The Singularity in the Zeitgeist

Too many SMBC comics to get all of them in one post, but here are four recent ones:

#1 The Fermi Paradox is resolved with reference to wireheading.

#2 About mind uploading.

#3 A different kind of singularity, and (naive) Fun Theory.

#4 Making fun of aging singularitarians.

I recently started a theoretically humorous webcomic about the Singularity entitled Singuhilarity. It's poorly drawn and a little rough in places, but it touches on a number of lesswrong-type subjects as well as some pretty standard science fiction tropes.

Here's the first comic and here's the latest.

The Big Bang Theory, Episode 402.

Sheldon (the most socially atypical character on a show full of them) plans a program of life-extension so that he will last until the Singularity, which he projects to occur around 2060 and chiefly involve the uploading of human consciousness into machines (his roommate Leonard describes the latter as become a "freakish self-aware robot". By the end of the episode Sheldon seems to have given up on the plan as too inconvenient/inadvertently dangerous.

T-Rex's birthday is tomorrow which happens to be MY birthday as well! Unlike T-Rex, however, I am not all emo about aging AND also unlike T-Rex, I have discovered a way to live forever: I will give you a hint, it involves liquid nitrogen and the boundless expanse of interstellar space and also entropy reversing somehow

From the Dinosaur Comics news post for 2010 October 19.

Today's xkcd: Future Timeline.

Edits:

• #888: uncritically-presented bad Fun Theory.
• #893: uncritically-presented straw-Vulcan argument for space travel.
• #894 references the increasingly visible progress being made in narrow AI.

Questionable Content mentioned the singularity in passing circa last night. The phrase "according to the Internet" made me think that there was some particular exaggerated article or press release making the rounds that it was referring to, but I couldn't find it and I was encouraged to refrain from asking directly.

Cracked.com's Jason Iannone, on the game-changing nature of AGI versus its portrayal in fiction:

Cortana, the Chief's AI sidekick [in Halo] [...] can control entire cities by remotely breaking into their battle nets and taking over their weaponry, without anyone ever setting foot in the area. Which is another way of saying that she really doesn't need the Chief at all.

[...] Master Chief, for all his badassitude, is really just a grunt whose life is being put unnecessarily at risk, although a video game featuring him doing nothing but sitting around eating tacos in the cafeteria while Cortana does all the work would probably not have sold as well.

From 6 Video Game Heroes Made Useless By Supporting Characters. High Challenge is relevant to the closing point.

At another thread: Greg Egan disses stand-ins for Overcoming Bias, SIAI in new book (specifically Zendegi).

Why the Singularity isn't going to happen on io9.com. Does that count as "in the culture at large"? Anyway, it's a really really weak article (apparently typical of io9) based on the idea that "singularity-level" technologies in the past (a misunderstanding in itself), such as the industrial revolution, didn't lead to the paradise on earth that some people said they would. The link between "it won't be that great" and "it won't even happen" is never bridged. Could be summed up as saying "it's geek religion, man".

Found via Reddit: On eutopia.

PvP 2009/03/24.

From last August: Scenes from a Multiverse takes on dualist objections to mind uploading.

Reddit comment thread in progress:

depending on when exactly we achieve this, this could be the best time to be born ever, because it will be the absolute earliest anybody will have achieved immortality. Someone born within 20 years of this moment could one day be the oldest human, sentient, or even living being in the Universe.

The comments are currently split between arguing and agreeing with this. So far, no mention of cryonics. One post presents a technical argument that "our current knowledge/technology is centuries away" from mind uploading/whole-brain emulation.

Further links, here and elsewhere:

• Cryoburn by Bujold, via NancyLebovitz
• CultureLab: Iain M. Banks: Upload for everlasting life, via Reddit
• Catastrophic risks from artificial intelligence (Onion Videos), via XiXiDu
• Bill Gates: "There are other potential problems in the future... Some would put super-intelligent computers on that list", via FkaRoko

At another thread: Stanford historian on the singularity.

Machine of Death, a recently released short-fiction anthology that hit #1 on Amazon, has a cover quote from Cory Doctorow containing the sentiment "Makes me wish I could die, too!".

