Hacker News Board Hacker News, browsed like a classic message board
Live data from Hacker News

The Problem of Thinking Too Much (2003) [pdf]

statweb.stanford.edu

Post reply on HN
Print view View original

41–50 of 84 posts

Re: The Problem of Thinking Too Much (2003) [pdf]

#41

Post by AstralStorm »

tachyonbeam wrote: post #38
thisoneworks wrote: post #17

Thinking should be a time boxed activity, with time proportional to the priority of the thing at hand. As a classic overthinker I've realized my overthinking is a weird amalgamation of perfectionism and procrastination. That took me a long time to accept, since I've always rationalized my overthinking as ultimately a good thing. If you assume that action (instead of thinking) is the primary unit in the world, then yo…

I found it also useful to realize that there are many situations in life where there are just too many unknowns. You're faced with two possible futures: should I go to University A or B? This choice is necessarily going to radically alter your life: depending which university you pick, you'll make different friends, date different people, hang out in different spots, maybe live in different cities. Based on this choi…

The thing is, and we do not collectively teach young, is that many decisions are workable but some have bad consequences that are nigh irredeemable. Then there's the sunk cost fallacy, all too common. Estimating the cumulative sunk costs and arriving at robust decisions is feasible.

Choosing a college may be a cost, but for now there are workable alternatives.

And you always have the way to know or at least roughly estimate which is good enough for a sane decision. You just take the cost of acquiring information into your decision time and you're gold. Apply hard limit for time and problem disappears. You did the best you could given the constraints.

In fact, the typical problem today is that you have too many ways to know and not enough ways to verify veracity.

Re: The Problem of Thinking Too Much (2003) [pdf]

#42

Post by lukevdp »

throwayEngineer wrote: post #14

This is a very Academia driven problem. They have been so detached from problem solving that they forget how to problem solve. Modeling a human randomness with math didn't give you perfect data and conclusions? What was the original problem? To solve something unsolvable with our current understanding of biology? Industry asks better questions and finds solutions. Academia waddles around in theoreticals, and makes mi…

Academia is a process of finding new knowledge and sharing it.

Industry is a process of commercialising added value.

They both add to society in different ways, and their outcomes aren’t really comparable.

Re: The Problem of Thinking Too Much (2003) [pdf]

#43

Post by Bartweiss »

mercer wrote: post #37
Bartweiss wrote: post #27

Earlier quoted context omitted.

This was a hugely helpful insight for me. First, because I found it much easier to accept "deliberating has a cost, factor that in" than a generalized claim of "stop thinking so much". Second, because thinking more really does help sometimes. "Go with your gut" doesn't sit well with me; every once so often I think about a decision and notice that I've made some major oversight in my first assessment. Aiming for the '…

I think going with your gut can be a fine solution, but it's probably worth assessing how well your gut has served you historically. For me, my gut has generally been more right than my over-analytical mind. But I have a number of friends where I think it's the other way around.

This definitely seems like a "competing needs" thing. People repeat the advice that helps them, but different people need to adjust in different directions. "Look before you leap" versus "he who hesitates", I suppose.

There's an interesting result from studying people's prediction accuracy in which they improve simply by being asked to think about their answers and whether they're sure. For me, overthinking often leads me to explain away 'irrational' reactions that were quite accurate, but taking at least one extra look to say "has my first reaction missed something fundamental" is a lifesaver.

Re: The Problem of Thinking Too Much (2003) [pdf]

#44

Post by kekebo »

gowld wrote: post #34
kekebo wrote: post #21

Earlier quoted context omitted.

Mindfulness type exercises / learning internal silence was the only effective strategy i found. I can still get carried away in my thoughts but there's a learned recurring return into silence that also gives room for meta cognitions of various sorts. The more I "practice" the more subconscious those silent moments appear, the higher their frequency and the easier it is to maintain longer stretches of them - and the m…

Doea that help you notice , or help you get better at suppressing it via practice and "muscle memory"?

The more quiet moments there are to reflect upon my thinking, the easier it becomes to notice and correct. I wouldn't call it suppressing, trying to apply 'force' in cognition usually leads to more internal conflict instead of resolution, at least in my experience.

Re: The Problem of Thinking Too Much (2003) [pdf]

#45

Post by vokep »

smallnamespace wrote: post #36

Is 'thinking about thinking' correctly actually an unsolvable problem? Let's say I'm trying to do X. - Level 0 might be: do the first thing that comes to mind - Level 1 may be 'consider your possible choices and do a cost-benefit analysis' Most people would say that Level 1 is preferable, but only for 'weighty' decisions where the cost of thinking is outweighed by the benefit of making a better decision. For example,…

Or maybe that is just what level 2 would be? Level 1 is doing cost-benefit analysis on all problems. Level 2 might be doing cheap meta analysis to determine if energy can be saved and a level 0 solution used, or if more thinking is required. I would think there is a top level, which is full memorization of the space in which thinking might occur. Than instead of any thinking taking place, a reference is made to this memory-truth-table and the correct action is produced in response. Such a truth-table sounds maybe a bit paradoxical though. Does it take into account its own presence in the situation?

