Jay Cornell (jay-cornell) Tue 17 Feb 15 22:53
AI in "narrow" forms has made huge strides. AGI (artificial general intelligence) has proven much more difficult. Self-driving cars, better versions of Siri, walking and flying robots: I'm pretty sure those aren't far away, but human intelligence/consciousness is a tougher challenge. I don't think AGI will be very much like us without great effort, and possibly not even then. The skeptic in me says that we don't fully understand human intelligence and consciousness, so how can we accurately replicate it? Yes, computer power is advancing rapidly, but our brains are more than collections of logic gates. Using silicon and software to replicate the behavior of a complex mass of tissue which is influenced by genes, the environment, social interaction, hormones, brain chemistry, diet... that's a tough job. Think of how shifts in brain inputs (light, sound, smells) can change emotions and behaviors: how do you handle that? Should the software emulate all that, or ignore it, or what? Or maybe we don't need to replicate all that. For many uses, simplified AGIs might be preferable. It seems to me that any AGI humans create will in some ways be different from us. Any way in which it is superior to us sets it apart, as does any human ability it lacks. In essential ways it will be alien, which is part of what makes the other side of the Singularity unknowable.
R.U. Sirius (rusirius) Tue 17 Feb 15 23:40
Well there ARE a number of brain emulation projects and biology presumably didn't have to understand what consciousness is to evolve the mechanism that experiences it. But I agree with Jay and Jon in the sense that what will probably emerge from advance AI projects will not be so much like humans but smarter, but different than humans and more useful, but less capable (well, incapable) of meaning and fun and valuing actual experience ad infinitum. So I fundamentally think it will wind up being a tool rather than our Mind Children exceeding us in the evolutionary process. But I could be wrong. One thing that comes up in the book is that the people who actually work on Artificial General Intelligence (as opposed to narrow AI) virtually all believe there will be greater than human intelligences. The ranges of when it will happen go from 10 years to hundreds of years, but they think it will happen.
Jon Lebkowsky (jonl) Wed 18 Feb 15 04:19
The way I see the AI problem is not that machines will evolve a consciousness and a will, and decide that humans are irrelevant or worse - that's the common sci-fi theme, but to me it's not much beyond the golem concept of Jewish folklore, an animated anthropomorphic entity created from clay and animated by magic. The more realistic problem, I think, is that we build a world that can only be managed by artificial or simulated intelligences, and evolve it beyond our ability to manage or pull the plug. What are the dangers in that world, which I think is the more practical vision of singularity? What does it mean to create articial technologies that evolve beyond our understanding? That perhaps think, but not like humans think?
Jeff Kramer (jeffk) Wed 18 Feb 15 07:44
As you guys have been watching this scene for a long while, are there certain technologies that have become commonplace in the last 20 years that have pushed us the furthest towards being different/better than what we were before? One might think of the cell phone and the ubitquity of the internet, but are there others that aren't as obvious that you think are precient of the kind of changes that may come in the next decades? The kinds of things we might have been dreaming about in the 80s?
Peter Meuleners (pjm) Wed 18 Feb 15 10:38
I will rephrase my "tool other than tech" comment. I think that learning how to control our surroundings at the quantum level is pretty much the only way we escape mortality. It will surely take the tools of tech to get very far in that process but eventually the tools must be left behind as conciousness comes to reside in the quantum potential energy flucuations that are always present, no matter the state of entropy. Of course, that could also be called a tool. But this may be utterly impossible. Introducing any kind of order to random fluctuations may be a self-canceling idea and entropy will always win. More conventional tech could lead to quantum tunneling from a dying Universe to a young Universe. That process could be repeated ad infinutum. (I know that these ideas are way out on the edge (or have fallen off of it) and that much of this discussion will be about what we actually see happening now and in ensuing decades, but these end-game sort of ideas fascinate me.)
Jay Cornell (jay-cornell) Wed 18 Feb 15 13:00
True, "biology presumably didn't have to understand what consciousness is to evolve the mechanism that experiences it." There's a chance that machine consciousness might arise in a way not fully "designed." But I suspect it's trickier than many believe. I am not too worried about the Skynet scenario, of intelligent machines deciding to exterminate humanity. Machines will need humans for maintenance and support for the foreseeable future, and I have a hard time imagining a motive for such slaughter. But as I said, their minds will not be fully human, so we have to take care with the control we give them.
