The world is not an engineering problem • Hacker News Board

Re: The world is not an engineering problem

#61

Post by MereInterest » Fri, Nov 09, 2018, 5:18 AM UTC

pmarreck wrote: post #7

MereInterest wrote: post #4

Earlier quoted context omitted.

Taking that argument to its logical conclusion, there should be no public policy whatsoever. Every single policy has at its roots an evidenceless belief. As an extreme example, "Seat belts save lives" is an objective truth, but "Saving lives is a good thing" is an evidenceless belief. That is it believed by the vast majority of the human race in no way constitutes evidence. Evidenceless beliefs are what need to dicta…

Good point technically but I think some subjective valuations are widespread enough to be able to be considered objective, such as "alive/animate things are better than dead/inanimate things", and "joy is superior to pain", and "murdering is bad". And regarding pain/pleasure, you can always do what doctors do and ask patients to assign the intensity a number from 1-10, which is basically a hack that takes a qualitati…

I think the distinction is a useful one to have, because it lets you identify the appropriate way to try to convince somebody. If people disagree about the effectiveness of policies, and which ones will bring about the desired effects, then additional evidence should be found and examined to determine who is correct. On the other hand, if people disagree about what the goal is, then no amount of evidence will convince them, because goals are unobservable.

Re: The world is not an engineering problem

#62

Post by natalyarostova » Fri, Nov 09, 2018, 5:42 AM UTC

TeMPOraL wrote: post #33

comonad-colaboy wrote: post #24

Earlier quoted context omitted.

Science is about verifiable proofs (verifiable within reasonable time limits) which somewhat means the same thing as formal proofs in the grand scheme of things. Correct me if I am wrong Edit: forgot to put in time limits

You're wrong. Science is about building and refining models (theories) in order to make them match observable reality as close as possible, in order to use those models to predict what happens in the future - both by itself, and in consequence of us poking stuff. That's what it means to know "how something works". Engineering takes these models and adds a "what's the best way to poke things to achieve a desired outco…

Mathematics is no different. Our brains are simply computers that verify that the system evaluates based on it's rule system. In fact, you can bootstrap an empirical verification of any mathematical proof holding if sufficient human brain-computes evaluate it, and determine it holds.

Re: The world is not an engineering problem

#63

Post by marchenko » Fri, Nov 09, 2018, 9:30 AM UTC

TeMPOraL wrote: post #44

Like hell it isn't. At the risk of going "no true scottsman", I hate this kind of anti-intellectual nonsense. The methods of science, engineering and rational reasoning are the best tools we have for solving world problems. They were literally invented for it. They're reified applied reasoning. You eschew them at your own peril. Now of course to solve problems involving humans, you have to factor in the human element…

I don't think this essay is anti-intellectual at all. I think it is an exploration of the longstanding philosophical problem of trying to rationally design a world for irrational beings. The excerpt from Dillow on consequentialism brings to mind an idea from James Scott: that elites control policy discussions by limiting the accepted terms of the argument to legible, rationalizable factors - generally economic and technocratic variables visible from a top-down view. This can place serious limits on our access to the store of human knowledge embedded in emotion and practice. What we consider "irrational" impulses are sometimes the products of generations of evolutionary tuning that would seem to be feats of engineering brilliance if they came with the proper documentation. Meta-rationality includes exploring the bounds and the weaknesses of the tools in our contemporary rationalist arsenal, as well as searching under the couch cushions of irrationality for the occasional bit of value.

Re: The world is not an engineering problem

#64

Post by TeMPOraL » Fri, Nov 09, 2018, 9:57 AM UTC

claudiawerner wrote: post #58

TeMPOraL wrote: post #44

Like hell it isn't. At the risk of going "no true scottsman", I hate this kind of anti-intellectual nonsense. The methods of science, engineering and rational reasoning are the best tools we have for solving world problems. They were literally invented for it. They're reified applied reasoning. You eschew them at your own peril. Now of course to solve problems involving humans, you have to factor in the human element…

>The methods of science, engineering and rational reasoning are the best tools we have for solving world problems. How do you feel about the fact that was is rational and irrational (indeed our whole meta-rational basis) has changed in time according not to new discoveries but merely the movement of the cultural zeitgeist[0]? What about the differences between instrumental reason (and rationality), technological rati…

The way I see it, rationality doesn't choose your values for you, the same way mathematical reasoning doesn't choose your axioms for you. It's not a fixed tool in a sense that we're still discovering its various methods, but it is a tool that should not depend on cultural context - because what works to achieve your goal or correctly predict consequences doesn't change even as your goal might.

Or, in other words, the principle of "garbage in, garbage out" applies.

