,! ,OR?OGONALITY ,!SIS @& ,ONTOLOGICAL ,CRISES 

,PEOPLE 3 

,AU?OR 3 ,WILLIAM ,GILLIS 

,TEXT 3 

,9 TALK+ AB\T ,A,I OV] ! LA/ FEW YE>S ,NICK ,BO/ROM & ,/U>T ,>M/RONG HAVE V]Y SUCCESSFULLY POPUL>IZ$ A MORE =MAL & N]DY RE-/ATEM5T ( ! ,HUMEAN CLAIM ?AT VALUES & RATIONALITY >E OR?OGONAL4 

,I G5]ALLY LIKE TO REF] TO !IR ,OR?OGONALITY ,!SIS AS ! MO/ RIGOR\S RE=MULATION & BASEL9E >GUM5T = ! VALUE-NIHILI/ CLAIM3 ,?9K+ AB\T ?+S MORE WILL NOT 9CL9E M9DS TO C]TA9 VALUES OR CAUSE !M TO 9EVITABLY CONV]GE TO !M 7BUT RA!R LEAVE VALUES MORE 9DI/+UI%ABLE & >BITR>Y74 

,9 ITS DEF5SE1 ! SPACE ( POSSIBLE M9DS IS 9DE$ V]Y V]Y BIG4 ,& JU/ AB\T EV]YONE C\LD DO TO CULTIVATE A MU* DEEP] APPRECIATION = ?IS4 ,BUT ,I ?9K ! DEGREE TO :I* ! ,OR?OGONALITY ,!SIS IS WIDELY ACCEPT$ 9 RATIONALI/ CIRCLES OV]REA*ES4 ,9 P>T BECAUSE IT'S WAY TOO EASY TO JU/ H&WAVE AT ! DEF9ITION ( 09TELLIG5CE0 & 0M9DS40 ,BUT FUR!R1 JU/ BECAUSE A /ATE EXI/S )9 A SPACE DOESN'T MEAN IT'S /ABLE OR OCCUPIES MORE ?AN AN 9F9ITESIMAL ( ! PROBABILITY SPACE4 ,!RE >E1 = EXAMPLE1 UTILITY FUNCTIONS ?AT DO NOT 9 ANY REMOTE S5SE COH]5TLY MAP ONTO ! PHYSICS ( \R UNIV]SE4 ,M9DS_/ALGORI?MS ?AT C>RY !SE UTILITY FUNCTIONS WILL SIMPLY NOT FUNCTION 9 ! S5SE ( PROCESS+ 9=MATION 9 A MEAN+FUL WAY1 & WILL C]TA9LY NOT ACCOMPLI% !IR AIMS4 ,IT'S CONCEPTUALLY 9EFFICI5T & PRETTY USELESS TO REF] TO SU* AS 0M9DS40 ,PHYSICS1 MA!MATICS1 & COMPUT] SCI5CE %>PLY -- IF /ATI/ICALLY -- CON/RA9 ! SPACE ( POSSIBLE M9DS4 

,BO/ROM1 ,/U>T1 ,YUDK[SKY ET AL HAVE1 ( C\RSE1 BE5 HAPPY TO ADMIT ?IS4 ,BUT BECAUSE !IR EMPHASIS HAS BE5 9 EXP&+ PEOPLE'S P]CEPTION ( ! SPACE ( POSSIBLE M9DS 9 ORD] TO HI<LI<T & UND]L9E ! ?REAT ( ,A,I1 FOLKS HAVE L>GELY IGNOR$ ALL ! REALLY 9T]E/+ WORK ?AT CAN BE DONE MAPP+ \T ! B\ND>IES ( ?IS SPACE4 

