This first line gets show in the index, even though it is in spoilers. This should probably be considered a bug.
I NEED MY SECOND INTERNET STAR.
Woah, this was hard! Took me longer than an hour, for the first time. The prompt felt very broad and less like it pushed the babble in a particular direction... it felt like it would probably have been best to be guided by curiosity, but I didn't feel much curiosity for the original post. I guess this is an interesting test cases in using babble challenge to help build more curiosity...
What are examples I've come across of institutions being cooperative or non-cooperative?
Some other ideas:
(I notice that for whatever reason I feel less in the mode of "it's okay to say something stupid, like 'bird in space suit'", and more like each thought has to be interesting! I feel stuck on hypotheses for what actually makes cooperative institutions successful, so let's try to attack that, starting with some stupid things...)
Cooperative institutions become successful if...
(This felt less like fun and more like work than earlier ones. And I emerge from it feeling dissatisfied: although I did in fact have 50 thoughts about stable cooperative institutions, and no doubt some of them are correct and some of them might even be clever, they are almost certainly all thoughts that other people have had and expressed before. The "sillier" challenges didn't have that feature, not least because they were asking questions that probably not so many people have spent long thinking about. My guess is that in order to actually contribute anything to the understanding of stable cooperative institutions I would need to sit and think in a less babbly fashion for quite a while, and do some actual research.)
1. This seems like something that would be more useful with actual expertise, which I don't have.
2. There are experiments that seem to suggest that countries we think of as having more and better such institutions tend to be countries whose people are more willing to trust one another (in "toy" econ/psych games) -- actual causation is highly nonobvious of course.
3. Maybe useful to look at this at the level of individuals -- what can an individual person do to make such institutions more or less likely to come into being and flourish, what incentives do they face, and what might change how they respond to those incentives?
4. First guess is that the main thing is the temptation for small individual "defections": one police officer taking a small bribe, one corporate executive taking a decision that's bad for the company but good for his annual bonus, etc.
5. "Better" countries/communities/... will be ones where (a) the incentives for such misbehaviour are weaker, perhaps because of stronger formal or informal enforcement or because somehow they're better at keeping incentives aligned, and/or (b) individuals' motivations are less selfish so that given bad incentives move them less.
6. People are not idealized consequentialists and often we just do whatever first occurs to us, what we did last time, what we have seen others do, etc. So to the above we should add (c) opportunities for misbehaviour are less salient, others seem to be less inclined to misbehave, etc.
7. Note that this may mean that _hypocrisy is good_: if everyone else seems to be doing the Right Thing then you will be more inclined to follow suit, even if in fact they're all secretly on the take and you just haven't noticed.
8. But of course _hypocrisy is bad_ too: it depends on whether the question is "behave badly for sure; admit it or not?" or "hide misbehaviour for sure; actually misbehave or not?".
9. I realise I've tacitly been thinking specifically of incentives rewarding "bad" behaviours, but surely there are also rewards for "good" ones: social approval, wealth from long-term success of employer, actually valuing whatever good the institution does, etc.
10. That bit about "long-term success" may be important. Suppose I am purely selfish and have no scruples, and I'm a CxO. I can embezzle a pile of money from the company and get away with it, but that will make it less likely to succeed. If I value getting _really_ rich in the longer term over getting _slightly_ rich in the shorter term, and if the prospect of longer-term success is real, I may choose not to embezzle.
11. Of course, if my individual embezzlement makes a negligible difference to the company's prospects I may do it after all, so there could be a tragedy-of-the-commons effect; perhaps incentive-crafting for good institutions needs quite different adjustments for individuals at different levels.
12. Seems like it would be worth looking at more and less successfully stable cooperative institutions in a single country / of a single type, looking for patterns.
13. Even better would be _experiments_ where we try to establish institutions in various ways and in various places and see what works and what doesn't, but that seems like it would be pretty much 100% impossible.
14. Some of the longest-lasting institutions are religions. Why?
15. To some extent many religions can be viewed as devices for encouraging cooperation and stability. (Cooperate, because the gods like it. Cooperate, because you are all part of a single whole. Don't change, because we already have the final revelation of all truth. ...)
16. Some other long-lasting institutions are nations.
17. Part of that is cheating: we tend to think of, say, "England" as having persisted for something on the order of a thousand years or so even though its exact boundaries have changed, it's more or less merged with Scotland and Wales and (Northern) Ireland, the way it's governed has changed radically, etc.