In one of the stories, "Flaming Marshmallow", "zvyyraavhz fcnpr ragebcl" vf vagrecergrq nf zrnavat gur crefba jba'g qvr gvy gur arkg zvyyraavhz, juvpu vf frra nf tbbq arjf engure guna gur greevoyr arjf vg zvtug or sbe fbzrbar rkcrpgvat cebcre vzzbegnyvgl gb unccra ol gura.

So far I haven't seen any of the stories deal with how the machine affects the cryonics movement (start testing rats and then putting them in suspension, and try to either fool the machine or get a prediction that doesn't match the cause of deanimation?) or physics (strong evidence against many-worlds?), or how it deals with mind copying (does it arbitrarily distinguish between the "original" and copies, thereby giving people a loophole to escape their deaths, or do all the copies end up dying in similar ways?), merging or reconstruction, or whether there's a large-scale study of the effects of praying and making sacrifices to the obviously intelligent Predictor. On the other hand I've only just started the book, and there's already been talk of another volume, so I hold out some hope.

The anthology's concept seems to rule out an imminent positive singularity, since you'd expect to have over a century's advance knowledge when predictions started reading "end of the universe"; but I wonder if an FAI could make a deal with the Predictor by precommitting to kill everyone in their appointed ways just before the universe ended. If the prediction required a person to be dead by a particular date, the AI would want to figure out how much of a person could be preserved while still having the Predictor consider them dead. Or maybe the Prediction Enforcer kills them in their appointed way heedless of the FAI, and the only difference the AI makes is in making it impossible for the Enforcer to disguise its actions as the arbitrary happenstance of life. (Back in 2006 I figured that was the only real possibility, since information going backward in time allows paradoxes; but now I see both as plausible.)

The AI-related sci-fi series Caprica was recently cancelled, following in the footsteps of The Sarah Connor Chronicles and Odyssey 5. The last five episodes are planned to air sometime in 2011.

I haven't watched the show, but comments I've seen on it include:

Genii Lodus, Stardestroyer.net, after a fairly harsh recap of the pilot:

The only interesting story this series might have had to tell - how Cylons come about has pretty much been done in the pilot- a 16 year old girl developed AI in her spare time outside of school and she was put into a magical plot device processor and plugged into a robot. (snip) I suppose if they'd had an actual progression towards developing true AI then it might well have looked too much like the Sarah Connor Chronicles but given both settings feature the significant plot of man builds AI which goes bad and then nukes man to fuck there would always be comparisons.

JoCoLa, Reddit:

it's taking its time with pretty complex subject matter- every movie or show about A.I. takes place after it was created, this is the only show I know of that addresses how it comes into being.

Earlier this year, the Reddit thread "Hey Reddit, how do you think the human race will come to an end?" received a fairly elaborate Singularitarian reply. Although it might be unfair to generalize given the context, he seems to be taking a basically fatalist stance, taking it as given that whoever creates AIs will program them to share our values.

The post currently has about 1165 karma (not necessarily representing votes of agreement), plus at least a few hundred replies taking probably every major position on the subject. This reply links to Robin Hanson's article If Uploads Come First.

I was disappointed that this subthread ended with the rhetorical question "What current AI projects aren't working with brains?" unanswered.

(Edit 11/25: corrected the assumption that the number of upvotes displayed by the script I had installed was accurate.)

The half-hour TV series "Sci Fi Science: Physics of the Impossible" recently aired an episode on AI risk, "A.I. Uprising". This is the airing schedule. It doesn't seem to be available online yet.

Via Cyan: John Scalzi blogged this semi-humorous story of a dubiously-friendly intelligence explosion, going into more detail than most about the path from intelligence to real-world power.

A 2006 Overcompensating post talks about the absurdity of how widely ignored existential risks are.

There's an unpleasant comedy sketch show take on cryogenic preservation here: http://www.youtube.com/watch?v=g7Lzr3cwaPs

Plot: a cryogenics company goes bust, the bodies they store are bought cheaply and revived to make low budget TV with no concerns about treating the revived people badly.

Relevance: What if cryogenic preservation works but the plan of waking up in a better future, doesn't.

The Post-Singularity Future Of Astronomy.

Yesterday's SMBC presents SIAI-style uFAI fears as essentially a self-fulfilling prophecy.