June 2011


Charlie Stross, author of a great book on the Singularity, has a post explaining why he doesn’t in fact believe in it.

The crux of his reasoning seems to be this passage, where he suggests that we don’t want ‘real’ AIs:

We clearly want machines that perform human-like tasks. We want computers that recognize our language and motivations and can take hints, rather than requiring instructions enumerated in mind-numbingly tedious detail. But whether we want them to be conscious and volitional is another question entirely. I don’t want my self-driving car to argue with me about where we want to go today. I don’t want my robot housekeeper to spend all its time in front of the TV watching contact sports or music videos. And I certainly don’t want to be sued for maintenance by an abandoned software development project.

As I argued in the PCK, human-interchangeable robots are in general a fairly dumb idea.    We don’t need a human-level AI because we have humans.  We could use various subsmart appliances; we can and do use specialized robots; we could likely use supersmart apps that help run countries and corporations.  About the only real use I can think of is as generalized workers in niche environments we can’t survive in, like deep space.

(There could be a market for tutors, butlers, and sexbots.  But this runs into the paradox of automation: replace enough jobs and humans become extremely cheap.  Using robots for these tasks would be another niche, for people too fastidious or paranoid to hire humans.  And Stross’s argument applies: you don’t want your children’s tutor, or your sexbot, to be scheming against you.  You want situational cleverness, not sapience.)

Plus, all the great features of AIs… why not just add them to humans?  That’s where I think we’ll end up going, and it’s what I put in the Incatena. 

(Stross also talks about the ethics of developing AIs… before you give voting rights to AIs, think a bit about when in the development process sentience occurs.  Long before product ship, for sure– but in continuing development, are you killing a sentient being?  See also my story on this.)

Stross quotes Hans Moravec as thinking that humans just won’t be able to compete with the nimble AIs:

A human would likely fare poorly in such a cyberspace. Unlike the streamlined artificial intelligences that zip about, making discoveries and deals, reconfiguring themselves to efficiently handle the data that constitutes their interactions, a human mind would lumber about in a massively inappropriate body simulation, analogous to someone in a deep diving suit plodding along among a troupe of acrobatic dolphins. Every interaction with the data world would first have to be analogized as some recognizable quasi-physical entity … Maintaining such fictions increases the cost of doing business, as does operating the mind machinery that reduces the physical simulations into mental abstractions in the downloaded human mind.

There seem to be two main elements to Moravec’s suggestion.  One: primate brains are slow.  Now in part I think this is an illusion caused by dwelling in the land of theory: electronics are sure faster than neurons, so electronic brains must be better!  Only the comparison is quite unfair: computers are fast because their basic operations are trivial.   Your brain can still do things in an instant that megacomputers still can’t, such as dealing with language and easily processing visual data.  (Not that I think AI is impossible.  I think it’s harder than is often assumed, but I’d be really surprised if we didn’t have it in a couple of centuries.)

Computers are better at, well, computerlike tasks.  Again, the obvious step here is not to hand over the keys to civilization, but to incorporate those advantages into our brains.  In AD 4901 you’ll be able to contemplate millions of chess moves, grep megabytes of data, and do vector math as fast as a computer too, since you’ll have one in your skull.

The other idea is that AIs will somehow not need the physical metaphors that we allegedly need.  Seriously?   Does Moravec really think that, say, computer programming is based on primate metaphors?  BASIC maybe, but surely not C++.

I think Moravec’s argument here is actually backwards.  He seems to have a priggish distaste for the biological… a long trope in science fiction, one that C.S. Lewis had a great time parodying back in That Hideous Strength, where his villains had a cringing disgust for the messiness, the fluidness, the grossness of organisms.  But that taste for dead, totalitarian order is way past its sell-by date.

To put it bluntly, I think it’s barmy to give up sex, sports, gardening, and eating, to say nothing of the aesthetics of music, dance, or the visual arts.  All that in order to do what?  Play a really good game of chess?  One you can play anyway once you have that calculation neurimplant?

Also see this post reflecting on another excellent point of Stross’s: there’s a limit to how far you can get anyway with pure thinking.  After enough of that you have to go back to the lab and check it out anyway. 

If you want to take a minor planet, perhaps Vesta, and turn it into computronium and think deep thoughts with it, fine.  But I just don’t see any convincing reason that we need more deep thinking than that.   (None of the singularity advocates seem able to explain what superhuman AIs will do that requires ever-increasing computational power.)

Advertisements

Venkatesh Rao has a fascinating, mind-blowing, and somewhat half-baked post on the history of the corporation from 1600 to 2100.  If you like Big Exciting Ideas you’ll like it.

