Accelerating Future

A short article I just wrote, mere seconds ago, for the Q&A site WiseGEEK:

***

A superintelligence is a theoretical smarter-than-human being. Despite the superficial implausibility of the idea, numerous scientists have not ruled out the possibility of a superintelligence being created within the next century or even the next 20 years. Some researchers have even stated it as their career goal. Some of the most interesting speculations regarding superintelligence involve scenarios where the superintelligence continues to further enhance its own intelligence and capability. Such an event has been termed a “Singularity” by theorists.

Recent movies such as The Matrix have popularized the notion of a superintelligence with the public. Artificial intelligence (AI) is one of the possible paths to superintelligence, and perhaps the most frequently discussed. With robotics as its “hands” and computers as its “brains”, a sophisticated artificial intelligence could think more thoughts and achieve more actions than a human would be capable of. This would prove especially true if the AI could redesign its own cognitive architecture, engineer and fabricate new robotics tools, and absorb more computing power. Many philosophers have expressed concern that such an entity could lack moral common sense and be a threat to humanity.

Other possible paths to superintelligence would be through human genetic engineering or Brain-Computer Interfacing. Brain-Computer Interfacing has improved to the point where mouse cursors and prosthetic arms can be controlled by thoughts alone. The military is pouring millions of dollars into developing the technology. Eventually we might create human beings capable of interfacing very closely with the global Internet, to the point where it would be difficult to distinguish between the two. This possibility has been explored in the movie and series Ghost in the Shell.

Human genetic engineering is probably the least likely path to lead to superintelligence, as such research on humans is heavily restricted in most developed countries. This path would take the longest to produce results - while a modification to an AI program or brain-computer interface might take only a few hours or days to test, a genetic modification would take 10 years or longer to test - the time it takes for the genetically engineered embryo to grow into an infant and then a child and adult. For this reason it is most likely that superintelligence will eventually emerge from Artificial Intelligence or Brain-Computer Interfacing research, possibly only within a few decades.

One additional possible route to superintelligence is the use of pharmacological or genetic therapies to boost human brain functioning. These approaches are less “radical” than Brain-Computer Interfacing, and therefore may have less power to truly augment intelligence, but their lesser degree of invasiveness may make the approach more acceptable for research and experimentation.

***

Comments?

7 Responses »

Those of you subscribed to the SL4 list have already seen this, but for everyone else:

Tim already had his bag and overcoat on and his keys in his hand and was about to leave when Diane stopped him at the door.

“I just got this thing working. You have to come and see it.”

“I have a bus to catch.”

“You can get the next one.”

“They’re every half an hour,” he objected. “This had better be good.”

“It’s super-duper. Look at the big screen, it’s easier than squinting at my terminal.”

“Will this take long?”

“A mere instant. Okay, quantum computing, right?”

Read the rest.

No Responses »

The Onion reports:

CHARLOTTE, NC—A top-of-the-line, third-generation Roomba Scheduler robotic floor-cleaning vacuum purchased in January by 35-year-old claims adjuster Ken Graney has inexplicably broken all three laws of Roombotics, a simple yet vital protocol programmed into every Roomba by its manufacturer, iRobot.

Continue.

Link c/o Carl Shulman.

No Responses »

I recently interviewed Robert A. Freitas Jr. on behalf of the Lifeboat Foundation. The interview had 12 questions and ranged from inquiries about his recent work to why hydrocarbons will be the ideal nanofactory feedstock. Freitas’ bio:

A research scientist with molecular nanotechnology company Zyvex, Robert Freitas Jr. has written more than 100 technical papers, book chapters and popular articles on a range of topics and most recently authored Nanomedicine, the first book on the potential medical applications of nanotechnology. In 2002, Freitas was a finalist for the Feynman prize in nanotechnology.

One of Dr. Freitas’ most notable recent efforts is the Nanofactory Collaboration project.

If you like the information Lifeboat is putting out, I encourage becoming a member to support our work.

6 Responses »

The gears are turning over at the Machine Phase blog.

2 Responses »

Videos of talks from the summer 2005 Technology, Entertainment, Design conference are now online. Talks from our friends Ray, Aubrey, and Nick are available and excellent as always. Just be sure to turn down your speakers for the initial intro is loud. For most of you, the material will be nothing new, but seeing them in action and hearing the audience’s response noises is interesting.

Meanwhile, the Google Foundation’s Executive Director, Larry Brilliant, thinks small by talking about fighting global pandemics.

From the “Bold Predictions, Stern Warnings” theme, I like this tagline: “These talks come from speakers who aren’t afraid to go negative — to name the problems they see, and propose bold solutions, with clear-eyed passion.”

There’s a funny thing about being negative. You’re always supposed to propose a solution right away. People tell me this all the time. But guess what: the solution doesn’t always exist. What then? True AI is coming and we have no idea how to set its goals and motivations such that it continues to be nice to us even when it becomes smarter than us and can reprogram itself. There are other unsolved problems, like, “how do we deal with genetically engineered superviruses?”, “what do we do if it becomes radically easier to isolate uranium-235?”, and so on. It’s fun and easy to think that some government employee or academic will work hard and come up with the solution just in time, like in the movies, but it just ain’t so.

