don't think you'll be able to tell if they've used the first or second method (they may even interchange the methods without thinking about it).

What's the big difference, then? From a data-centric perspective, the difference is clear: the first method has the possibility of generating new information, and the second method absolutely doesn't. Creating new information is the only way to create value. By finding new answers, your project will therefore create value, and your project will be more likely to succeed. By relying on existing information, your project will never add more value to the data environment, and therefore there's no new information to distribute, and you have in fact manufactured no innovation, and it could be said that you haven't even made a decision.

In a world where only good ideas are rewarded, people recycling old ideas will only ever ride the tails of succeeding projects, and they will never surpass the projects they are trying to imitate or learn from. It goes against common sense, in some ways: learning from the past is not actually valuable, unless the lesson you learn is that you should not rely on the past for the answers to future questions. It's a subtle ecosystem, though... you can imagine a spectrum of ideas that all look like they were generated the same way, but in actuality some are riding in the wake of other, succeeding in limited ways that make them scratch their heads wondering why they're not seeing the same success as the wake-generators, when in fact it's just a simple matter of seeing who is generating new information and who isn't.

Comments

[Don]

There's one distinct advantage to building on the ideas of others: presumably they haven't broken yet, so you have a reasonably sturdy foundation.

Given your A and B choices for approaches, I'd be more inclined to pick C, which is somewhere between the two. Innovation is good, new ideas and information are good, but that doesn't necessarily mean old information is inherently bad. If I was going to rest 15,000 pounds on a slab of a certain material, I'd really appreciate seeing examples of earlier attempts to do so in related situations that did or did not work.

Alternately, perhaps there's a problem with the definition of A and B. If the past can be learned from at all (putting too much weight on this sort of block makes the block break, for example), then looking at the solutions to related problems (B) is part of finding the information required to solve your problem (A). So there is no B, since it's a subset of A. B only comes in when the solution one comes up with is sufficiently similar to solutions to similar problems that someone could say, "Hey, you just copied that guy!"

If that works, let 'em say that. Can't argue with results.

[Erik Benson]

Yeah, I know it's not as clear cut as I pretended it was, and it's definitely not that easy to tell if you're using A or B without really thinking about when. One thing I may not have said though is that when trying to solve a problem, the only difference between A and B is that B tries to find answers to similar problems. A might look for answers to small parts of the problem (like choosing the right type of block), but presumably the problem they're trying to solve isn't "Which type of block doesn't break under too much weight?" If it were, then obviously they'd be using some version of method B to solve this problem and they should not expect any new information or value to come from the solution to this problem.

That said, I guess there are a lot of problems that aren't designed to create new value, and so method B will work fine for them. For the purposes of this entry, I was assuming that people wanted to create new, valuable, products.

[capitalist]

you linked to your wishlist !

[ricardo vacapinta]

There's a standard toolset for solving problems in the world of mathematics. One of the most common is to rely on isomorphisms, or analogies across problems which at first would seem to have no relation.

Kaluza-Klein theories were constructed on the premise that Einstein's general theory which likened gravity to curvature in space-time might be extended to other forces such as electromagnetism. These KK theories are the forerunners of modern day string-theories which also rely on extra dimensions.

I do believe however that A is the path towards greater innovation. Whenever I hear stuff like "Here's this swarm theory which mimics how ants solve problems" I think, bah, what other sorts of metaphors are there in the world, undiscovered by evolution, which, if we apply pure force of intellect and discard our preconceptions about the world and our worn analogies may allow us to truly surprise ourselves.