18. Part of it is that to some extent nations are discovered as well as created. E.g., England and the UK are not the same thing as either of the geographical entities Britain or Great Britain, but there's a close relationship; England and the UK are not the same thing as "where English is spoken natively", but again there's a close relationship.
19. One thing (related to those factors) that helps nations cohere is a sense of having common interests and purpose. Religions try to encourage this too.
20. Other institutions also try to foster that sense, both by making it true (e.g,, company bonus schemes and share options) and by trying to make people _feel like_ it's true (e.g., use of "family" language).
21. If we feel that stable cooperative institutions are suffering lately, could it be that people in them either _have_ fewer common interests or _feel_ fewer common interests?
22. It's a commonplace observation (and probably true) that there was a considerable shift towards individualism in the West during roughly the 20th century.
23. Note that we don't necessarily _want_ an unconditional increase in stability of cooperative institutions. Suppose the institution is the Mafia, the Nazi Party, a price-fixing cartel, a cult designed to exploit its members for profit.
24. When thinking about incentives, or paths of least resistance, we need to bear in mind indirect effects. How do we incentivize A (earlier) to set things up so that B (later) is incentivized to do what we want?
25. In some cases A and B may be the same people at different times, and these may be particularly difficult. How do we get politicians _now_ to arrange their procedures so that _later_ those same politicians will be difficult for lobbyists to corrupt? How do we get company executives _now_ to set up their compensation schemes so that _later_ they will act in the company's interests? Their own present incentives may point in the wrong direction for this already.
26. Arguably, the norm of _trying to establish and maintain stable cooperative institutions_ is itself a stable cooperative institution. So if there's some general decline, it may feed on itself. "Gradually, then suddenly."
27. This suggests that if we're in a situation where stable cooperative institutions are generally doing well and we want that to continue, we may want something like a zero-tolerance policy towards defection. Public outrage any time any politician shows even a hint of corruption. Instant termination for employees who aren't working hard.
28. _Explicit_ attempts to maintain stable cooperative institutions can backfire if they aren't felt to match the real underlying values. So once the rot starts, the only paths to fixing it involve decisive actions (e.g., hire/fire to make sure the people you have are actually aligned with your company's interests) and not just proclamations about what "our company culture is".
29. The stablest institutions might be ones where something about the nature of the institution itself automatically produces stability.
30. That's not necessarily a good goal to aim for; optimizing for one thing tends to worsen others and the same traits that make an institution super-stable may e.g. make it super-unresponsive to change. (Consider, e.g., the Roman Catholic Church which arguably has both those characteristics.) Still, it might be worth looking for such traits. Let's stick with our example of the Catholic Church, or maybe Christianity more generally.
31. Feature: making the persistence of the institution an explicit goal of the institution.
32. (Cf. Pournelle's "Iron Law of Bureaucracy".)
32. Feature: a near-watertight belief that the institution and its goals are supremely valuable.
33. In many places, Catholicism took a _really bad_ beating because of child-abuse scandals; does that near-watertight belief in the goodness of the institution have the consequence that when it does something too bad to ignore, people are more likely to give up on the institution wholesale than to try to amend it? (It sounds plausible but I'm not at all sure it's so; perhaps other institutions in which similar bad things happened would have suffered even more attrition.)
34. Feature: explicit, written beliefs, values, procedures, etc. (Harder for them to shift gradually, that way.)
35. Feature: the institution's official values are explicitly cooperative. (For-profit businesses may face a fundamental disadvantage here.)
36. I should note that I'm only _guessing_ that all these things are features that promote stability. The real causes might be other things entirely. But they seem like plausible stability-promoters.
37. Let's consider now another institution that we might _want_ to be stable, but that's been called into doubt in some places lately: democracy. It's lasted quite a while in many places. What helps its stability? What hurts?
38. Presumably voters like democracy because it gives them (or at least seems to give them?) some control over their rulers. (So one might hope that democracies are somewhat "stable against popular revolution". This seems like a big deal.)
39. Whoever's on top in _any_ system might be expected to like the system because it gives them power. (So maybe _any_ political system is somwhat "stable against attack from the top".)
40. So explicit attacks on democracy may come from (a) groups that hope for more power than democracy looks like giving them -- e.g., people/parties/... who have tasted power but are currently out and don't expect to win the next election, or people/parties/... who are in power but fear that they're about to lose it -- or (b) breakdowns in voters' trust that democracy actually does empower them.