3 Responses »

At the foundation of Singularity theory lies the idea of recursive self-improvement. Advances in artificial intelligence or the augmentation of humans is expected to lead to further advances in intelligence enhancement, and so on, until some unknown barrier, perhaps set only by the laws of nature. This unfolding “intelligence explosion” has been called a Singularity by some. Whether or not Singularity is the best word, it has stuck and will continue to stick, so trying to wrestle with the terminology is pointless at this stage.

As you can see by the content on this blog, I’m obviously some sort of a believer of the Singularity idea. Most of the earth’s inhabitants have never heard of the idea and never thought of it, though it did originate in 1965. Since then, it has mostly been discussed by computer scientists working in artificial intelligence, but a decade or two ago it started slowly creeping into the public consciousness - through science/technology and futurist enthusiast types, mostly. The challenge is taking it beyond the geek crowd, and into the scene of regular everyday smart people, like academics, educated professionals, venture capitalists, and the like.

The assumptions underlying the Singularity are pretty basic.

• Intelligence is a phenomenon based on electrochemical activity in the brain.
• Other mediums, like computers, can theoretically implement intelligence.
• The study of intelligence will eventually let us create these implementations.
• At some point, they’ll gain the ability to replicate and self-improve.
• The impact on the world will be huge.

Unfortunately, I’d bet that over 95% of humanity disagrees with the first assumption. The main reason is because the relevant science was only discovered in the last couple hundred years, but humanity has been busy believing that our brains have been running based on divine mandate for tens, if not hundreds of thousands of years. The first book of the Bible and the Torah, Genesis, strongly implies that only God can create humans, animals, plants, etc., from raw materials. According to a recent Newsweek poll, 91% of American adults believe in God.

Most people are dualists. They believe the mind and body are made of two different things, and work in two different ways. The body can be physically repaired by doctors because it’s natural and understandable. The mind can’t be physically repaired because it’s special and immaterial. Nevermind that brain surgery happens every day and thousands of people have brain implants. The Judeo-Christian religions - Christianity, Judaism, and Islam - are fundamentally based on dualistic philosophy.

It’s not easy to convey a five-step argument to the public when 95% of your audience won’t even accept point #1.

The key is to find an audience where the majority accepts the first point. This can be done at places like Google HQ, the computer science department at Stanford, and atheist philosophy clubs. All are good places to start talking about the Singularity.

Conciseness demands tossing out as much unnecessary material as possible when making your argument. Jettisoning superfluous discussion over dualism vs. monism is absolutely essential if you’re going to get anywhere in transmitting the Singularity idea to newbies. You’ve probably argued about it with others for hours already, and there’s a limit to what most people can stand.

A good way to jettison ideas you don’t intend to argue about is to be dismissive and make fun of them. For example, one can say that just because dualists’ ideas are disconnected from reality doesn’t mean that the physics of their brain is. Another gem is this chart from Wikipedia, which describes Descartes’ lame, 17th-century ideas about the mind and body:

God of the gaps, anyone? Many atheists have not resigned from dualist views yet. If you know an atheist whom you suspect is a closet dualist, present them with this chart and ask them what to label the upper triangle, if not “God”. Your “soul”? If the box has no label, then mental and physical events are connected in the same causal fabric, and follow the same basic rules. Which means we are utterly guaranteed to eventually determine the brain’s workings, like we did those of the heart and stomach, and create artificial replacements with superior performance.

Anyone who can reasonably claim that this reverse-engineering process would take over 100 years is almost certainly a closet dualist. I propose that all monists personally expect AI to arrive before 2100.

But many monists, such as the majority of the transhumanist community, are still sketchy regarding the idea of an AI self-replicating and self-enhancing autonomously to the point where it becomes the primary shaper of the world. I propose that such doubts are based on a human superiority complex, plain and simple. Our species is so full of itself that it slaughters millions of animals daily with impunity - why would it acknowledge the possibility of superintelligent AI with abilities greater than ours in all domains?

It is extremely difficult to get people to relinquish their religious beliefs. Dualism is a component of religious belief, thus it seems like attacks on theistic dualism are doomed to failure. As a result of this, I suggest that those who wish to spread acceptance of the Singularity meme focus on the other philosophical flaw I mentioned - the human superiority complex.

Interestingly, one of the best tools for fighting this superiority complex is probably science fiction, which regularly features alien races of all shapes, sizes, and intelligence levels. Greater-than-human intelligence is not hard to imagine for most science fiction fans. But by science fiction, I mean science fiction stories, not Hollywood movies, which portray nonhumans in a way that only reinforces the human superiority complex.

If you want to encourage people to think about the Singularity rationally, getting them to read a little science fiction may be a good bet. But too much science fiction can cause people to get lazy when it comes to considering the nuts and bolts of the situation. People have to find their own balance. Let me just remark that reading no science fiction at all is definitely reading too little.

Featured resources:

Students with easy internet access are favouring going for online courses than attending the courses in person. This is partially because of the ultra-friendly online software available on the site. These sites also have subject specialists who can be consulted via voip. In case of opting out of a course, there are added options of data recovery as well. This kind of information is much more easily available now, thanks to the million web hosting blog fanatics out there.

41 Responses »