Re: The world is not an engineering problem

#65

Post by ajuc » Fri, Nov 09, 2018, 11:58 AM UTC

> For my part, I prefer things a little messy because not only are the solutions so often dependent on coercion but they also require that the ordinary citizen's faith and feelings are denied. Maximising utility seems a good thing but it is not the main reason why people do things like set up business, create charities, build village halls, paint, sing, create or innovate. Technocracy treats the world as an engineering problem when it's an unfolding story, explorers in a dense jungle not white-coated scientists in a laboratory.

There's a very famous fragment of Polish romantic ballad "Romantyczność" by Adam Mickiewicz:

    Feeling and faith stronger speaks to me,
    than the eye and the glass of a wise man.

It was written in 1821 in occupied Poland, and (together with other similar literature) was responsible for creating nationalist romantic movement that caused several failed uprisings, countless deaths, and whole generations of educated patriots being forced to migrate abroad escaping repressions.

When Poland got independent in 1918 (mostly thanks to a good luck and WW1, just like other countries in the region) - this attitude claimed the success ("if not for failed uprisings we wouldn't be here"), and people believed it. Failed uprisings are celebrated to this day, and a few rational generals who wanted to prevent the useless massacre and were hanged because of that - are still considered traitors.

Then it caused governments of interwar Poland to pursue unrealistic and opportunistic strategy that resulted in 1/6th of the population and 1/7th of the territory being lost in WW2 despite supposedly being on the winning side. But it sure felt nice to be brave and be the "first to fight". People are still boasting "Poland - first to fight" like it's a good thing to be stupid and die for no reason.

This national romanticism is still very much defining the public debate in Poland together with the only mainstream alternative - positivism and pragmatism. And romanticism is still winning - 200 years later. We've only got +-25 years of pragmatism after 1989, but it's over now.

It's why populists can win elections - because people want to ignore reality and stop analyzing it. "Just do what feels right, it will be OK for sure." "Winning trade wars is so easy". Everything is easy if you ignore reality because it's too complex.

It's a very harmful attitude. Don't let it take over your culture, it's very hard to get back to the enlightenment once you leave it behind.

Re: The world is not an engineering problem

#66

Post by naasking » Fri, Nov 09, 2018, 12:16 PM UTC

MereInterest wrote: post #60

naasking wrote: post #46

Earlier quoted context omitted.

> "Saving lives is a good thing" is an evidenceless belief. Is it? Any kind of value system must presuppose the value of life because only life can hold something valuable. Value systems that do not are therefore inconsistent.

I don't see any inconsistency there. For example, I could imagine a value system that solely values the existence of paper clips. Actions that increase the number of paper clips in the universe are good, while actions that decrease the number of paper clips in the universe are bad. Life has value only in as much as it causes paper clips to exist. It would be a rather silly value system for a human to have, but there…

If paper clips are the highest good, then ensuring that paper clips will be made in perpetuity is also of utmost importance. Only life is capable of knowing the goodness of paper clips, ergo we should ensure life's continuance so as to ensure the continued production of paper clips.

With the invention of machines we could replace life with more efficient automated systems that produce paper clips, but stupid machines cannot adapt to changing circumstances, like outside threats in the form of asteroids, climate change, earth quakes, and so on. Therefore the machines should be at least as intelligent as possible so as to adapt to any circumstances and ensure the continued production of paper clips.

Intelligent machines are simply another form of life. This obviously wouldn't happen overnight, but with the gradual introduction of more and more technology, the creators would more and more integrate with their machines.

Any value system that denies the value of life is simply inconsistent.

Re: The world is not an engineering problem

#67

Post by specialist » Fri, Nov 09, 2018, 1:01 PM UTC

TeMPOraL wrote: post #44

Like hell it isn't. At the risk of going "no true scottsman", I hate this kind of anti-intellectual nonsense. The methods of science, engineering and rational reasoning are the best tools we have for solving world problems. They were literally invented for it. They're reified applied reasoning. You eschew them at your own peril. Now of course to solve problems involving humans, you have to factor in the human element…

You're a better person than me. Since you responded, I tried to read the OC. Word salad.

To your point, I now believe all design problems are human problems, which sometimes involve math and science.

I read Neal Postman's Technopoly and Donald Norman's Design of Everyday Things around the time I was figuring out that computers weren't delivering on their promises in the areas I care about (CAD for A/E/C, education), and in some cases made things worse. Validation for what I was experiencing, observing.

I credit Postman and Donald Norman for my conversion from technophile to humanist.

I don't have anything to say about rationality. I gave up (forfeited) after the 2016 USA elections.