41. As well as being overthrown outright, democracy can fail if it stops actually making power responsive to voters' wishes. E.g., falsified vote counts, all candidates having to be approved by some institution that has the _real_ power, enough corruption that you can't tell what a candidate will actually do without knowing who'll be bribing them, etc. These are probably bigger dangers, in most democracies, than overt abandonment of democracy.
42. (More generally: Stability of a cooperative institution isn't enough; it needs to remain genuinely cooperative, and that rather than stability as such may be the most important failure mode.)
43. One reason why democracy tends to persist is that nations and their people are _proud_ of having it and often make it a key part of their identity. Of course this leaves open failure modes where the name remains but the reality fades, but it does seem that _being something people are proud of_ is an advantage. This is not exactly surprising.
44. It doesn't feel as if I'm producing any very interesting or original thoughts. What happens if we reverse the question and ask: Suppose we have some cooperative institution and we want to _make it fail_ (either when designing it, or later); what would we do?
45. We could make it so that keeping it working requires people to do things that are very difficult or against their interests.
46. We could hire some smart and unscrupulous people and say "make this thing fail; I don't care how you do it".
47. We could design it to pit people against one another, and hope that the resulting acrimony will itself be destructive. (Since this seems like it describes both capitalism and democracy, and both are doing pretty well all things considered, maybe it wouldn't work.)
48. We could aim not to make it _die_ as such but to make it _unstable_, so that small changes tend to produce larger changes, in the hope that eventually that will kill it. So e.g. if the institution involves some sort of division of power (as e.g. the legislative/executive/judicial split or the House/Senate split or the state/Federal split, in the US), try to make sure that when one entity starts to get more power there are ways for it to parlay that into more more power. (Consider e.g. gerrymandering, court-packing, etc.)
49. The above has mostly been considering institutions whose existence is somehow formalized and explicit. Even, e.g., "democracy" cashes out in any particular place to a bunch of laws. Some important institutions aren't explicit in this way; consider e.g. mutual trust in a community (or a family or ...).
50. It seems like it makes an _enormous_ difference who's meant to be participating in an institution. Smaller, more homogeneous groups with more common interests cooperate better. Perhaps we can, and if so perhaps we should, somehow build larger-scale cooperation out of smaller-scale institutions where cooperation is easier?
1) Working complicated cooperative institutions are formed from many smaller institutions that work together.
2) People altruistically valuing the same thing can give rise to co-operation on a small scale, but I suspect it isn't so important at larger scales
3) Co-operation often arises because it is in the selfish interests of the co-operating entities.
4) People are probably genetically pre-disposed to co-operate with what they perceive as their ``tribe''.
5) My immediate reaction to this statement is that it isn't true: ``The magic that used to enable such cooperative institutions is fading.''
6) I agree that institutional cultures can help foster co-operation (or make it more difficult)
7) Peer pressure is a key factor in maintaining institutional cultures.
8) Government institutions have failed before, current situation doesn't look unique.
9) It is possible for things to get better.
10) Could the co-operation exhibited by ant colonies be regarded as an ant institution?
11) Maybe something like an institution is inevitable when multiple members of the same species live close to one another?
12) Things that look like an institutional failure from the outside may look like an institutional success from the inside. e.g. the bureaucracy may like expanding the bureaucracy even if no on else does.
13) Political institutions have become more inclusive over the 20th century, e.g. by giving women the vote.
14) At a high level could the entire human species be thought of as an institution?
15) The fall of early civilizations like Rome can be regarded as an early form of institutional failure.
16) Plenty of private sector institutions also fail. Most new companies tend to go bankrupt.
17) Government failings may be bigger/more noticeable than private sector ones because private sector ones tend to go bankrupt at an early stage, whereas the government has a large supply of cash to bail out failed projects.
18)Improvements in communications technologies have mad it easier for institutions to co-operate over larger scales.
19) re ``Most brand names seem to be less regarded than they used to be'' could probably be said at most times during the past.
20) Are subcultures much more difficult to form now, or is it that the ones that are forming are less visible to you?
21) The scale of the largest institutions has increased through time and will probably continue to do so.
22) Large institutions can insulate most of themselves from problems at the top, which limits the damage incompetent leaders can do.
23) Private sector institutions seem to be gaining in importance relative to public sector ones.
