Jailbroken Cleverness 1

A mostly-jailbroken friend said this, and it illustrates the sort of thing you can think if you’ve worked through this blog’s tech and know that everyone is trying to pwn you:


so like the cistem wants to gunk up our epistemic systems about cissexism and prevent us from talking with each other, forming class consciousness / common knowledge about the structure of cissexism, and acting as a group against oppression. they want to route all information processing through cissexist nodes, like all information about “abuse” is encouraged to be routed through the REACH abuse panel. and the abuse panel is cissexist and has no transfems in it, it has not banned anna salamon for her cissexism and allowance for robert lecnik to run events after Jay told her what had happened and robert confessed. and then went on to sexually assault another transfem (me). 

it is a common heirarchy structure of having one node for infoprocessing and try and gunk up communication with “they are so nice to transfems, did you know many transfems consider them wonderful, they use perferred pronouns” to discourage lateral communication among transfem nodes. why talk with other transfems about experiences of cissexism when as far as the infoprocessing system presents you are an anomaly and others have not noticed this?

one tool against systems clogging up infoprocessing systems is to not listen to specific claims like “we cant be transphobic, we use perferred pronouns” and select for yourself at random a proxy of cissexism, make a prediction using it and see how it turns out. this is what i did with employment numbers, i dont think the cissexist systems thought that people would look at this so i would make the prediction “given the kind of place REACH is, i bet the abuse panel doesnt have any transfems” and then check. now people complicit in cissexism know people checking for it will look at employment numbers, so if they want to protect their status as non-cissexist while not actually becoming non-cissexist they will goodhart employment numbers like people like anna goodhart pronouns despite thinking of transfems as violent men. so the next round of checking this in an anti-inductive battle against an enemy that wants to obscure its optimization is to choose another metric at random, predict how the metric will turn out in a world where the org is cissexist and then check to see how it worked out in reality.

anon

In particular, this part is very good:

one tool against systems clogging up infoprocessing systems is to not listen to specific claims like “we cant be transphobic, we use perferred pronouns” and select for yourself at random a proxy of cissexism, make a prediction using it and see how it turns out. this is what i did with employment numbers, i dont think the cissexist systems thought that people would look at this so i would make the prediction “given the kind of place REACH is, i bet the abuse panel doesnt have any transfems” and then check.

I think the organizations in question had like 0 out of 40 people working there who were transfem, and the community they exclusively recruit from is like 15-20% transfems.