Re: The Problem of Thinking Too Much (2003) [pdf]

#46

Post by thisoneworks »

tachyonbeam wrote: post #38
thisoneworks wrote: post #17

Thinking should be a time boxed activity, with time proportional to the priority of the thing at hand. As a classic overthinker I've realized my overthinking is a weird amalgamation of perfectionism and procrastination. That took me a long time to accept, since I've always rationalized my overthinking as ultimately a good thing. If you assume that action (instead of thinking) is the primary unit in the world, then yo…

I found it also useful to realize that there are many situations in life where there are just too many unknowns. You're faced with two possible futures: should I go to University A or B? This choice is necessarily going to radically alter your life: depending which university you pick, you'll make different friends, date different people, hang out in different spots, maybe live in different cities. Based on this choi…

Fascinating. Thanks for sharing. It reminds me of Sylvia Plath's fig tree - [https://www.goodreads.com/quotes/7511-i-saw-my-life-branchin...] which I always think about. I've observed that it's useful to think of the one future that you would sacrifice all other futures for. Which would be what exactly do you value?

On regrets and mistakes, Jordan Peterson talks about this sort of thing in a beautiful way as the idea of "redemptive mistakes", and that "people are a lot more unhappy when they look back in their lives about the things they didn't do than they are about the mistakes they made while they were doing things" - [https://www.youtube.com/watch?v=ZwGDnSWmqhM]. It has stuck with me as I try to be more permissive of myself making mistakes and being imperfect.

Re: The Problem of Thinking Too Much (2003) [pdf]

#47

Post by openfuture »

throwayEngineer wrote: post #14

This is a very Academia driven problem. They have been so detached from problem solving that they forget how to problem solve. Modeling a human randomness with math didn't give you perfect data and conclusions? What was the original problem? To solve something unsolvable with our current understanding of biology? Industry asks better questions and finds solutions. Academia waddles around in theoreticals, and makes mi…

This is hopelessly narrow minded.

I won't defend all of academia but it is clear that sometimes there's a cache of 'useless knowledge' that suddenly gets cashed in when industry reaches a point where it is actually needed (most famous example being number theory and cryptography).

The whole point of academia is for it to be a space for the free exploration of unknown intellectual domains. This exploration will often be worthless but may occasionally bear fruit in completely civilization changing ways.

Meanwhile, industry is too busy manipulating the evaluation criteria to make problems magically go away to actually solve anything (this is the counter strawman to your strawman).

Seriously though, our econonic systems do not seem to be able to evaluate survival as an important concept. There is no evidence that industry will move by itself in a direction that sacrifices profits to avoid pollution, someone else will just say he did the same and offer a cheaper product.

What we want is for our political systems to be more robust and for the gradient between academia and industry to be more smooth. What we have is broken in too many ways to count.

Re: The Problem of Thinking Too Much (2003) [pdf]

#48

Post by earthboundkid »

taneq wrote: post #22
js8 wrote: post #5

I have realized some time ago that when some decision is difficult to make, it is likely that both outcomes are equally good (or bad), or at least, equally indeterminable. If one of them was much better, then the decision would be obvious! So indeed, if you can't decide, spin a penny. You save time and the outcome will be similar.

The advice my mum gave me was similar: If you don't know which way to choose, flip a coin, and see how you feel about the outcome . That will tell you what you really wanted to choose.

A variant on that which I will use when no coin is handy is to assign odd and even to whatever options and then look at the time. If you haven't looked at the time recently, the final digit is random to you.

Re: The Problem of Thinking Too Much (2003) [pdf]

#49

Post by AstralStorm »

thisoneworks wrote: post #46
tachyonbeam wrote: post #38

Earlier quoted context omitted.

I found it also useful to realize that there are many situations in life where there are just too many unknowns. You're faced with two possible futures: should I go to University A or B? This choice is necessarily going to radically alter your life: depending which university you pick, you'll make different friends, date different people, hang out in different spots, maybe live in different cities. Based on this choi…

Fascinating. Thanks for sharing. It reminds me of Sylvia Plath's fig tree - [ https://www.goodreads.com/quotes/7511-i-saw-my-life-branchin... ] which I always think about. I've observed that it's useful to think of the one future that you would sacrifice all other futures for. Which would be what exactly do you value? On regrets and mistakes, Jordan Peterson talks about this sort of thing in a beautiful way as the id…

This is in no way an endorsement of doing without thinking, which will catch up with you and result in you being forced to skip options. At least I hope it isn't.

There's a clear spectrum between recklessness on one end, indecision and inaction on the other and doing right things right in the middle, spontaneously or in a planned way.

Relevancy is key, nobody remembers things that are irrelevant by definition as a regret... Cost is a kind of relevancy.

Going back to the article, you almost completely lost contact with the original problem once you're past third or rarely fourth order thinking because you will lack enough particular data to decide on any measure of success, with exception of particularly structured recursive problems. Welcome to decision variant of Goedel's second. Decisions and axioms (data points) are countable.

Re: The Problem of Thinking Too Much (2003) [pdf]

#50

Post by yeahitslikethat »

gowld wrote: post #28
Bartweiss wrote: post #26

Earlier quoted context omitted.

Feynman tells an intriguingly similar story from another angle. When he moved from Cornell to Caltech, he had a sort of Buridan's Ass situation where each one would beat the other's offer and leave him with no way to make a decision. Even the weather wouldn't cooperate, with the snow driving him out of Cornell and smog driving him out of Caltech. He eventually landed at Caltech, basically arbitrarily, and realized th…

Young people are experimental and open to novelty. Old people are conservative and prefer the status quo. The combination selects for a robust society. That's an evo-bio trope

Is because of memory. Old people remember how things were and want those memories to be continuously valuable.

If everything changes they have nothing over the younger competition.

Post reply on HN