Surprised by Brains

Fol­lowup to: Life’s Story Continues

Imag­ine two agents who’ve never seen an in­tel­li­gence - in­clud­ing, some­how, them­selves—but who’ve seen the rest of the uni­verse up un­til now, ar­gu­ing about what these new­fan­gled “hu­mans” with their “lan­guage” might be able to do...

Believer: Pre­vi­ously, evolu­tion has taken hun­dreds of thou­sands of years to cre­ate new com­plex adap­ta­tions with many work­ing parts. I be­lieve that, thanks to brains and lan­guage, we may see a new era, an era of in­tel­li­gent de­sign. In this era, com­plex causal sys­tems—with many in­ter­de­pen­dent parts that col­lec­tively serve a definite func­tion—will be cre­ated by the cu­mu­la­tive work of many brains build­ing upon each oth­ers’ efforts.

Skep­tic: I see—you think that brains might have some­thing like a 50% speed ad­van­tage over nat­u­ral se­lec­tion? So it might take a while for brains to catch up, but af­ter an­other eight billion years, brains will be in the lead. But this planet’s Sun will swell up by then, so -

Believer: Thirty per­cent? I was think­ing more like three or­ders of mag­ni­tude. With thou­sands of brains work­ing to­gether and build­ing on each oth­ers’ efforts, whole com­plex ma­chines will be de­signed on the timescale of mere mil­len­nia—no, cen­turies!

Skep­tic: What?

Believer: You heard me.

Skep­tic: Oh, come on! There’s ab­solutely no em­piri­cal ev­i­dence for an as­ser­tion like that! An­i­mal brains have been around for hun­dreds of mil­lions of years with­out do­ing any­thing like what you’re say­ing. I see no rea­son to think that life-as-we-know-it will end, just be­cause these ho­minid brains have learned to send low-band­width sig­nals over their vo­cal cords. Noth­ing like what you’re say­ing has hap­pened be­fore in my ex­pe­rience -

Believer: That’s kind of the point, isn’t it? That noth­ing like this has hap­pened be­fore? And be­sides, there is prece­dent for that kind of Black Swan—namely, the first repli­ca­tor.

Skep­tic: Yes, there is prece­dent in the repli­ca­tors. Thanks to our ob­ser­va­tions of evolu­tion, we have ex­ten­sive knowl­edge and many ex­am­ples of how op­ti­miza­tion works. We know, in par­tic­u­lar, that op­ti­miza­tion isn’t easy—it takes mil­lions of years to climb up through the search space. Why should “brains”, even if they op­ti­mize, pro­duce such differ­ent re­sults?

Believer: Well, nat­u­ral se­lec­tion is just the very first op­ti­miza­tion pro­cess that got started ac­ci­den­tally. Th­ese new­fan­gled brains were de­signed by evolu­tion, rather than, like evolu­tion it­self, be­ing a nat­u­ral pro­cess that got started by ac­ci­dent. So “brains” are far more so­phis­ti­cated—why, just look at them. Once they get started on cu­mu­la­tive op­ti­miza­tion—FOOM!

Skep­tic: So far, brains are a lot less im­pres­sive than nat­u­ral se­lec­tion. Th­ese “ho­minids” you’re so in­ter­ested in—can these crea­tures’ han­daxes re­ally be com­pared to the majesty of a di­vid­ing cell?

Believer: That’s be­cause they only just got started on lan­guage and cu­mu­la­tive op­ti­miza­tion.

Skep­tic: Really? Maybe it’s be­cause the prin­ci­ples of nat­u­ral se­lec­tion are sim­ple and el­e­gant for cre­at­ing com­plex de­signs, and all the con­volu­tions of brains are only good for chip­ping han­daxes in a hurry. Maybe brains sim­ply don’t scale to de­tail work. Even if we grant the highly du­bi­ous as­ser­tion that brains are more effi­cient than nat­u­ral se­lec­tion—which you seem to be­lieve on the ba­sis of just look­ing at brains and see­ing the con­voluted folds—well, there still has to be a law of diminish­ing re­turns.

Believer: Then why have brains been get­ting steadily larger over time? That doesn’t look to me like evolu­tion is run­ning into diminish­ing re­turns. If any­thing, the re­cent ex­am­ple of ho­minids sug­gests that once brains get large and com­pli­cated enough, the fit­ness ad­van­tage for fur­ther im­prove­ments is even greater -

Skep­tic: Oh, that’s prob­a­bly just sex­ual se­lec­tion! I mean, if you think that a bunch of brains will pro­duce new com­plex ma­chin­ery in just a hun­dred years, then why not sup­pose that a brain the size of a whole planet could pro­duce a de novo com­plex causal sys­tem with many in­ter­de­pen­dent el­e­ments in a sin­gle day?

Believer: You’re at­tack­ing a straw­man here—I never said any­thing like that.

