I, Nanobot — Bionanotechnology is Coming

Maria Williams writes "Alan H. Goldstein, inventor of the A-PRIZE, and popular science columnist, says: Scientists are on the verge of breaking the carbon barrier — creating artificial life and changing forever what it means to be human. And we're not ready... Nanofabricated animats may be infinitesimally tiny, but their electrons will be exactly the same size as ours — and their effect on human reality will be as immeasurable as the universe. Like an inverted SETI program, humanity must now look inward, constantly scanning technology space for animats, or their progenitors. The first alien life may not come from the stars, but from ourselves." Yes it's an older article, but it's a fairly quiet sunday today.

Comments

On the verge of breaking the carbon barrier?

[Timesprout]

Sounds like someone at the lab has a hot date lined up......

The carbon barrier will be broken by silicon.

[fyngyrz]

A pretentious, overheated, paranoid dupe, at that. :) But since we're here:

Life will almost certainly arise from Ai (Ai=artificial intelligence), but it won't be "nano" anything. My feeling as an Ai researcher is that we're already well past the computing threshold for Ai/AL (AL=artificial life.) Ai will always equal AL, though the reverse is not implied. A typical desktop today seems to me to have more than enough power to "come to life." What we are missing is the algorithm, no more.

My reasoning for this is as follows:

The complexity of the part of the brain that we use to think is not as high as commonly supposed. Much of what is in there handles what are really (to a computer) mundane things; pattern recognition (sound, light, touch), movement, of which none of which are components of intelligence, per se (ex, a blind, deaf immobile person is still intelligent.) A computer's sensorium can simply be a text stream; and in this realm, pattern recognition, retrieval and communications are relatively natural to it, and extremely high powered in comparison to how we do the same thing. Likewise, a computer has no autonomic system to regulate, no balance to keep, no hungers to assuage. We know very well how to do associative memory, and we know how to make those associations carry associations of their own; I'm talking of the coloration of memory with any set or range of characteristics one might imagine or preconceive into being. This stuff is all relatively easy. Processing it and making sense of it, that's really the issue. We don't have that.

But: Processing isn't likely to be a hard problem. I say this because the human mind (and all the other minds we are aware of) achieve what they do with massive parallelism of very simple structures. Parallel processing is no different than serial processing, in terms of what it can and cannot do in the end. Speed-wise, yes, parallelism is very powerful, but computationally, it is no more effective than going one step at a time and accumulating results. Further, there is no parallel structure that cannot be emulated or represented with serial processing of those structures in a manner that stages results to the appropriate level of parallelism. Processing is, instead, probably an esoteric problem, in that the way it works is not obvious to our higher-level or aggregate way of considering information, memory, and reason.

Some have argued that because of the massive parallelism in the brain, creating something comparable will require a huge amount of resources that we do not yet have. There is a basic misconception at work here, and it is a critical one: If, as they argue, your desktop is not fast enough to compute serially what the brain does in parallel in the same time frame, this in no way undermines the idea that your desktop can still do the computation. In other words, If you ask me a question, and I give you an answer you deem to be intelligent within, say, 10 seconds of reflection; and you type the same question to your presumptive Ai, and it gives you essentially the same answer in say, an hour, that answer is no less intelligent for being more slowly produced. Intelligence isn't about speed - it never has been. Intelligence is about the nature of the question, the nature of the conversation, the nature of the reflection.

The day that someone comes up with an algorithm that produces answers of a nature that we can begin to argue about how well they compare - or exceed - the human capacity, will be the same day that the hardware companies begin to examine the software and build hardware optimized to make those algorithms faster, compile better, etc. Though this will in no way make anything more intelligent, it will make the intelligence easier to converse with for us, right up until the point when the computer is faster than we are. At that point, again, it will be important for us to remember that speed isn't the issue, and never was. Otherwis

Very Poor Science

[YetAnotherBob]

So this guy has a PHD in Biochemestry? Where does he get his facts??

He states that the human body has a potential of thousands of Volts. In reality it seldom reaches 1 volt. (A potential of 1/2 volt over a distance of a 50 angstroms is not a potential of millions of volts, it may be high in terms of Volts per Meter, but over any larger distance, it just dissapears. That's why we can't power pacemakers or laptop computers off your neural energy. Zero power is just zero power, no matter how clever you think your argument is.) He states that there are strong similarities between Carbon and Silicon chemistry. Yes, but there are also energy differences that are profound. Reality, there are very few living creatures that can use Silicon. Most of those that can are bacteria, and they use it only to create a shell or frame.

A little reality here, there are good reasons to believe that the first engineered bio machines will not be too greatly different than the ones we've already been living with. We call them bacteria. Control is a problem. It has been a problem for a very long time.

The article is just an attempt to scare people with little knowlege of the underlying science. The Author appears to be ignorant of basic physics or chemistry. His biology may or may not be suspect. I don't know that area as well. If this is the best that the critics of nanobot research can do, then they should be doomed to failure.

Rank superstition and scary fiction are a very poor way to make technical policy.