Towards optimal play as Villager in a mixed game

Link post

On Twit­ter, Freyja wrote:

Things cap­i­tal­ism is trash at:
Valu­ing prefer­ences of any­thing other than adults who earn money (i.e. fu­ture peo­ple, non-hu­mans)
Pric­ing non-stan­dard­is­able goods (i.e. in­for­ma­tion)
Play­ing nicely with non-quan­tifi­able val­ues + ob­jec­tives (i.e. love, rit­ual)
Things cap­i­tal­ism is good at:
In­cen­tivis­ing the pro­duc­tion of novel goods and ser­vices
Co­or­di­nat­ing large groups of peo­ple to pro­duce com­plex bun­dles of good­s
The ob­vi­ous: mak­ing value fungible
Any­one know of work on -
a) in­te­grat­ing the former into ex­ist­ing eco­nomic sys­tems, or
b) de­vel­op­ing new sys­tems to provide those things while in­clud­ing cap­i­tal­ism’s ex­ist­ing benefits?

This in­ter­sected well enough with my cur­rent in­ter­ests and those of the peo­ple I’ve been dis­cours­ing with most closely that I figured I’d try my hand at a quick ex­pla­na­tion of what we’re do­ing, which I’ve lightly ed­ited into blog post form be­low. This is only a loose sketch, I think it does rea­son­ably pre­cisely out­line the ar­gu­ment, but many read­ers may find that there are sub­stan­tial in­fer­en­tial leaps. Ques­tions in the com­ments are strongly en­couraged.

Any se­ri­ous at­tempt at (b) will first have to un­wind the dis­in­for­ma­tion that claims that the thing we have now is cap­i­tal­ism, or re­motely effi­cient.

The short ver­sion of the pro­ject: learn­ing to talk hon­estly within a small group about how power works, both sys­tem­i­cally and as it ap­plies to us, with­out try­ing to hold onto in­for­ma­tion asym­me­tries. (There’s per­va­sive temp­ta­tion to with­hold poli­ti­cal in­for­ma­tion as part of a zero-sum priv­ilege game, like Plato’s philoso­pher-kings.)

Some back­ground: post-WWII elite in­sti­tu­tions (e.g. corps) are com­pet­i­tive to en­ter, but not un­der perfor­mance pres­sure, be­cause of US gov­ern­ment policy. This strongly se­lects for zero-sum games, which mimic but wreck dis­course. (See Mo­ral Mazes for more, es­pe­cially the case stud­ies that make up most of the book, start­ing around chap­ter 3.)

This cre­ates op­por­tu­nity in two ways.

First, in­sti­tu­tions are mostly too stupid to model their en­vi­ron­ment be­yond the zero-sum games they spe­cial­ize in, so a small group that’s able to main­tain in­for­ma­tion hy­giene and not turn on each other should be able to take & hold ter­ri­tory. “And not turn on each other” turns out to be re­ally hard, be­cause all our role mod­els and in­tu­itions for how to sur­vive in this world in­volve do­ing that all the time. But we’re learn­ing!

(A mun­dane ex­am­ple of a de­ci­sive ad­van­tage due to in­for­ma­tion hy­giene: Paul Gra­ham writes about how his startup did bet­ter be­cause it used an el­e­gant pro­gram­ming lan­guage. That’s only in­for­ma­tion hy­giene on the purely tech­ni­cal level, but that was enough to out­ma­neu­ver huge cor­po­ra­tions with a strong per­ceived in­cen­tive to ruin them, for quite a while. For a less mun­dane ex­am­ple, the story of how Elisha out­ma­neu­vered mul­ti­ple rul­ing dy­nas­ties is a per­sonal fa­vorite − 2 Kings 5-10. The nar­ra­tive dis­torts the “mir­a­cles” a bit but it’s not hard to re­con­struct how he ac­tu­ally did it.)