Jay Cornell (jay-cornell) Wed 18 Feb 15 13:14
Jeff Kramer: We're both very interested in the rapid advances of 3D printing, which has the potential to be quite disruptive. I'm not sure everyone will have one in their house, but it will spur decentralized and personalized manufacturing. I think we'll see places like FedEx Office (formerly Kinko's) buying 3D printers, to rent like copiers. Medical advances are also rapid. This year there will be a trial of anti-cancer DNA nanobots, basically an anti-cancer drug inside a DNA capsule designed to open when it detects cancer cells. This should allow precise delivery of the drug. If that works, and they can develop versions that target heart disease and dementia and other hard-to-treat conditions, it will be revolutionary.
R.U. Sirius (rusirius) Wed 18 Feb 15 15:03
There's the stuff that is already available, which in terms of enhancement, isn't much by 21st C. standards. Obviously, the info tech that's already been mentioned. There's the raw beginnings of replacement parts becoming better than the original. They're not really, but you have the runner Oscar Pistorius (also handy with a gun) and his artifical legs raising questions about whether it would be unfair to allow him to compete in an olympics competition (when he did compete, he never bested second place). There's a story in the book about a man with a mind-controlled bionic leg who climbed the 103 story Willis Tower in Chicago. There's a renewed enthusiasm around nootropics i.e. smart drugs much ballyhooed and sneered at during my days on The Well in the early '90s. I don't think there's been a test that shows permanent intelligent increase although the term mental enhancement has been used around some Alzheimer drugs. My sense is that progress is being made in this area but we can't be sure. There's a telescopic contact lens. Lots of that sort of thing. External enhancers that are slow to get to market and get distributed.
R.U. Sirius (rusirius) Wed 18 Feb 15 15:11
The most exciting stuff is, as usual, always tantalizingly in the experimental stage. Jay mentioned the nanobots. There's been success manipulating brain activity with optogenetics (pulses of light) There's been the actual reversal of cellular aging in mice. It's always the damn mice. Still, being able to do that in a mammal should be seen as proof of concept, even if we never figure out how to do it in the much more complex human. Jay mentioned 3d printing. There's the company Organovo that has 3d printed liver tissue. They're already talking about full organs (and about a way of introducing the tissue into the body so that it functions, step by step, as a full organ replacement. This is an enhancement in the sense of health extension and possibly life extension. That's kind of off the top but there's so much more. It may all sound abstract, but there's a lot of stuff that's a lot closer to actual usage than a few years ago, never mind back when we were dreaming quasi-utopian dreams 20 years ago.
R.U. Sirius (rusirius) Wed 18 Feb 15 19:01
Jon commented >> The more realistic problem, I think, is that we build a world that can only be managed by artificial or simulated intelligences, and evolve it beyond our ability to manage or pull the plug. What are the dangers in that world, which I think is the more practical vision of singularity? >> While we're not utterly pwned yet to artificial or simulated intelligences, we're already in a sort of social/financial singularity in which all our data is available and has probably been pwned by various groups of hackers working for state entities or for themselves or by big finance or some combination thereof. And of course we have the periodic panics about the "power grid" and so forth. We're really on a razor's edge where that balance between order and chaos feels like it's going to tilt over into utter chaos. Which, in fact, may be a good reason to get very powerful AIs on the case, adding resilience to systems that are brittle and thin. I mean, we should all have legitimate encrypted backup identities, each one made up of a distinct code... at least one.
Jon Lebkowsky (jonl) Wed 18 Feb 15 20:07
"Which, in fact, may be a good reason to get very powerful AIs on the case, adding resilience to systems that are brittle and thin." What would that look like? How broadly defined would these AI agents be?