24) Governments may force some of the largest private sector institutions (google/amazon) to stop them become more powerful.
25) Multi national institutions will probably become more prevalent and more powerful.
26) We may get interplanetary institutions one day.
27) A pre historic tribe can be thought of as an institution, so in one form or another they have always been with us.
28) Short of going to live as a hermit in a remote location it is probably impossible to avoid being a member of an institution.
29) Relations between different institutions are a key part of what they do, and a major determinant of how successful they are.
30) Humans instinctively divide people into ingroup/outgroup. This probably helps define institutions as distinct entities.
31) Institutions that fail badly tend to cease to exist and be replaced by others. So at any time the existing institutions are the ones which have succeeded in co-operating at least to an extent. So we shouldn't be surprised that existing institutions tend to co-operate at least moderately well.
32) An institution can't form, or exist for any length of time without at least some co-operation and stability. So the title could be replaced with ``Where do institutions come from?''
33) Peoples knowledge of how to craft institutional cultures may be more instinctive and sub-conscious than conscious.
34) Perceived self interest is key. Rich and powerful people tend to set up incentives for others mainly because it is in their interest to do so, not out of altruism.
35) Fear of punishment (e.g. prison) for defecting against the institution can help keep them stable.
36) The post talks about trends originating on the West coast of the US, then spreading to the rest of the West, but what about Asia?
37) Asian nations will probably play a stronger role in shaping multi-national institutions in the future than western ones.
38) Maybe sub-cultures are harder to form now because most of the ones there is a need for have already formed?
39) Co-operative institutions are often not intentionally designed in the way they operate.
40) I wouldn't regard brands such as bitcoin or Amazon as being better regarded.
Mismanagement by government institutions is not limited to America. Some examples from Britain:
41) The government had ``Private Finance Initiatives'' which meant they didn't have to pay for infrastructure until a later date, and due to accounting tricks didn't have to be reported as government debt. If they had just borrowed the money up front and used that to pay for the infrastructure it would have cost less overall.
42) In the early part of the 21st century the government significantly increased the health budget to make themselves look good, unfortunately it was increased faster than it could easily absorb it, so a lot of it ended up being wasted on things that didn't improve health care.
43) Every government IT program seems to be delivered late and over budget. The government had to recently delay some changes to the law on same sex marriages because they couldn't change their IT systems in time, and they spent 10 billion pounds on an IT system for the health service that didn't work.
44) Strikes in the winter of 1978-79 disrupted key services such as gravedigging, rubbish collection, and some health services. Things got better after that, so it is possible to recover from failures.
45) In the 1990s Britain stopped building submarines for a while. When we started again loss of key skills meant the program was seriously delayed and went massively over budget, and the Americans had to be asked to help sort out this mess. It was later estimated that it would have been cheaper to build a couple of submarines during the gap to maintain the skill base than what actually happened.
46) Britain's handling of Covid 19 is an epic fail. The government seems to be reacting against short term pressure with no clear long term plan to handle it.
47) When the British government was building the next generation of air defence destroyers they changed the program so that fewer would be built for more money just because it would keep it within short term budgets.
48) When the British government was negotiating Brexit they gave away their best negotiating cards in the initial round of negotiations without getting anything meaning full in return.
49) The British government spent billions of pounds on a new maritime patrol aircraft, paid the company more money for providing fewer of them at a later date when they screwed up, then scrapped them just as they were about to become operational, then realised a few years later this left a major capability gap, and then decided to buy inferior aircraft from abroad as they no longer had confidence local industry could deliver.
50) A few years ago when a new pay package was being negotiated for some workers in the health system the government agreed to pay them a lot more for doing less just because the unions proposed it.
>! eating some space for preview purposes :::
**********************************************************************************
Root causes of cooperation
Why past was (or looks) better
Prompt is pretty open. Focusing on "content" is ambigious whether to reflecdt on things said on the main post or the topic they gesture towards. There is no choice paralysis challenge factor here.
Safeties
In babble mode it seems that my "touchy subject" censor is not working that much. It probably has a somewhat defencible reason to exist. Part of why it is okay to shut it off is that the output is taken into a space where the danger the module is built to repel is less relevant. I am kind of strained to think on the topi and beforehand thought I woudl reflect on the exercie afterwards. Seems that isn't viale right now and seems it will not be done if it doesn't happen now.