I like his little diagram of, well, Things That Influence History:

Venk diagram

 He explains:

On the scale of days or weeks, culture, politics and war matter a lot more in shaping our daily lives. But those forces fundamentally cancel out over longer periods.  They are mostly noise, historically speaking. They don’t cause creative-destructive, unidirectional change (whether or not you think of that change as “progress” is a different matter).

Business though, as an expression of the force of unidirectional technological evolution, has a destabilizing unidirectional effect. It is technology, acting through business and Schumpeterian creative-destruction, that drives monotonic, historicist change, for good or bad. Business is the locus where the non-human force of technological change sneaks into the human sphere.

Which however suggests that there should really be an even smaller, harder circle called “Technology”.  Science and technology don’t always have to go through business to get things done.

He divides the history of the corportaion neatly into two periods: one in which corporations largely invested in the control of space, which largely meant getting direct and exclusive access to physical resources in a zero-sum mercantilist world; and one in which they sought to control time– i.e. what we usually call productivity, which involves that deep embrace of technology. 

I have some quibbles over some of his details.  I think he focuses on the East India Company too much, because he’s part of the Anglosphere.  He mentions the Dutch VOC but only barely; but the story he tells of a corporation taking over a key supply region (the EIC, Bengal, 1757) was foreshadowed a hundred years earlier by the VOC in the Spice Islands.  (Indeed, just as the EIC is largely why there was a British India, the VOC is largely why there is an Indonesia.)

He has an interesting idea, Peak Attention.  Time-mining, he says, has its limits.  The premodern lifestyle had huge sinks of wasted time which technology could save: think of all the time it took to chop down trees, wash clothes, cook meals.  This fits in with Yglesias’s point about surprising unproductivity in mid-20C America: insurance used to be done by door-to-door sales, backed up by hordes of pencil pushers and people who looked things up in printed tables.

But, says Rao, there are limits to how much more attention can be mined.  Corporations today are reduced to remining time already saved: e.g. convincing people to watch your new TV show on your computer instead of going to a movie theater.

I’m not so sure about this.  For one thing, it seems to be confusing productivity with leisure time.  So far as I can see, productivity in work can keep increasing for a very long time.  Whatever your job, there’s going to be something that can be done to speed it up; and if this process is repeated enough, entirely new job categories will be created. 

Productivity can increase leisure time, but only if that’s a social value… in the US, so far as I can see, it isn’t; if anything companies are demanding more from their workers.  

The article kind of falls apart when it comes to explaining what’s next.  Rao thinks the heyday of the corporation is over, but doesn’t really explain why, besides referencing Ronald Coase.  He starts to talk like this:

How do we measure Coasean growth? I have no idea. I am open to suggestions. All I know is that the metric will need to be hyper-personalized and relative to individuals rather than countries, corporations or the global economy. There will be a meaningful notion of Venkat’s rate of Coasean growth, but no equivalent for larger entities.

The fundamental scarce resource that Coasean growth discovers and colonizes is neither space, nor time. It is perspective.

Digging a bit deeper, this seems to refer to “cloudworking” and other blue-sky notions that the mid-20C corporate workforce is over.  I haven’t read enough to be sure what he’s talking about, but my initial reaction is that he’s mistaken an increase in corporate power for its opposite.  The cradle-to-grave job such as my father had is gone not because workers wanted to be free, but because labor is nearly powerless, and because job requirements are far less static.

Yeesh.  I decided to try the ending of Dragon Age Origins again, this time with infantry.  And this happened:

Giving Mr. Dragon a headache for all the ones he caused me

I.e., victory.  But good lord was it a squeaker.  The battle is in three parts; in the first part you just fight the dragon, which isn’t too horrible so long as you make everyone attack at range.  There’s also some ballistae to use.  Then the dragon retreats out of melee range and sends in a load of darkspawn.  I held off on sending in the dwarves till then; I figured they’d do better than the elves had.  And they did, though I still got swarmed a bit.

I thought it was another failure when my PC died… especially as she had the best area of effect spells, Inferno and Fireball.  Still, everyone was in good health and I continued.  I was cursing a lot because the characters didn’t always seem to respond or do damage… due to lag, I think.

I made it to stage three, when the dragon moves back into range; I’d never got this far in previous tries.  But then I lost Leliana, which meant that I couldn’t use the ballistae much.  (Only a rogue can repair them.)  Now I was down to Alistair and Morrigan… and I think I had one dwarf left.  And it looked like there were plenty of darkspawn.

And then Morrigan was gone.  No magic at all now, and I’m not even good with warriors.  But hell, Alistair was still alive, and I had plenty of health potions.  I dashed to a ballista and fired it at Mr. Dragon… also knocking off a few darkspawn that chased after me.  Fortunately they were just grunts and went down fast.    

