4 ms·

> More knowledge and options are always better. That's not even true in theory. Given a simple problem A, when adding more options, at some point, choosing amo

> More knowledge and options are always better. That's not even true in theory. Given a simple problem A, when adding more options, at some point, choosing among the options requires more effort than solving the simple problem, if only by brute force. There's a reason why RDBMS sometimes skip indexes. And in practice, GP is right: with humans, too many options eventually only distract from the problem.

SapporoChris 4y ago
Well illustrated by Munroe as always. https://xkcd.com/309/ https://xkcd.com/309/
ragle 4y ago
> Given a simple problem A, when adding more options, at some point, choosing among the options requires more effort than solving the simple problem, if only by brute force. Hick's Law, more or less: https://en.wikipedia.org/wiki/Hick%27s_law https://en.wikipedia.org/wiki/Hick%27s_law
niclo 4y ago
But you are considering an increase only in options without the increase in knowledge to better evaluate options parent is suggesting, and you are putting the constraint that only single option at time may be adopted while in reality multiple may work in parallel/together just fine.
kbenson 4y ago
> Given a simple problem A, when adding more options, at some point, choosing among the options requires more effort than solving the simple problem Only if you require that every option must be assessed. You can always choose a random subset, or the first X entries, as use those as the solution space. Additional solutions provide for additional strategies in choosing a solution, but at any time you can ignore additional solutions and treat them as if they don't exist. > And in practice, GP is right: with humans, too many options eventually only distract from the problem. The whole point of government is to provide a structure such that decisions like this can be made be reducing the set of people that needs to reach consensus to a manageable about. Those people can choose to accept or ignore any solution presented, and call on whatever expertise they need to make that decision. The whole idea of there being too many solutions only makes sense when you you apply it towards the general public and people whose job isn't to make such a decision or be domain experts. Since those people aren't the ones actually making the decision, providing too many solutions to the problem to them is a non-issue, and for the actual decision makers and domain experts they have on call, additional solutions are beneficial. Whether you or I are getting overwhelmed with the possible solutions is irrelevant. We aren't ultimately the ones that will decide what action to take (at least directly), and we aren't the domain experts advising those that will decide the plans (or present the ultimately small set of options that are valid). At least I don't think so, unless you're a domain expert, but even then, your role would just be to provide advice (or if you're a politician in charge of making this decision, but that's even less likely).
- dllthomas 4y ago
  If you're choosing a random subset, then new worse options may crowd out existing better options, so it remains the case that more options isn't always better.
- roughly 4y ago
  > Only if you require that every option must be assessed. You can always choose a random subset, or the first X entries, as use those as the solution space. Additional solutions provide for additional strategies in choosing a solution, but at any time you can ignore additional solutions and treat them as if they don't exist. This presumes that options are of equal value, or at the least that there are more than N actual viable solutions, where N is sufficient to ensure whatever sample size you’re using includes one of the viable ones. “Ignoring additional solutions” presupposes that you know which solutions you should ignore and still requires some form of assessment of the solution space. As for the rest - while indeed it is the job of government to structure these decisions, recent history suggests that the opinions of the broader public still play a role in those decisions, that decisions are rarely made based purely off their technical merits, and that snake-oil salesmen can still do measurable harm to decision-making in representative governments.
  - kbenson 4y ago
    > This presumes that options are of equal value, or at the least that there are more than N actual viable solutions, where N is sufficient to ensure whatever sample size you’re using includes one of the viable ones. That that presumes we're looking through ideas sequentially and thus are actually limited with the number of choices we're generally presented with, when really it's more akin to mapreduce when assessed at the level of leadership and domain experts. Between billions of people and thousands of domain experts, I don' think we're actually reaching any real limits, given the logarithmic nature of filtering good ideas in this way. What that means is that in effect there's a pool of solutions already filtered for viability and usefulness to choose from. All we're doing by entertaining the idea that we should stop looking for and presenting solutions is giving people an excuse to prevent ideas they don't agree with from being spread to others. The whole idea of one person thinking that because they think an idea doesn't have merit that others shouldn't see it is itself harmful. > while indeed it is the job of government to structure these decisions, recent history suggests that the opinions of the broader public still play a role in those decisions, that decisions are rarely made based purely off their technical merits, and that snake-oil salesmen can still do measurable harm to decision-making in representative governments. There are many possible reasons why we don't see progress from our leaders on certain issues. I'm not convinced that there being 1000 possible solutions is any worse than there being 10. I suspect that the same forces that prevent progress work the same in both cases, and if we're talking about less than 10 possible solutions, then I think we haven't examined the problem well enough. In any case, I don't think the problem of there being too many solutions and too much information is as obviously the problem as others are presenting it, and would want to see a real argument presented and defended before I would accept it at face value. That's not because I'm trying to be overly argumentative, just that as I noted, I just don't think it's as obvious a conclusion as others seem to.