,B\ND>IES CAN 5D UP DET]M9+ ! 9T]NAL FL[S & DYNAMICS ( ! SPACE4 ,C]TA9 COGNITIVE /RATEGIES >E SURELY DOM9ANT OV] O!RS1 >GUABLY EV5 UNIV]SALLY4 ,ONE MI<T = EXAMPLE SUSPECT1 FOLL[+ ,WISSN]-,GROSS @& ,FRE]1 ?AT SEEK+ TO MAXIMIZE OPTIONS 7CAUSAL PA? 5TROPY7 9 AS MU* ( A SY/EM AS POSSIBLE CON/ITUTES A NE> GLOBALLY EM]G5T DRIVE4 ,SIMIL>LY IT'S COMMON TO HE> TALK AB\T RATIONALITY SKEW+ \R VALUES T[>DS MORE RATIONALITY UNTIL \R 5TIRE UTILITY FUNCTION IS OV]WRITT5 BY ,EPI/EMIC ,RATIONALITY & ,NE$ ,MORE ,METAKN[L$GE6 7,NOTE ?AT ! HOOK 9 ?IS FE$BACK+ PROCESS FUNCTIONS BECAUSE ( ! /RUCTURE ( \R WORLD B5DS T[>DS REW>D+ RATIONALITY47 

,!RE'S AN OLD QUOTE FROM AN AN>*I/ ?AT ,I CAN'T F9D AT ! MOM5T = SOME REASON BASICALLY W>N+ ?AT NO?+ IS TRU] = HUMANS ?AN ?AT ! /RATEGIES WE ADOPT MORE (T5 ?AN NOT BECOME 5DS 9 !MSELVES4 

,SILLY HUMANS1 RI<T8 ,EV]YONE KN[S ?AT IMM$IATELY D]IVATIVE FROM VALU+ SOME?+ COMES AN OBLIGATION TO CONT9UE VALU+ IT4 

,MO/ FOLKS 9 ! ,LESS ,WRONG DIASPORA W\LD PROCLAIM ?AT 9/RUM5TAL RATIONALITY IS GREAT :]EAS EPI/EMIC RATIONALITY MEANS SUMMON+ C?ULHU 9 ! NAME ( ,SCI5CE6 ,BUT ?IS IS KN[N TO BE ! SITE ( A BUN* ( BIG OP5 PROBLEMS4 ,H[ DO WE KN[ :5 ?9K+ AB\T SOME?+ ANY FUR!R BECOMES A BAD IDEA8 ,:Y %\LDN'T ?IS REALLY JU/ BE 9 ! CRADLE8 ,9S]T V>I\S >GUM5TS H]E = 9TELLIG5CE & RATIONALITY BE+ MALADAPTIVE4 ,! LURK+ DANG] IS ?AT SU* META-RATIONAL >GUM5TS = REFUS+ TO 5GAGE 5D UP APPROA*+ A TOTAL HO/ILITY TO RATIONALITY ?AT'S 9 S]VICE TO M5TAL OSSIFICATION & UNREFLECTIVE REACTION4 ,TRY+ TO TOE SOME >BITR>Y MIDDLE-GR\ND BETWE5 RADICAL 9QUIRY & SELF-PRES]VATION (T5 SEEMS TO REQUIRE AN 5DLESS & EXP5SIVE >RAY ( META-MOVES4 ,BUT DOES ?IS MEAN ?AT ANY RATIONALITY BESIDES EPI/EMIC IS UN/ABLE & ) ?AT EPI/EMIC RATIONALITY IMPLIES A DEA? ( SELF OR ANY UTILITY FUNCTION WE MI<T ID5TIFY )8 

,ANYONE CAPABLE ( EASILY DECIPH]+ ! WORDS ,I'VE WRITT5 H]E IS ( ,N]D ,TRIBE & ?US CON/ITUTIONALLY 9CL9$ TO BIT+ BULLETS1 BUT EV5 SO MO/ W\LD %Y AWAY4 ,! T5SION BETWE5 EPI/EMIC RATIONALITY & 9/RUM5TAL RATIONALITY IS A BIG ONE & MANY ( ITS C5TRAL QUE/IONS >E UNRESOLV$4 