Se­cond, be­cause most sup­posed pro­duc­tive ac­tivity is done in the con­text of huge sta­ble cor­po­ra­tions, peo­ple are try­ing to max­i­mize the num­ber of jobs and com­plex­ity per unit of out­put. This im­plies that many things can be done much more eas­ily.

So that im­plies that if we can have good enough in­for­ma­tion hy­giene and group co­he­sion not to fall vic­tim to the per­verse im­pulse to do the kind of make-work or ar­tifi­cial scarcity that cre­ates much of cost dis­ease, we can learn how to build a nearly full-stack civ­i­liza­tion in a small city-state. Ob­vi­ously there are many steps be­tween here and there, but since lots of them in­volve get­ting col­lec­tively smarter, a de­tailed plan would be in­ap­pro­pri­ate.

What does good in­for­ma­tion hy­giene and group co­he­sion look like? The game Were­wolf is a good ex­am­ple. Play­ers are se­cretly as­signed the iden­tity of Villager (ini­tially the ma­jor­ity) or Were­wolf (minor­ity). Each round all play­ers vote one player out, and Were­wolves se­cretly do the same. There are other de­tails that al­low villagers to make some in­fer­ences about who the were­wolves are. But they have to play the first few rounds right or they lose.

Op­ti­mal play for Were­wolves in­volves (a) tar­get­ing whichever villagers are the most helpful to pub­lic de­liber­a­tion, for ex­clu­sion, and (b) dur­ing pub­lic de­liber­a­tion, be­ing as un­helpful as they can get away with while ap­pear­ing to try to help at other times. I re­al­ized a lot of things about how so­cial skills feel from the in­side when I fi­nally figured out how to play cor­rectly as a Were­wolf.

Op­ti­mal play for Villagers in­volves cre­at­ing as much clar­ity as pos­si­ble, as soon as pos­si­ble, and be­ing will­ing to as­sume that peo­ple who seem to be fool­ishly gum­ming up the works are Were­wolves if there’s no other clear tar­get.

With op­ti­mal play, Villagers usu­ally win, but in prac­tice, at best one or two peo­ple try to cre­ate clar­ity and are picked off in the first round by the Were­wolves. The other Villagers are re­signed to try­ing to die last, so they lose.

The thing I said about elite cul­ture fa­vor­ing zero-sum games can be re­cast as: the so­cial en­vi­ron­ment fa­vors play­ing Were­wolf over play­ing Villager. In case it’s not ob­vi­ous, op­ti­mal real-world play for Villagers can of­ten in­volve leav­ing the Were­wolves alone. In real life there are bet­ter things to do than mur­der your en­e­mies, like hang out. Villagers just need to defend them­selves if and when they’re ac­tu­ally threat­ened.

We’re try­ing to learn how to play the Villager strat­egy suc­cess­fully, in a con­text where we’ve mostly been ac­cul­turated to play as Were­wolves, es­pe­cially among elites. This has to in­volve figur­ing out how to do in­ter­per­sonal fault anal­y­sis (iden­tify when peo­ple are be­ing Were­wolfy) with­out scape­goat­ing (as­sum­ing that fault → blame → ex­clu­sion).

In other words, jus­tice seeks truth, but in­tends to leave no one be­hind; peo­ple who can’t con­tribute need to feel safe ad­mit­ting that, and peo­ple who hurt the group need the op­tion to re­pent & heal the breach.

We don’t have great fi­nesse yet but op­ti­mal play in our world seems to be some fluid in­te­gra­tion of talk­ing about poli­tics, heal­ing per­sonal trauma, and in­ter­sub­jec­tive open­ness.

Havel’s The Power of the Pow­er­less de­scribes a similar (but less self-aware) strat­egy which he calls “dis­si­dence.” He (ac­cu­rately, I think) pre­dicts that the situ­a­tion in Cap­i­tal­ist coun­tries will be more difficult than the situ­a­tion in Com­mu­nist ones, be­cause Cap­i­tal­ist ide­ol­ogy is more per­sua­sive be­cause it’s more plau­si­bly true.