R.U. Sirius (rusirius) Wed 18 Feb 15 23:37
hmmm. i'm just thinking that a complex enough AI... not necessarily even AGI but just unthinkable data churn... could probably churn out nearly infinite variations of encrypted code that shields each persons personal data. How the end user could actually use this to supplant the good old easily cracked social security numbers and like that... it's beyond my pay grade. Bring in the cypherpunks. Now... about this book...
Jon Lebkowsky (jonl) Thu 19 Feb 15 05:34
The book is an overview and guide to the impact of science fictional concepts leaking into the real world. The fictional subgenre of what-if scenarios always depended on an assumption that ideas about the future are potentially predictive, that science fiction worked in a field of possibility. Somewhere in the coming together of the cyberpunk near-future narratives and the evolution of the Internet, the sense of possibility was hyped into overdrive. How did you build this guide to all the stuff that fell out of that collision of imagination with real-world invention? How did you decide what to include, what to exclude? And how did the two of you juggle voice and authorial responsibility in putting the book together?
R.U. Sirius (rusirius) Thu 19 Feb 15 10:25
The selections were largely based on a familiarity with the obsessions of people in the transhumanist culture. Leaving aside the personalities and groups, most of the science and technology topics could be linked to ideas of human enhancement and many of them can tie into NBIC -- Nano Bio Info Cogno -- one of the notions of convering technologies that might radically alter our collective and individual futures. Gotta run. More later...
Jay Cornell (jay-cornell) Thu 19 Feb 15 12:23
R.U. came up with the original table of contents, which we largely stuck to. As the book progressed we made some additions and combined a few topics when that seemed warranted. Regarding voice, that just seemed to happen pretty smoothly. Our styles aren't too dissimilar, at least when writing about these topics. We also edited each other, which probably helped, but we really didn't change a lot that way. I suppose a careful reader might be able to guess who wrote what, but I think it all came out without any jarring shifts.
R.U. Sirius (rusirius) Thu 19 Feb 15 17:32
There are the sort of really broad topics like Consciousness, Artificial Intelligence, Evolutionary Psychology, Genomics, Moore's Law, The Singularity, Cognitive Science, Open Source and so on... And there are the things that are extant or are in process of occurring like Augmented Reality, Brain-Building Projects, Graphene, Implants, Neurobotics, Optogenetics, good ol' VR... Then you've got the really extreme out there stuff that actually draws the most attention and feeds SF writers (at the same time as they often sneer at it) like Abolitionism ... the idea of ending suffering of all sentient beings... or Mind Uploading, Simulation Theory, Transbemanism, and, of course, Psychedelic Transhumanism. And there are the quirky personalities and groups like the Mormon Transhumanist Association...
Gail Williams (gail) Thu 19 Feb 15 18:36
R.U. Sirius (rusirius) Thu 19 Feb 15 19:39
Jon Lebkowsky (jonl) Thu 19 Feb 15 20:35
In the book, you note that "living beyond the perceived limits of an individual human life seems to be the central obsession of transhumanist culture." Transhumanists believe that we can live longer with a higher quality of life and some believe we can ultimately defeat death, that there are people living today who will live forever. Where do you fall on the spectrum of belief between longevity and immortality? If and when you die, will you go the Alcor route? (For the benefit of those unfamiliar with Alcor, it's a cryonics company that freezes human bodies just after death, hoping to be able to resurrect them when future tech/medicine supports that sort of thing.)
Jay Cornell (jay-cornell) Thu 19 Feb 15 22:58
Certainly, we have been living longer and with a higher quality of life than ever, due to improved hygiene and nutrition and health care and generally more health-conscious attitudes. It's fascinating to me to see how age is portrayed in 1930s films, compared to now. Back then, people in their 40s or 50s were usually portrayed as elderly, and anyone in their 60s was considered pretty ancient. Nowadays, we have actresses who are sex symbols at ages that would have shocked people back then. (OK, maybe some of that is due to plastic surgery, but not all of it.) The Alcor thing is interesting and I wish them luck, but it never appealed to me, personally. Lots of transhumanists are interested in mind uploading, which is something I'm rather skeptical about. Like AGI, I think it'll be far harder than many believe. Here's one reason: OK, your mind has been uploaded to a computer and you can live forever (maybe). You can now think 10 million faster than before! That's great, except doesn't that mean your subjective sense of time is also 10 million times faster? Does every hour feel like 10 million hours? When an old-fashioned meat human speaks to you, won't it seem like it takes them a year to speak 3-4 words? Here's another issue: how do you handle pension schemes like Social Security if people are living for hundreds of years? Many pension systems are already going broke. It's a big social and financial problem that any longevity breakthrough will vastly complicate.