Skep­tic: Yeah? Let’s hear you as­sign a prob­a­bil­ity that a brain the size of a planet could pro­duce a new com­plex de­sign in a sin­gle day.

Believer: The size of a planet? (Thinks.) Um… ten per­cent.

Skep­tic: (Muffled chok­ing sounds.)

Believer: Look, brains are fast. I can’t rule it out in prin­ci­ple -

Skep­tic: Do you un­der­stand how long a day is? It’s the amount of time for the Earth to spin on its own axis, once. One sun­lit pe­riod, one dark pe­riod. There are 365,242 of them in a sin­gle mil­len­nium.

Believer: Do you un­der­stand how long a sec­ond is? That’s how long it takes a brain to see a fly com­ing in, tar­get it in the air, and eat it. There’s 86,400 of them in a day.

Skep­tic: Pffft, and chem­i­cal in­ter­ac­tions in cells hap­pen in nanosec­onds. Speak­ing of which, how are these brains go­ing to build any sort of com­plex ma­chin­ery with­out ac­cess to ri­bo­somes? They’re just go­ing to run around on the grassy plains in re­ally op­ti­mized pat­terns un­til they get tired and fall over. There’s noth­ing they can use to build pro­teins or even con­trol tis­sue struc­ture.

Believer: Well, life didn’t always have ri­bo­somes, right? The first repli­ca­tor didn’t.

Skep­tic: So brains will evolve their own ri­bo­somes?

Believer: Not nec­es­sar­ily ri­bo­somes. Just some way of mak­ing things.

Skep­tic: Great, so call me in an­other hun­dred mil­lion years when that evolves, and I’ll start wor­ry­ing about brains.

Believer: No, the brains will think of a way to get their own ri­bo­some-analogues.

Skep­tic: No mat­ter what they think, how are they go­ing to make any­thing with­out ri­bo­somes?

Believer: They’ll think of a way.

Skep­tic: Now you’re just treat­ing brains as magic fairy dust.

Believer: The first repli­ca­tor would have been magic fairy dust by com­par­i­son with any­thing that came be­fore it -

Skep­tic: That doesn’t li­cense throw­ing com­mon sense out the win­dow.

Believer: What you call “com­mon sense” is ex­actly what would have caused you to as­sign neg­ligible prob­a­bil­ity to the ac­tual out­come of the first repli­ca­tor. Ergo, not so sen­si­ble as it seems, if you want to get your pre­dic­tions ac­tu­ally right, in­stead of sound­ing rea­son­able.

Skep­tic: And your be­lief that in the Fu­ture it will only take a hun­dred years to op­ti­mize a com­plex causal sys­tem with dozens of in­ter­de­pen­dent parts—you think this is how you get it right?

Believer: Yes! Some­times, in the pur­suit of truth, you have to be coura­geous—to stop wor­ry­ing about how you sound in front of your friends—to think out­side the box—to imag­ine fu­tures fully as ab­surd as the Pre­sent would seem with­out benefit of hind­sight—and even, yes, say things that sound com­pletely ridicu­lous and out­ra­geous by com­par­i­son with the Past. That is why I boldly dare to say—push­ing out my guesses to the limits of where Truth drives me, with­out fear of sound­ing silly—that in the far fu­ture, a billion years from now when brains are more highly evolved, they will find it pos­si­ble to de­sign a com­plete ma­chine with a thou­sand parts in as lit­tle as one decade!

Skep­tic: You’re just dig­ging your­self deeper. I don’t even un­der­stand how brains are sup­posed to op­ti­mize so much faster. To find out the fit­ness of a mu­ta­tion, you’ve got to run mil­lions of real-world tests, right? And even then, an en­vi­ron­men­tal shift can make all your op­ti­miza­tion worse than noth­ing, and there’s no way to pre­dict that no mat­ter how much you test -

Believer: Well, a brain is com­pli­cated, right? I’ve been look­ing at them for a while and even I’m not to­tally sure I un­der­stand what goes on in there.

Skep­tic: Pffft! What a ridicu­lous ex­cuse.

Believer: I’m sorry, but it’s the truth—brains are harder to un­der­stand.

Skep­tic: Oh, and I sup­pose evolu­tion is triv­ial?

Believer: By com­par­i­son… yeah, ac­tu­ally.

Skep­tic: Name me one fac­tor that ex­plains why you think brains will run so fast.

Believer: Ab­strac­tion.

Skep­tic: Eh? Ab­strah-shun?

Believer: It… um… lets you know about parts of the search space you haven’t ac­tu­ally searched yet, so you can… sort of… skip right to where you need to be -

Skep­tic: I see. And does this power work by clair­voy­ance, or by pre­cog­ni­tion? Also, do you get it from a po­tion or an amulet?

Believer: The brain looks at the fit­ness of just a few points in the search space—does some com­pli­cated pro­cess­ing—and voila, it leaps to a much higher point!

Skep­tic: Of course. I knew tele­por­ta­tion had to fit in here some­where.