Finally the ballista jammed.  But the dragon was down to a sliver of health.  I switched to a crossbow– not Alistair’s forte, but I didn’t want him getting near the thing.  I whittled down that sliver one arrow at a time, restoring Alistair’s health as needed.  And finally the damned thing crumpled.  I had won.

A cutscene played… my PC had somehow revived, and dashed over, grabbing a mondo sword and plunging it into the dragon’s brain.  Pretty cheeky, considering she’d been out cold for several minutes and also, being a mage, didn’t know a thing about swords.

The game offers you a few final choices in a post-coronation scene.  The only really interesting bit would be what would happend with Leliana.   She seemed to be playing it cool… the game let me flirt with her, but she had plans to go off on a quest of her own.  Well, at least I could invite myself along.  The little Orlesian tease.  (Checking the wiki, this is because Leliana was in a romance, but wasn’t in love.  I never got the option to do her personal quest.)

Forgot to mention this: there’s a new grammar up, Lé.  It’s isolating, and also fun because it belongs to a female-dominant culture.

Well, that’s about it, again, for Dragon Age Origins.  It’s frustrating; I’ve tried the final boss battle half a dozen times, and can’t get past the huge wave of darkspawn that show up.  And it doesn’t help that the battle is so laggy on my computer that it’s a complete chore to do anything.  Or that you can’t save partway through. 

Edit: but see this.

Damned dirty dragon

I have mixed feelings about the game.  It was compelling enough to put up with the lag and endless sloggy dungeons.   I think the combat system completely sucked for a warrior, but it’s not bad when you have two mages; you just get into a rhythm of pausing every couple of seconds to select spells.  There was even one really difficult battle where I lost everyone but Morrigan, and she managed to wrap up the remaining baddies… very satisfying. 

The voice acting and character interaction are the best part of the game; unfortunately you get less and less of it as the game goes on.  They’ve made a real effort to suggest a complicated world with realistic power struggles, fallible leaders, and rampant prejudice… and then made every problem something to be solved with a brawl.  It seems like a cheat sometimes… you’re presented with some moral choice, or a tasty bit of intrigue (e.g. which dwarven king to support, or how to defeat the evil usurper without alienating his powerful daughter), but whatever you decide, the real result comes from one more battle.  (The dwarven kings don’t matter, and the only real concern with the usurper plot is to not lose one of your companions.)  All this is one reason I appreciate a good stealth game: if well done, it’s an alternative to constant combat.

4. Anyone ever tell you that you look hot covered in blood?

The romance options are interesting, though mine got screwed up… I was going for Leliana, and simply never got any options to do anything interesting, despite having her at 100%.  Probably just as well; I’ve seen the Leliana lovemaking session online, and it’s, well, less than erotic.  The DAO character models are way too clunky; it looks like robots grappling.  Mass Effect 1 did this a lot better.

I appreciate what they were trying to do… this world feels under threat in a way that Oblivion didn’t manage (with its people complaining about mud crabs even as Hell is invading).  But I can’t get past a few basic problems:

  • It’s a very conventional fantasy world… there’s really nothing amazing about the art and architecture, the weapons, the fantasy races, or the gameplay.  It’s a nice touch that the elves are the underclass, but they really have no character besides the pointy ears.  Jade Empire was way better in terms of interesting combat mechanics and a gameworld that seemed novel and lively.
  • The combat system seems to combine the worst aspects of real-time and turn-based systems.   A fully turn-based mechanic as in King’s Bounty actually works better: you fully focus on the different capabilities of each warrior in turn and never lose track of anyone.
  • The quests are just so damn long.  I have zero desire to replay any of them, they were just slogs.  And that in turn means that there are huge chunks of companion interaction I’m missing out on, or alternative character builds that I’m not going to try. 

Another gruntle: the inventory system.  I felt like I never found enough good stuff, and inventory management becomes a chore rather than a challenge.  You had to worry about full inventories in Oblivion too, but you had Feather spells there, and the dungeons weren’t so long.  The crafting system seems half-done… I could never find enough lyrium dust or elfroot to make much.

All this sounds rather negative, which is a little misleading.  There really are a lot of nice little touches.  I like when the companions tease or annoy each other; I like the dynamic where your choices may please one companion and offend another; the war dogs are a great touch; and the really weird bits (mostly the Fade sections) are quite good.  The multiple origins are also quite clever.

When last we visited Against Peace and Freedom, my wife was reading it.  She took her time, but she finished, and had lots of interesting comments, as did my old friend Harry.  I finished another round of revisions, and now I’m getting another printed copy.  I changed the margin size, so I needed to see it in physical form again.

It always amazes me how long this process lasts… this was supposed to be the novel that was pretty near ready to go.  I’m a little alarmed at what I’ll find when I open up Babblers again.

I’m also thinking of making a nice color book of the Historical Atlas.  At least a few people have said they’d get one, so why not?