R.U. Sirius (rusirius) Thu 19 Feb 15 23:30
Speaking for myself, you'll notice in the book that transhumanists are "they" rather than I, and my general lack of personal enthusiasm for living hundreds or thousands or nearly infinite numbers of years may be telling. I would like to live a very long time in good health in a human civilization that doesn't disgust or exhaust me, and to be honest, I don't think we're there and I'm a bit skeptical about seeing it happen. On the other hand, if treatments become available that can effectively alleviate some of the problems that come with aging, of course I will take the treatments, just as I take my medicine now when it comes to the choice of feeling sick or dying or feeling a bit better and not dying. And in so doing, my life may be extended beyond the generally accepted 120-some year limit thus far. Indeed, this is not really the topic but... most of the people who are sort of blowhards about hating tech will usually wind up using the tech when they need it. Technically speaking, I'm very skeptical about immortality. The explanations that you find for the idea that if your body doesn't die of natural causes you can find a way to go on until the heat death of the universe seem like fantasies to me... although we have an excellent one in the book. (I like fantasies)
R.U. Sirius (rusirius) Thu 19 Feb 15 23:33
And oh yeah, the cryonics thing is not for me for much the same reason that I have qualms about hyperlongevity... this out-of-death experience I'm having is ok, but I'm not currently craving a return engagement.
Jon Lebkowsky (jonl) Fri 20 Feb 15 04:43
Obviously extended quality of life is a thing, but I spend some time hanging out in medical circles, and I've learned that medical and surgical interventions that are more aggressive tend to have rather pronounced down sides - side effects, potential for infection, etc. So, for instance, if more people choose elective surgery to place implants, to me that's a risk indicator, and not a positive move toward immortality. So I'm skeptical, too, but when I've had conversations with somone like (Extropian cofounder and Alcor CEO) Max More, he seems completely level-headed and intelligent. What are your thoughts about Max's work, and the concept of extropy? Immortality and life extension aside, is his thinking practical philosophy or a utopian dream?
Jeff Kramer (jeffk) Fri 20 Feb 15 07:48
Getting back to the pensions bit, it seems like if we can extend life indefinitely, it's still going to take upkeep dollars to live, and at some point you have a giant population of people who've lived for a super-long time and are super-connected but possibly not super-motivated (they've put in their time, have their money in the bank). If a giant bubble comes along (let's call it Neptunian Methane Futures, because that sounds cool) and wipes out everybodys checking account and they can no longer afford to keep living, that's an interesting problem. Is the assumption that the market makes stay-alive-pills super-cheap because everybody wants them? Or that by then we've transcended meatspace, so we all just fit inside diamond computers orbiting the sun?
R.U. Sirius (rusirius) Fri 20 Feb 15 10:26
Jonl Medical technology isn't pretty. Side effects, unintended consequences and the expense can be prohibitive... If we're talking surgical organ replacement, then the access is probably limited to the wealthy. Hopefully, targeted medicines and cellular repair and replacement and a whole host of other effective treatments will improve things in these areas. There are converging sciences and technologies ranging from nanotechnology and genomics to Optogenetics to forms of self monitoring that could get so sophisticated that you're all over a problem as soon as it develops. Now if they can get vegetables to taste like Buffalo wings, my personal health will be guaranteed for decades to come. Max More is a pretty well balanced and thoughtful guy to pay attention to. Regarding ALCOR, I was a bit cynical about cryonics in the body of the book, but a few months later when we were able to add an addendum, this had come up... " Doctors at UPMC Presbyterian Hospital in Pittsburg will place ten patients with life threatening gunshot or knife wounds in suspended animation, theoretically allowing them more time to fix the injuries.
Members: Enter the conference to participate