Competent Elites

Fol­lowup to: The Level Above Mine

(Any­one who didn’t like yes­ter­day’s post should prob­a­bly avoid this one.)

I re­mem­ber what a shock it was to first meet Steve Jurvet­son, of the ven­ture cap­i­tal firm Draper Fisher Jurvet­son.

Steve Jurvet­son talked fast and ar­tic­u­lately, could fol­low long chains of rea­son­ing, was fa­mil­iar with a wide va­ri­ety of tech­nolo­gies, and was happy to drag in analo­gies from out­side sci­ences like biol­ogy—good ones, too.

I once saw Eric Drexler pre­sent an anal­ogy be­tween biolog­i­cal im­mune sys­tems and the “ac­tive shield” con­cept in nan­otech­nol­ogy, ar­gu­ing that just as biolog­i­cal sys­tems man­aged to stave off in­vaders with­out the whole com­mu­nity col­laps­ing, nan­otech­nolog­i­cal im­mune sys­tems could do the same.

I thought this was a poor anal­ogy, and was go­ing to point out some flaws dur­ing the Q&A. But Steve Jurvet­son, who was in line be­fore me, pro­ceeded to de­mol­ish the ar­gu­ment even more thor­oughly. Jurvet­son pointed out the evolu­tion­ary trade­off be­tween viru­lence and trans­mis­sion that keeps nat­u­ral viruses in check, talked about how greater in­ter­con­nect­ed­ness led to larger pan­demics—it was very nicely done, de­mol­ish­ing the sur­face anal­ogy by cor­rect refer­ence to deeper biolog­i­cal de­tails.

I was shocked, meet­ing Steve Jurvet­son, be­cause from ev­ery­thing I’d read about ven­ture cap­i­tal­ists be­fore then, VCs were sup­posed to be fools in busi­ness suits, who couldn’t un­der­stand tech­nol­ogy or en­g­ineers or the needs of a frag­ile young startup, but who’d got­ten ahold of large amounts of money by dint of seem­ing re­li­able to other busi­ness suits.

One of the ma­jor sur­prises I re­ceived when I moved out of child­hood into the real world, was the de­gree to which the world is strat­ified by gen­uine com­pe­tence.

Now, yes, Steve Jurvet­son is not just a ran­domly se­lected big-name ven­ture cap­i­tal­ist. He is a big-name VC who of­ten shows up at tran­shu­man­ist con­fer­ences. But I am not draw­ing a line through just one data point.

I was in­vited once to a gath­er­ing of the mid-level power elite, where around half the at­ten­dees were “CEO of some­thing”—mostly tech­nol­ogy com­pa­nies, but oc­ca­sion­ally “some­thing” was a pub­lic com­pany or a siz­able hedge fund. I was ex­pect­ing to be the youngest per­son there, but it turned out that my age wasn’t un­usual—there were sev­eral ac­com­plished in­di­vi­d­u­als who were younger. This was the point at which I re­al­ized that my child prodigy li­cense had offi­cially com­pletely ex­pired.

Now, ad­mit­tedly, this was a closed con­fer­ence run by peo­ple clue­ful enough to think “Let’s in­vite Eliezer Yud­kowsky” even though I’m not a CEO. So this was an in­cred­ibly cherry-picked sam­ple. Even so...

Even so, these peo­ple of the Power Elite were visi­bly much smarter than av­er­age mor­tals. In con­ver­sa­tion they spoke quickly, sen­si­bly, and by and large in­tel­li­gently. When talk turned to deep and difficult top­ics, they un­der­stood faster, made fewer mis­takes, were read­ier to adopt oth­ers’ sug­ges­tions.

No, even worse than that, much worse than that: these CEOs and CTOs and hedge-fund traders, these folk of the mid-level power elite, seemed hap­pier and more al­ive.

This, I sus­pect, is one of those truths so hor­rible that you can’t talk about it in pub­lic. This is some­thing that re­porters must not write about, when they visit gath­er­ings of the power elite.