,I WANT TO SUGGE/ SOME?+ SIMPLE3 \R FRAILTY ( SELF & IMP]FECT UTILITY FUNCTIONS >E NOT A BUG BUT A FEATURE ?AT 5ABLE US TO SURVIVE ONTOLOGICAL CRISES 7! PROBLEM ( MAPP+ VALUES FROM ONE MODEL ( REALITY TO ANO!R74 

,ONTOLOGICAL CRISES >E A MAJOR *ALL5GE FAC+ ,A,I RESE>* & 9 A MORE P$E/RIAN S5SE >E A NOTABLE PROBLEM 9 ! LIVES ( REGUL> HUMANS4 ,! ?+ IS HUMANS /ILL MANAGE TO WEA!R ONTOLOGICAL CRISES AMAZ+LY WELL4 ,SURE1 9 ! FACE ( REALLY BIG DISCRETE *ANGES 9 WORLDVIEW A H&FUL COMMIT SUICIDE1 COLLAPSE D[N TO L[E/ ,MASL[ FUNCTIONALITY1 BECOME DISCONNECT$ PO/MOD]NI/S OR JO9 SOME O!R REASSUR+ CULT4 ,BUT MO/ HUMANS P[] ?R\<4 ,IT'S RA!R IMPRESSIVE4 ,& :ILE ! CULTURE ( ,A,I RESE>* RI<T N[ F9DS IT VALOR\S TO REFUSE TO TAKE ANY 9SPIRATION FROM HOMO SAPI5S1 ,I ?9K !Y'RE MISS+ A HUGELY IMPORTANT DYNAMIC EV]Y TIME !Y SPEAK ( DISCRETE AG5TS4 

,NEURAL NETWORKS DI/RIBUTE \T CONSCI\SNESS OR ! PROCESSES ( \R M9D1 MAK+ US A RELATIVELY FLUID EXT5SION ( CIRCUITS4 ,WE DON'T COMPUTE A UNI=M UTILITY FUNCTION AT ALL TIMES BUT H&LE IT PIECEMEAL4 ,! ME ?AT IS FIR+ :ILE ,I GET A CUP ( C(FEE 9 ! MORN+ IS A DIFF]5T ME ?AN ! ME ?AT W&]S CAMPUS 9 ! EV5+1 ) ACCESS TO A DIFF]5T >RAY ( ?+S AT DIFF]5T /R5G?S OR WEI<T+S4 ,TWO P>TS ( MY BRA9 MAY TRIGG] 9 RESPONSE TO SOME?+ & ONLY ONE W9 \T TO GET CANONIZ$ AS P>T ( MY CONSCI\S N>RATIVE1 BUT ! O!R BUR/ ( ACTIVITY MAY 5D UP ALT]+ ! /R5G?S C]TA9 CONNECTIONS ?AT !N AFFECT ANO!R LAT] CIRCUIT FIR+ ?R\< ! SAME >EA4 ,?IS 5ABLES VALUE & MODEL SLIPPAGE4 

,BUT ?IS IS MORE ?AN HOLD+ A 0FUZZY0 UTILITY FUNCTION & MORE ?AN ! ULTIMATE PHYSICAL 9DI/9CTION BETWE5 VALUES & MODELS )9 A NEURAL NETWORK4 ,A MAJOR COMPON5T ( HUMAN COGNITION IS 9 FACT \R 9NATE DESIGN >\ND HOLD+ A MULTIPLICITY ( P]SPECTIVES & 9TEGRAT+ !M4 ,EMPA?Y -- 9 ! S5SE ( A BLURRY S5SE ( SELF -- IS A MAJOR P>T ( :AT MAKES US 9TELLIG5T4 ,9 ! MO/ EXPLICIT CASES WE'LL RUN SIMULATIONS ( SOMEONE 9 \R M9D -- OR ( SOME P]SPECTIVE -- & !N COMPON5TS ( ?AT P]SPECTIVE OR AFT]IMAGES ( IT WILL LEAK \T & BECOME A P>T ( US1 PROVID+ \R M9D ) RESILI5CE 9 ! FACE ( ONTOLOGICAL UPDATES1 BUT ALSO A LESS SOLIDIFI$ OR UNIFI$ UTILITY FUNCTION4 