Re: The world is not an engineering problem

#68

Post by claudiawerner » Fri, Nov 09, 2018, 2:01 PM UTC

TeMPOraL wrote: post #64

claudiawerner wrote: post #58

Earlier quoted context omitted.

>The methods of science, engineering and rational reasoning are the best tools we have for solving world problems. How do you feel about the fact that was is rational and irrational (indeed our whole meta-rational basis) has changed in time according not to new discoveries but merely the movement of the cultural zeitgeist[0]? What about the differences between instrumental reason (and rationality), technological rati…

The way I see it, rationality doesn't choose your values for you, the same way mathematical reasoning doesn't choose your axioms for you. It's not a fixed tool in a sense that we're still discovering its various methods, but it is a tool that should not depend on cultural context - because what works to achieve your goal or correctly predict consequences doesn't change even as your goal might. Or, in other words, the…

Isn't rationality the principle of GIGO, though? Let's say I have a garbage idea, but my methods for acheiving it are rationally constructed - isn't this even a little problematic? Surely me must judge not only the means (and ignore any aspect of the goal) but to look and examine whether the goal is a good too.

But if your claim is that rationality is independent of culture (which, as I've already noted, is rubbish, since the idea of what is rational and irrational changes with cultural zeitgeist, unless you have some claim to objective rationality which I don't) then we can't use a static methods of rationality to judge the goodness of a particular goal, which is obviously context-dependent. So if not rationality, then what? I suggest critical thinking in place.

Re: The world is not an engineering problem

#69

Post by ajuc » Fri, Nov 09, 2018, 2:16 PM UTC

compiler-guy wrote: post #10

pmarreck wrote: post #2

Public policy should only be based on objective truths and not evidenceless beliefs or feelings. It's one thing to be for or against marijuana usage; it's another to still have a law on the books banning its use and sale when a growing plethora of science says it's far safer than substances already legalized and regulated (such as alcohol) and actually therapeutic in some cases (such as cancer). It's one thing to be…

"Damming this river will hurt the surrounding ecosystem" and "Damming this river will provide power for thousands of people" are both objective truths. The only way to mediate between them is by discussing non-objective things like how we value fish and the services provided by electricity. Do we value keeping fish alive more than keeping old people alive via power-hungry air conditioning? There are millions of simil…

> The only way to mediate between them is by discussing non-objective things like how we value fish and the services provided by electricity.

No, it's not the only way. It's the easy way.

The hard way is to consider other consequences of both choices, including long-term effects, including changes that aren't 0%-100%, but slightly change probability of other events.

Hurting the ecosystem might increase the probability of flooding the area, which will hurt some of the people that you are trying to comfort by giving them cheaper energy. So, maybe it's objectively irrational to build the dam, no matter whether you value people or fish more?

Maybe the population in the area is decreasing anyway and cheap power won't help? We can predict such things pretty reliably. So why build a huge dam and hurt environment if in 40 years it will be useless?

It's impossible to know for sure, but it's possible to estimate. Abandoning the rationality altogether just because it's not immediately obvious what to do is very irresponsible.

I know it was just an example, but IMHO it's a representative example. Many of the issues in the center of the public debate now are objectively scientifically solved, or so close to be solved, that it doesn't matter what the exact details are.

Yet people continue to argue about them because of tribal thinking, religion, corporate interests, etc. See global warming, equality for sexual orientations, war on drugs, public vs private healthcare, etc.

OC proposed that human element is inherently unpredictable, so rationality isn't the solution. Yet - thanks to the law of big numbers - big groups of people are more predictable than small groups of people or individual people. So we may not know who will vote X or Y, but we can predict how many people will vote X and Y with reasonably small error. We have big data, that can predict you're pregnant before you know you're pregnant, just basing on you browser history FFS.

Re: The world is not an engineering problem

#70

Post by gglitch » Fri, Nov 09, 2018, 2:50 PM UTC

marchenko wrote: post #63

TeMPOraL wrote: post #44

Like hell it isn't. At the risk of going "no true scottsman", I hate this kind of anti-intellectual nonsense. The methods of science, engineering and rational reasoning are the best tools we have for solving world problems. They were literally invented for it. They're reified applied reasoning. You eschew them at your own peril. Now of course to solve problems involving humans, you have to factor in the human element…

I don't think this essay is anti-intellectual at all. I think it is an exploration of the longstanding philosophical problem of trying to rationally design a world for irrational beings. The excerpt from Dillow on consequentialism brings to mind an idea from James Scott: that elites control policy discussions by limiting the accepted terms of the argument to legible, rationalizable factors - generally economic and te…

Thanks for the succinct account of irrational impulses, and for the James Scott ref. Can you recommend any particular book of his? My local library appears to have _Against the Grain_ on hand.