Be­cause the last news your read­ers want to hear, is that this per­son who is wealthier than you, is also smarter, hap­pier, and not a bad per­son morally. Your reader would much rather read about how these folks are over­worked to the bone or suffer­ing from ex­is­ten­tial en­nui. Failing that, your read­ers want to hear how the up­per ech­e­lons got there by cheat­ing, or at least smarm­ing their way to the top. If you said any­thing as hideous as, “They seem more al­ive,” you’d get lynched.

But I am an in­de­pen­dent scholar, not much be­holden. I should be able to say it out loud if any­one can. I’m talk­ing about this topic… for more than one rea­son; but it is the truth as I see it, and an im­por­tant truth which oth­ers don’t talk about (in writ­ing?). It is some­thing that led me down wrong path­ways when I was young and in­ex­pe­rienced.

I used to think—not from ex­pe­rience, but from the gen­eral memetic at­mo­sphere I grew up in—that ex­ec­u­tives were just peo­ple who, by dint of su­pe­rior charisma and butt-kiss­ing, had man­aged to work their way to the top po­si­tions at the cor­po­rate hog trough.

No, that was just a more com­fortable meme, at least when it comes to what peo­ple put down in writ­ing and pass around. The story of the hor­rible boss gets passed around more than the story of the boss who is, not just com­pe­tent, but more com­pe­tent than you.

But en­ter­ing the real world, I found out that the av­er­age mor­tal re­ally can’t be an ex­ec­u­tive. Even the av­er­age man­ager can’t func­tion with­out a higher-level man­ager above them. What is it that makes an ex­ec­u­tive? I don’t know, be­cause I’m not a pro­fes­sional in this area. If I had to take a guess, I would call it “func­tion­ing with­out re­course”—liv­ing with­out any level above you to take over if you falter, or even to tell you if you’re get­ting it wrong. To just get it done, even if the prob­lem re­quires you to do some­thing un­usual, with­out any­one be­ing there to look over your work and pen­cil in a few cor­rec­tions.

Now, I’m sure that there are plenty of peo­ple out there bear­ing ex­ec­u­tive ti­tles who are not ex­ec­u­tives.

And yet there seem to be a re­mark­able num­ber of peo­ple out there bear­ing ex­ec­u­tive ti­tles who ac­tu­ally do have the ex­ec­u­tive-na­ture, who can thrive on the fi­nal level that gets the job done with­out re­course. I’m not go­ing to take sides on whether to­day’s ex­ec­u­tives are over­paid, but those ex­ec­u­tive ti­tles oc­cu­pied by ac­tual ex­ec­u­tives, are not be­ing paid for noth­ing. Some­one who can be an ex­ec­u­tive at all, even a be­low-av­er­age ex­ec­u­tive, is a rare find.

The peo­ple who’d like to be boss of their com­pany, to sit back in that com­fortable chair with a lovely golden parachute—most of them couldn’t make it. If you try to drop ex­ec­u­tive re­spon­si­bil­ity on some­one who lacks ex­ec­u­tive-na­ture—on the the­ory that most peo­ple can do it if given the chance—then they’ll melt and catch fire.

This is not the sort of un­pleas­ant truth that any­one would warn you about—at least not in books, and all I had read were books. Who would say it? A re­porter? It’s not news that peo­ple want to hear. An ex­ec­u­tive? Who would be­lieve that self-valu­ing story?

I ex­pect that my life ex­pe­rience con­sti­tutes an ex­tremely bi­ased sam­ple of the power elite. I don’t have to deal with the ex­ec­u­tives of ar­bi­trary cor­po­ra­tions, or form busi­ness re­la­tion­ships with peo­ple I never se­lected. I just meet them at gath­er­ings and talk to the in­ter­est­ing ones.

But the busi­ness world is not the only venue where I’ve en­coun­tered the up­per ech­e­lons and dis­cov­ered that, amaz­ingly, they ac­tu­ally are bet­ter at what they do.