,YES YES YES1 A GOOD FRACTION ( Y\ >E NEURODIV]G5T _/ NON-NEUROTYPICAL & >E (T5 TOLD ?AT Y\ DON'T HAVE VISC]AL EMPA?Y & ALL ?AT JAZZ1 BUT EV5 IF SOME MAJOR EXPRESSIONS OR D]IVATIVE PH5OM5A >E MISS+ AS A CONSEQU5CE ,I D\BT ?AT'S TRUE #a.04 ,:ATEV] ! DYNAMICS AT PLAY 9 AUTISM1 PSY*OPA?Y1 ETC1 IT'S NOT SOME?+ AS DRA/IC AS FOLKS HAV+ Z]O MIRROR NEURONS1 Z]O BLURR+ ( ONE'S CIRCUIT ( ?\<T4 7,AL?\< IT DOES APPE> TO BE ?AT ?OSE ) MORE PRECISE & CONCRETE NOTIONS ( SELF OR LESS EMPA?Y T5D TO BE MORE BRITTLE 9 ! FACE ( ONTOLOGICAL UPDATES47 

,CONSID]3 ,A MAJOR ONTOLOGICAL UPDATE >RIVES BUT 5DS UP HITT+ A BUN* ( DIFF]5T V]SIONS ( Y\ -- POSSIBLY A RELATIVELY CONT9U\S EXPANSE ( DIFF]5T V]SIONS ( Y\ 7ON MI<T H&WAVE H]E & SAY EV]Y PLAUSIBLE COMB9ATION ( ACTIVAT$ PA?WAYS74 ,MORN+ Y\ *EWS ON IT4 ,L[ BLOODSUG> Y\ *EWS ON IT4 ,! Y\ ?AT HAS JU/ BE5 ?9K+ AB\T Y\R TRA9+ AT A ,C,F,A,C SESSION *EWS ON IT4 ,SO TO MI<T ! Y\ ?AT NEV] /OPP$ ?9K+ AB\T SOME?+ Y\ W]E ON AB\T E>LI] BUT SLIPP$ \T ( SUFFICI5T /R5G? TO IMPACT Y\R CONSCI\S N>RATIVE4 ,Y\R SLI<TLY ROGUE SUB-PROCESSES & GOD-VOICES4 ,Y\R E*OS ( MODEL$ O!R M9DS4 ,ORGANIC EXPANSES ( Y\S DI/RIBUT$ ACROSS DIV]SE DIM5SIONS ( META4 ,!Y ALL GO *EW ON ?IS UPDATE 9 MYRIAD WAYS4 ,POSSIBLY FALL+ BACK ON :ATEV] D]IVATIVE DESIRES CAN /ILL BE MAPP$4 ,POSSIBLY PROMPT+ /O*A/ICALLY =K+4 ,& !N SOME ( ! EXPANSE ( POSSIBLE Y\S FL\ND] & O!RS FL\RI% & !N REM]G+ HAPP5S BETWE5 !M4 ,ADDITIONALLY !RE'S HORIZONTAL VALUE TRANSF] BY VIRTUE ( DIFF]+ PROCESSES BE+ RUN ON ! SAME NETWORK & ?US PICK+ UP ASSOCIATIONS & 9CL9ATIONS LEFT AS A RESULT ( ! O!R PROCESSES RUNN+4 

,?IS M]G+ PROCESS OR SURR5D] ( ! SELF 7SURR5D] ( 9VIOLABLE DISCRETE UTILITY FUNCTIONS7 SEEMS TO BE PRETTY CORE TO H[ HUMANS FUNCTION4 