SapporoChris 4y ago

Well illustrated by Munroe as always. https://xkcd.com/309/ https://xkcd.com/309/

ragle 4y ago

> Given a simple problem A, when adding more options, at some point, choosing among the options requires more effort than solving the simple problem, if only by brute force. Hick's Law, more or less: https://en.wikipedia.org/wiki/Hick%27s_law https://en.wikipedia.org/wiki/Hick%27s_law

niclo 4y ago

But you are considering an increase only in options without the increase in knowledge to better evaluate options parent is suggesting, and you are putting the constraint that only single option at time may be adopted while in reality multiple may work in parallel/together just fine.

kbenson 4y ago

> Given a simple problem A, when adding more options, at some point, choosing among the options requires more effort than solving the simple problem Only if you require that every option must be assessed. You can always choose a random subset, or the first X entries, as use those as the solution space. Additional solutions provide for additional strategies in choosing a solution, but at any time you can ignore additional solutions and treat them as if they don't exist. > And in practice, GP is right: with humans, too many options eventually only distract from the problem. The whole point of government is to provide a structure such that decisions like this can be made be reducing the set of people that needs to reach consensus to a manageable about. Those people can choose to accept or ignore any solution presented, and call on whatever expertise they need to make that decision. The whole idea of there being too many solutions only makes sense when you you apply it towards the general public and people whose job isn't to make such a decision or be domain experts. Since those people aren't the ones actually making the decision, providing too many solutions to the problem to them is a non-issue, and for the actual decision makers and domain experts they have on call, additional solutions are beneficial. Whether you or I are getting overwhelmed with the possible solutions is irrelevant. We aren't ultimately the ones that will decide what action to take (at least directly), and we aren't the domain experts advising those that will decide the plans (or present the ultimately small set of options that are valid). At least I don't think so, unless you're a domain expert, but even then, your role would just be to provide advice (or if you're a politician in charge of making this decision, but that's even less likely).

dllthomas 4y ago
If you're choosing a random subset, then new worse options may crowd out existing better options, so it remains the case that more options isn't always better.
roughly 4y ago
> Only if you require that every option must be assessed. You can always choose a random subset, or the first X entries, as use those as the solution space. Additional solutions provide for additional strategies in choosing a solution, but at any time you can ignore additional solutions and treat them as if they don't exist. This presumes that options are of equal value, or at the least that there are more than N actual viable solutions, where N is sufficient to ensure whatever sample size you’re using includes one of the viable ones. “Ignoring additional solutions” presupposes that you know which solutions you should ignore and still requires some form of assessment of the solution space. As for the rest - while indeed it is the job of government to structure these decisions, recent history suggests that the opinions of the broader public still play a role in those decisions, that decisions are rarely made based purely off their technical merits, and that snake-oil salesmen can still do measurable harm to decision-making in representative governments.
- kbenson 4y ago
  > This presumes that options are of equal value, or at the least that there are more than N actual viable solutions, where N is sufficient to ensure whatever sample size you’re using includes one of the viable ones. That that presumes we're looking through ideas sequentially and thus are actually limited with the number of choices we're generally presented with, when really it's more akin to mapreduce when assessed at the level of leadership and domain experts. Between billions of people and thousands of domain experts, I don' think we're actually reaching any real limits, given the logarithmic nature of filtering good ideas in this way. What that means is that in effect there's a pool of solutions already filtered for viability and usefulness to choose from. All we're doing by entertaining the idea that we should stop looking for and presenting solutions is giving people an excuse to prevent ideas they don't agree with from being spread to others. The whole idea of one person thinking that because they think an idea doesn't have merit that others shouldn't see it is itself harmful. > while indeed it is the job of government to structure these decisions, recent history suggests that the opinions of the broader public still play a role in those decisions, that decisions are rarely made based purely off their technical merits, and that snake-oil salesmen can still do measurable harm to decision-making in representative governments. There are many possible reasons why we don't see progress from our leaders on certain issues. I'm not convinced that there being 1000 possible solutions is any worse than there being 10. I suspect that the same forces that prevent progress work the same in both cases, and if we're talking about less than 10 possible solutions, then I think we haven't examined the problem well enough. In any case, I don't think the problem of there being too many solutions and too much information is as obviously the problem as others are presenting it, and would want to see a real argument presented and defended before I would accept it at face value. That's not because I'm trying to be overly argumentative, just that as I noted, I just don't think it's as obvious a conclusion as others seem to.