Case in point: Pro­fes­sor Rod­ney Brooks, CTO of iRobot and former di­rec­tor of the MIT AI Lab, who spoke at the 2007 Sin­gu­lar­ity Sum­mit. I had pre­vi­ously known “Rod­ney Brooks” pri­mar­ily as the pro­moter of yet an­other dread­ful nou­velle paradigm in AI—the em­bod­i­ment of AIs in robots, and the for­sak­ing of de­liber­a­tion for com­pli­cated re­flexes that didn’t in­volve mod­el­ing. Definitely not a friend to the Bayesian fac­tion. Yet some­how Brooks had man­aged to be­come a ma­jor main­stream name, a house­hold brand in AI...

And by golly, Brooks sounded in­tel­li­gent and origi­nal. He gave off a visi­ble aura of com­pe­tence. (Though not a thou­sand-year vam­pire aura of ter­rify­ing swift perfec­tion like E.T. Jaynes’s care­fully crafted book.) But Brooks could have held his own at any gath­er­ing I at­tended; from his aura I would put him at the Steve Jurvet­son level or higher.

(In­ter­est­ing ques­tion: If I’m not judg­ing Brooks by the good­ness of his AI the­o­ries, what is it that made him seem smart to me? I don’t re­mem­ber any stun­ning epipha­nies in his pre­sen­ta­tion at the Sum­mit. I didn’t talk to him very long in per­son. He just came across as… formidable, some­how.)

The ma­jor names in an aca­demic field, at least the ones that I run into, of­ten do seem a lot smarter than the av­er­age sci­en­tist.

I tried—once—go­ing to an in­ter­est­ing-sound­ing main­stream AI con­fer­ence that hap­pened to be in my area. I met or­di­nary re­search schol­ars and looked at their poster­boards and read some of their pa­pers. I watched their pre­sen­ta­tions and talked to them at lunch. And they were way be­low the level of the big names. I mean, they weren’t visi­bly in­com­pe­tent, they had their var­i­ous re­search in­ter­ests and I’m sure they were do­ing pass­able work on them. And I gave up and left be­fore the con­fer­ence was over, be­cause I kept think­ing “What am I even do­ing here?”

An in­ter­me­di­ate stra­tum, above the or­di­nary sci­en­tist but be­low the or­di­nary CEO, is that of, say, part­ners at a non-big-name ven­ture cap­i­tal firm. The way their aura feels to me, is that they can hold up one end of an in­ter­est­ing con­ver­sa­tion, but they don’t sound very origi­nal, and they don’t sparkle with ex­tra life force.

I won­der if you have to reach the Jurvet­son level be­fore think­ing out­side the “Out­side the Box” box starts to be­come a se­ri­ous pos­si­bil­ity. Or maybe that art can be taught, but isn’t, and the Jurvet­son level is where it starts to hap­pen spon­ta­neously. It’s at this level that I talk to peo­ple and find that they rou­tinely have in­ter­est­ing thoughts I haven’t heard be­fore.

Hedge-fund peo­ple sparkle with ex­tra life force. At least the ones I’ve talked to. Large amounts of money seem to at­tract smart peo­ple. No, re­ally.

If you’re won­der­ing how it could be pos­si­ble that the up­per ech­e­lons of the world could be gen­uinely in­tel­li­gent, and yet the world is so screwed up...

Well, part of that may be due to my bi­ased sam­ple.

Also, I’ve met a few Con­gressper­sons and they struck me as be­ing at around the non-big-name ven­ture cap­i­tal level, not the hedge fund level or the Jurvet­son level. (Still, note that e.g. Ge­orge W. Bush used to sound a lot smarter than he does now.)