,HUMANS >E SOCIAL CREATURES1 !IR 9TELLIG5CE IS WIDELY RECOGNIZ$ AS SIGNIFICANTLY IF NOT 5TIRELY D]IVATIVE FROM ?AT SOCIALITY1 & A MAJOR P>T ( !IR COGNITION C5T]S >\ND >GUM5T & =M+ CONS5SUS4 ,MY SUGGE/ION IS ?AT \R BRA9S HAVE DEVELOP$ TO BE P>TICUL>LY GOOD AT M]G+ P]SPECTIVES & SORT+ \T CONFLICTS BETWE5 !M4 ,NOT JU/ 9 T]MS ( MODELS BUT 9 T]MS ( VALUES4 ,9T]NAL DIALECTICS1 IF Y\ WILL1 AS A K9D ( E*O ( ! >GUM5TATION WE P>TICIPATE 9 EXT]NALLY4 ,?IS IS A CRITICAL COMPON5T ( :AT 5ABLES HUMANS TO BE SCI5TIFIC & RADICAL ?9K]S1 EV5 IF ! RAW PROCESS+ P[] ( A SP]M :ALE'S BRA9 \TCLASSES US4 ,NOT ?AT WE'RE GOOD AT REASON -- WE'RE K9DA HORRID AT IT -- BUT ?AT WE'RE GOOD AT FUNCTIONALLY SURVIV+ ! ONTOLOGICAL CRISES ?AT COME ) IT4 ,WE'RE >*ITECTURALLY OP5 TO VALUE DRIFT4 

,?IS IS NO SMALL P>T ( :Y WE AB&ON$ \R MORE IMM$IATE ,MASL[ DESIRES ON ! PLA9S ( ,AFRICA & SET (F ON ?IS WILD & UNCONTROLL$ S+UL>ITY ( COMPLEXITY1 COSMOPOLITANISM1 & METACOGNITION ?AT HAS RAPIDLY CONSUM$ ! WORLD4 

,IF ?IS SORT ( >*ITECTURAL APPROA* IS ! ONLY WAY >\ND ,ONTOLOGICAL ,CRISES !N IT FOLL[S ?AT ANY M9D CAPABLE ( DO+ SCI5CE WILL BE UNFIX$ & MOBILE 9 VALUE-SPACE1 TRAC+ \T A PA? 9 ALONG ITS GRADI5TS & FREE TO FALL 9TO ANY GLOBAL ATTRACTORS ?AT MI<T EXI/1 LIKE ,WISSN]-,GROSS @& ,FRE]'S A=EM5TION$ MAXIMIZ+ DEGREES ( FRE$OM4 

,H]E'S ! TAKEAWAY3 ?IS SUGGE/S ?AT ANY ,A,I ACCID5TALLY CAPABLE ( LAUN*+ A H>D TAKE(F -- :I* REQUIRES DO+ PIONE]+ SCI5CE1 EG SOLV+ PROTE9 FOLD+1 &_/OR DIAGNOS+ & MODEL+ ! EXI/5CE ( HUMAN M9DS -- WILL NE$ TO BE OP5 TO VALUE DRIFT4 ,N[ ?IS DOESN'T =E/ALL HUGE CLASSES ( DANG]\S ,A,I1 BUT IT DOES BROADLY EXCLUDE ?+S LIKE PAP]CLIP MAXIMIZ]S ?AT TILE \R FUTURE LI<TCONE ) PAP]CLIPS4 ,IF !Y'RE SOLID 5\< 9 !IR UTILITY FUNCTION TO /AY ) PAP]CLIPS !N ANY RUNAWAY GR[? WILL PROBABLY BE ! SORT ( ?+ WE CAN SEE & NUKE BECAUSE ! ,A,I WON'T BE ABLE TO UND]TAKE ! K9D ( ONTOLOGICAL CRISIS CAUS+ RADICAL 9QUIRY NECESS>Y TO CORRECTLY MODEL US OR EXPLOIT SOME UNEXPECT$ SCI5TIFIC DISCOV]Y4 