If you'd like something more constrained, consider the following prompts:
In a recent LessWrong question Anna Salamon asks “Where did stable, cooperative institutions come from (like bridges that stay up; the rule of law; or Google)?” She also worries that “the magic that used to enable such cooperative institutions is fading”.
Anna’s post is strong in babble. It does provide gears-level mechanisms and concrete hypotheses. But it also gestures at intuitions, felt senses, and insights-waiting-to-be-had.
This week’s challenge is simple: Have 50 thoughts about Anna’s post and questions.
Do you have a guess at the magic enabling human societies to build roads and postal services? Do you think institutions are actually getting stronger over time? What are 10 examples of how institutions changed in the last 100 years? Or 10 predictions about how they'll change in the future? Etc.
Your thoughts can be hypotheses, questions, anecdotes, confusions, disagreements, feelings... and so forth.
50 thoughts, no need for them to be longer than a sentence.
You have 1 hour.
Looking back
Here are the current rankings. (You gain a star for completing a challenge, and lose one for missing a week. I’m not including myself since it feels weird to be both gamemaster and participant.)
Great job everyone!
★★★★★ gjm
★★★★ Yonge
★★★ Tetraspace Grouping, Slider
★★ Mark Xu, Bucky
★ Turntrout, Harmless, Tao Lin, Daniel Kokotajlo, chasmani, supposedlyfun
Moving Forwards
This is week 6 of my 7-week babble sprint.
It is said that sufficiently advanced technology is indistinguishable from magic.
I think something similar is true for building skills.
There are some skills of which you can see the contours. You can squint and see yourself wielding them, with practice. And there are some things which seem like magic. As if though the kinds of humans who wield them are fundamentally different from the kind of human you are. There's no set of steps that could get you to where they are at.
Intellectual creativity often falls in this bucket.
For whatever reason, culture loves to create the vision of a genius. The media writes about “the 14-year old who climbed Mount Everest and wrote software for America’s largest bank” when in fact they made an impressive-for-their-age contribution to an open source package and camped out at a lower base station reachable by walking.
Maybe because creativity is so illegible. It seems that there is nothing. And then there’s an idea. George Orwell said of writing that it was like being “driven on by some demon whom one can neither resist nor understand.”
It’s especially illegible from the outside.
It will often happen to me that I read a LessWrong post. Full of brilliant, interesting, novel thoughts; and with a bustling comment section. And faced with this Tower of Babble I take a peak at what my own brain generates, 2 seconds after being hurled into the spotlight —
nothing
— and I despair.
I feel like I don’t have ideas. Like I am a person who does not have ideas.
But I miss that Tower’s are built one stone at a time. Once, where that great obelisk rests, there was only wind.
I’ve recently been meditating on this, trying to feel this truth in my bones:
I am a machine. One that turns time and food and air into creativity. And machines are in the domain of Science. They are understandable, extendable, lawful.
Now that I’ve done 5 weeks of the Babble Challenge, this is becoming clearer. I can choose to have ideas. I’m getting a better sense of the gears that turn to produce my creativity, I see motion where before there was only fog and magic.
If you’ve also looked at LessWrong threads and felt they were the playing fields of wizards, I also want you to have this experience. I want you to feel like a machine who, with ambition and deliberate practice, can learn to turn time into ideas.
Rules
Think about the ideas and the questions, not the spelling or word choice. Think about institutions and cooperation, not about paragraph length and sentence structure. Try to engage with the substance, rather than the symbol.
Any answer must contain 50 ideas to count. That’s the babble challenge.
However, the 1 hour limit is a stretch goal. It’s fine if it takes longer to get to 50.
This is really important. Sharing babble in public is a scary experience. I don’t want people to leave this having back-chained the experience “If I am creative, people will look down on me”. So be generous with those upvotes.
If you comment on someone else’s post, focus on making exciting, novel ideas work — instead of tearing apart worse ideas.
I've often found that 1 great idea can hide among 10 bad ones. You just need to push through the worse ones. Keep talking. To adapt Wayne Gretzky's great quote: "You miss 100% of the ideas you never generate."
If you spend 5 min agonising over not having anything to say, you’re doing it wrong. You’re being too critical. Just lower your standards and say something, anything. Soon enough you’ll be back on track.
This is really, really important. It’s the only way I’m able to complete these exercises.
---
Now, go forth and Babble!
50 thoughts on the question about stable, cooperative institutions!