But mainly: It takes an as­tro­nom­i­cally high thresh­old of in­tel­li­gence + ex­pe­rience + ra­tio­nal­ity be­fore a screwup be­comes sur­pris­ing. There’s “smart” and then there’s “smart enough for your cog­ni­tive mechanisms to re­li­ably de­cide to sign up for cry­on­ics”. Ein­stein was a deist, etc. See also Eliezer1996 and the ed­ited vol­ume “How Smart Peo­ple Can Be So Stupid”. I’ve always been skep­ti­cal that Jeff Skil­ling of En­ron was world-class smart, but I can eas­ily vi­su­al­ize him be­ing able to sparkle in con­ver­sa­tion.

Still, so far as I can tell, the world’s up­per ech­e­lons—in those few cases I’ve tested, within that ex­tremely bi­ased sam­ple that I en­counter—re­ally are more in­tel­li­gent.

Not just, “it’s who you know, not what you know”. Not just per­sonal charisma and Machi­avel­lian ma­neu­ver­ing. Not just pro­mo­tion of in­com­pe­tents by other in­com­pe­tents.

I don’t say that this never hap­pens. I’m sure it hap­pens. I’m sure it’s en­demic in all sorts of places.

But there’s a flip side to the story, which doesn’t get talked about so much: you re­ally do find a lot more cream as you move closer to the top.

It’s a stan­dard idea that peo­ple who make it to the elite, tend to stop talk­ing to or­di­nary mor­tals, and only hang out with other peo­ple at their level of the elite.

That’s easy for me to be­lieve. But I sus­pect that the rea­son is more dis­turb­ing than sim­ple snob­bery. A re­porter, writ­ing about that, would pass it off as snob­bery. But it makes en­tire sense in terms of ex­pected util­ity, from their view­point. Even if all they’re do­ing is look­ing for some­one to talk to—just talk to.

Visit­ing that gath­er­ing of the mid-level power elite, it was sud­denly ob­vi­ous why the peo­ple who at­tended that con­fer­ence might want to only hang out with other peo­ple who at­tended that con­fer­ence. So long as they can talk to each other, there’s no point in tak­ing a chance on out­siders who are statis­ti­cally un­likely to sparkle with the same level of life force.

When you make it to the power elite, there are all sorts of peo­ple who want to talk to you. But un­til they make it into the power elite, it’s not in your in­ter­est to take a chance on talk­ing to them. Frus­trat­ing as that seems when you’re on the out­side try­ing to get in! On the in­side, it’s just more ex­pected fun to hang around peo­ple who’ve already proven them­selves com­pe­tent. I think that’s how it must be, for them. (I’m not part of that world, though I can walk through it and be rec­og­nized as some­thing strange but sparkly.)

There’s an­other world out there, richer in more than money. Jour­nal­ists don’t re­port on that part, and in­stead just talk about the big houses and the yachts. Maybe the jour­nal­ists can’t per­ceive it, be­cause you can’t dis­crim­i­nate more than one level above your own. Or maybe it’s such an awful truth that no one wants to hear about it, on ei­ther side of the fence. It’s eas­ier for me to talk about such things, be­cause, rightly or wrongly, I imag­ine that I can imag­ine tech­nolo­gies of an or­der that could bridge even that gap.

I’ve never been to a gath­er­ing of the top-level elite (World Eco­nomic Fo­rum level), so I have no idea if peo­ple are even more al­ive up there, or if the curve turns and starts head­ing down­ward.

And re­ally, I’ve never been to any sort of power-elite gath­er­ing ex­cept those or­ga­nized by the sort of per­son that would in­vite me. Maybe that world I’ve ex­pe­rienced, is only a tiny minor­ity carved out within the power elite. I re­ally don’t know. If for some rea­son it made a differ­ence, I’d try to plan for both pos­si­bil­ities.

But I’m pretty sure that, statis­ti­cally speak­ing, there’s a lot more cream at the top than most peo­ple seem will­ing to ad­mit in writ­ing.

Such is the hideously un­fair world we live in, which I do hope to fix.

Part of the se­quence Yud­kowsky’s Com­ing of Age

Next post: “Above-Aver­age AI Scien­tists

Pre­vi­ous post: “The Level Above Mine