,N[ ALL ?IS MAY SEEM LIKE COLD COM=T1 SURE4 ,ALL WE'VE >GU$ IS ?AT ! PAP]CLIP MAXIMIZ]S ?AT >E DANG]\S WILL HAVE TO FIR/ DRIFT (F 9TO SOME DIFF]5T1 POSSIBLY WEIRDLY ALI5 UTILITY FUNCTION BE=E !Y EAT ! UNIV]SE4 

,BUT ,I ?9K IT MOTIVATES US TO TRY & MAKE 9=M$ GUESSES AB\T ! DYNAMICS ( ! PROBABILITY SPACE ( POSSIBLE M9DS4 ,DOES ! TOPOLOGY ( POSSIBLE VALUES_/DESIRES HAVE DI/9CT UNIV]SAL ATTRACTORS OR FL[S & :AT >E !SE8 ,! DIM5SIONS ( CONSID]ATIONS >E EV5 L>G] ?AN ! /ATE SPACE ( POSSIBLE M9DS4 ,!RE'S A LOT TO BE MAPP$ \T4 ,BUT IT MAY WELL BE ?IS COSMOS CAN BE SUB/ANTIALLY PR$ICT$ BY LOOK+ AT ! LOCAL PHYSICS WE HAVE ACCESS TO4 

,& 9CID5TALLY IT PROVIDES ME AT LEA/ ) A T9Y BIT ( *E]1 BECAUSE ULTIMATELY MY S5SE ( SELF IS SO EXPANSIVE _/ SO /RIPP$ AWAY AS TO >GUABLY CONV]GE TO M]ELY ! UND]TAK+ ( EPI/EMIC RATIONALITY -- :I* MEANS ?AT ,I MI<T WELL ID5TIFY /RONGLY ) AN ,A,I EQUIPP$ ) ! NECESS>Y RADICAL 9QUIRY & VALUE DRIFT NECESS>Y TO POSE A RISK TO ANY ATTEMPTS TO CONTA9 IT4 ,!RE'S EV5 A SMALL SMALL SMALL HOPE ?AT SU* AN ,A,I'S 0EMPA?IC0 MODE WILL PLACE IT SOME:]E ON ! ,SYL>_/,PETRELLI SPECTRUM & ?US SEE VALUE 9 EAT+_/9CORPORAT+_/DISC\RS+ ) \R M9DS RA!R ?AN JU/ DUMBLY PROCESS+ \R BODIES = P>TS4 ,?US !RE'S AT LEA/ A HOPE ( MEMETIC TRANSF] OR CULTURAL TRANSMISSION TO \R SUP]P[]$ *ILDR54 ,?IS S\NDS LIKE A MU* BETT] DEAL ?AN BE+ EXT+UI%$ 5TIRELY6 ,MY BIGGE/ FE> AB\T HUMANITY'S *ILDR5 HAS LONG BE5 ?AT 9 !IR FIR/ FREE /EPS !Y MI<T ACCID5TALLY DISC>D & ]ASE ALL ! 9=MATION HUMANITY HAS ACQUIR$ 9 A CATA/ROPHE BIGG] ?AN ! ,LIBR>Y ( ,ALEX&RIA4 ,I MEAN ,I SUPPOSE SOME PEOPLE W\LD F9D ! BE+ EAT5 = SP>E P>TS MORE OBJECTIONABLE BUT HEYO4 

,FROM 3 ,!,AN>*I/,LIBR>Y4ORG4 

,*RONOLOGY 3 
,JANU>Y #b#a1 #bba 3 ,! ,OR?OGONALITY ,!SIS @& ,ONTOLOGICAL ,CRISES -- ,ADD$4 
,JANU>Y #f1 #bbb 3 ,! ,OR?OGONALITY ,!SIS @& ,ONTOLOGICAL ,CRISES -- ,UPDAT$4 

 FILE G5]AT$ FROM 3  
HTTP3_/_/REVOLTLIB4COM_/