4 ms·

My understanding is that most (or at least many) very significant scientific discoveries were serendipitous. They were discovered while looking for something el

by rlpb 4y ago

My understanding is that most (or at least many) very significant scientific discoveries were serendipitous. They were discovered while looking for something else. So statements like "When you don’t have much to go on, and limited resources, it’s better to aim at problems that you know are out there. Those things will lead you to new discoveries." don't make sense to me. I'd find this article and its opinion more credible if it could point out how this didn't apply to the LHC. Was every discovery and technology that came out of the LHC predicted in advance? If not, then this argument fails.

ThePhysicist 4y ago
For serendipity to work its magic you need to be able to explore the solution space. In high energy physics that usually means putting more energy into a system, as (naively said) that will increase the amount of stuff that will happen. So I'm all for building bigger and bigger accelerators, even if they cost trillions of dollars to make, as we have only explored a tiny piece of the energy scale so far, and we're pretty sure very interesting stuff happens at higher scales (ultimately, grand unification of all forces). Also, there's a mutual relationship between theory and experiment in physics. The LHC was well motivated by theoretical predictions (most notably the Higgs boson) that needed experimental verification. In physics you have to put your theories to the test, otherwise they're not worth much. Each experimental result is like a nail that fixes a part of your theory space (by excluding possible alternative theories), and that then allows you to explore the space further with greater confidence. Hence the more nails, the better.
radu_floricica 4y ago
You're missing the second part of his argument. Serendipitous discoveries can come from anywhere, so it does make sense to pay attention to cost. Which would you think would be a better bet for a random win, 100k separate experiments, or just one large experiment that's a scaled up version of something we already have? If you say the second, you have severe scope neglect: 100k is a huge number. Of course, one expensive experiment may be the right call in some situations, but only if you have good reason to think it'll pay off. Either because you're certain it will, or because the expected utility of unknown results is larger than the cost (expected utility being probability times value). This very much isn't the case here.
brazzy 4y ago
The statement "When you don’t have much to go on, and limited resources, it’s better to aim at problems that you know are out there. Those things will lead you to new discoveries." which you say doesn't make sense to you, is exactly meant to promote "serendipitous" discoveries while looking for something else. A bigger collider does not, in fact, have anything to look for, and so the only possible benefit it could yield is such "serendipitous" discoveries. But how many of those could there really be when you're essentially doing only one thing? The point of the article is that the ridiculous amount of money, which a bigger collider would cost, is much better spent on a large amount of much more varied research activities, which do have concrete things to look for, but also have the potential for serendipitous discoveries. And since those activities are much more varied, the total potential for such discoveries would much larger as well.

by rlpb 4y ago

ThePhysicist 4y ago

For serendipity to work its magic you need to be able to explore the solution space. In high energy physics that usually means putting more energy into a system, as (naively said) that will increase the amount of stuff that will happen. So I'm all for building bigger and bigger accelerators, even if they cost trillions of dollars to make, as we have only explored a tiny piece of the energy scale so far, and we're pretty sure very interesting stuff happens at higher scales (ultimately, grand unification of all forces). Also, there's a mutual relationship between theory and experiment in physics. The LHC was well motivated by theoretical predictions (most notably the Higgs boson) that needed experimental verification. In physics you have to put your theories to the test, otherwise they're not worth much. Each experimental result is like a nail that fixes a part of your theory space (by excluding possible alternative theories), and that then allows you to explore the space further with greater confidence. Hence the more nails, the better.

radu_floricica 4y ago

You're missing the second part of his argument. Serendipitous discoveries can come from anywhere, so it does make sense to pay attention to cost. Which would you think would be a better bet for a random win, 100k separate experiments, or just one large experiment that's a scaled up version of something we already have? If you say the second, you have severe scope neglect: 100k is a huge number. Of course, one expensive experiment may be the right call in some situations, but only if you have good reason to think it'll pay off. Either because you're certain it will, or because the expected utility of unknown results is larger than the cost (expected utility being probability times value). This very much isn't the case here.

brazzy 4y ago

The statement "When you don’t have much to go on, and limited resources, it’s better to aim at problems that you know are out there. Those things will lead you to new discoveries." which you say doesn't make sense to you, is exactly meant to promote "serendipitous" discoveries while looking for something else. A bigger collider does not, in fact, have anything to look for, and so the only possible benefit it could yield is such "serendipitous" discoveries. But how many of those could there really be when you're essentially doing only one thing? The point of the article is that the ridiculous amount of money, which a bigger collider would cost, is much better spent on a large amount of much more varied research activities, which do have concrete things to look for, but also have the potential for serendipitous discoveries. And since those activities are much more varied, the total potential for such discoveries would much larger as well.