Believer: See, the fit­ness of one point tells you some­thing about other points -

Skep­tic: Eh? I don’t see how that’s pos­si­ble with­out run­ning an­other mil­lion tests.

Believer: You just look at it, dammit!

Skep­tic: With what kind of sen­sor? It’s a search space, not a bug to eat!

Believer: The search space is com­press­ible -

Skep­tic: Whaa? This is a de­sign space of pos­si­ble genes we’re talk­ing about, not a fold­ing bed -

Believer: Would you stop talk­ing about genes already! Genes are on the way out! The fu­ture be­longs to ideas!

Skep­tic: Give. Me. A. Break.

Believer: Ho­minids alone shall carry the bur­den of des­tiny!

Skep­tic: They’d die off in a week with­out plants to eat. You prob­a­bly don’t know this, be­cause you haven’t stud­ied ecol­ogy, but ecolo­gies are com­pli­cated—no sin­gle species ever “car­ries the bur­den of des­tiny” by it­self. But that’s an­other thing—why are you pos­tu­lat­ing that it’s just the ho­minids who go FOOM? What about the other pri­mates? Th­ese chim­panzees are prac­ti­cally their cous­ins—why wouldn’t they go FOOM too?

Believer: Be­cause it’s all go­ing to shift to the level of ideas, and the ho­minids will build on each other’s ideas with­out the chim­panzees par­ti­ci­pat­ing -

Skep­tic: You’re beg­ging the ques­tion. Why won’t chim­panzees be part of the econ­omy of ideas? Are you fa­mil­iar with Ri­cardo’s Law of Com­par­a­tive Ad­van­tage? Even if chim­panzees are worse at ev­ery­thing than ho­minids, the ho­minids will still trade with them and all the other brainy an­i­mals.

Believer: The cost of ex­plain­ing an idea to a chim­panzee will ex­ceed any benefit the chim­panzee can provide.

Skep­tic: But why should that be true? Chim­panzees only forked off from ho­minids a few mil­lion years ago. They have 95% of their genome in com­mon with the ho­minids. The vast ma­jor­ity of op­ti­miza­tion that went into pro­duc­ing ho­minid brains also went into pro­duc­ing chim­panzee brains. If ho­minids are good at trad­ing ideas, chim­panzees will be 95% as good at trad­ing ideas. Not to men­tion that all of your ideas be­long to the far fu­ture, so that both ho­minids, and chim­panzees, and many other species will have evolved much more com­plex brains be­fore any­one starts build­ing their own cells -

Believer: I think we could see as lit­tle as a mil­lion years pass be­tween when these crea­tures first in­vent a means of stor­ing in­for­ma­tion with per­sis­tent digi­tal ac­cu­racy—their equiv­a­lent of DNA—and when they build ma­chines as com­pli­cated as cells.

Skep­tic: Too many as­sump­tions… I don’t even know where to start… Look, right now brains are nowhere near build­ing cells. It’s go­ing to take a lot more evolu­tion to get to that point, and many other species will be much fur­ther along the way by the time ho­minids get there. Chim­panzees, for ex­am­ple, will have learned to talk -

Believer: It’s the ideas that will ac­cu­mu­late op­ti­miza­tion, not the brains.

Skep­tic: Then I say again that if ho­minids can do it, chim­panzees will do it 95% as well.

Believer: You might get dis­con­tin­u­ous re­turns on brain com­plex­ity. Like… even though the ho­minid lineage split off from chim­panzees very re­cently, and only a few mil­lion years of evolu­tion have oc­curred since then, the chim­panzees won’t be able to keep up.

Skep­tic: Why?

Believer: Good ques­tion.

Skep­tic: Does it have a good an­swer?

Believer: Well, there might be com­pound in­ter­est on learn­ing dur­ing the mat­u­ra­tional pe­riod… or some­thing about the way a mind flies through the search space, so that slightly more pow­er­ful ab­stract­ing-ma­chin­ery can cre­ate ab­strac­tions that cor­re­spond to much faster travel… or some kind of feed­back loop in­volv­ing a brain pow­er­ful enough to con­trol it­self… or some kind of crit­i­cal thresh­old built into the na­ture of cog­ni­tion as a prob­lem, so that a sin­gle miss­ing gear spells the differ­ence be­tween walk­ing and fly­ing… or the ho­minids get started down some kind of sharp slope in the ge­netic fit­ness land­scape, in­volv­ing many changes in se­quence, and the chim­panzees haven’t got­ten started down it yet… or all these state­ments are true and in­ter­act mul­ti­plica­tively… I know that a few mil­lion years doesn’t seem like much time, but re­ally, quite a lot can hap­pen. It’s hard to un­tan­gle.

Skep­tic: I’d say it’s hard to be­lieve.

Believer: Some­times it seems that way to me too! But I think that in a mere ten or twenty mil­lion years, we won’t have a choice.