Tuesday, March 31, 2009

Invasive vs. Non-invasive Augmentation

The focus of biological technologies will remain on the understanding of ourselves and other lifeforms, treating diseases, and extending high quality lifespans. On the other hand, actual augmentation of our intellectual and physical powers will, for the most part, happen by proxy via non-invasive devices.

In addition to carefully applying existing and well-established trends for understanding the future, there is another characteristic to all of the future technologies in this blog. They are all noninvasive, that is, they do not rely upon directly "plugging in" to the human brain. There are several reasons for this:

1. It is unnecessary. The value-add of these technologies can be achieved without direct connections between the technologies in question and the human brain.

2. Although many aggressive technology enthusiasts (such as frequent futurist forums) seem to be enamored with this idea of direct tech-to-brain interfaces, I do not believe that the vast majority of the consumer base will be anywhere near so enthused. As Woody Allen said in Sleeper, "No one touches my brain - that's my second favorite organ."

3. Direct brain-to-technology interfaces will be subject to far greater legal penalties if something goes wrong. And it would not take many such suits to move the market from such devices.

4. The technologies discussed in this blog are intended to be utilized for an arbitrary period of time, from a few minutes to several hours or more. Even something as simple as 3D glasses start to annoy after a short while. So although the technologies in this blog are completely compatible with such things as 3-D glasses, virtual-reality helmets, etc., and the software would be virtually identical, they do not require such devices to fulfill their design purposes of utility and/or imagination actualization.

5. The technologies in this blog are meant to be utilized or enjoyed alongside "reality", not exclude the real world altogether. For example, a virtual reality-helmet in essence removes you from reality, it's difficult to impossible to know what's going on in the real world while wearing such an enveloping device. The technologies in this blog provide an extremely realistic virtual experience, without shutting out the real world entirely - you can still hear your baby cry, for instance.

6. For a given amount of computational "augmentation", a separate, stand-alone device will generally be more economical than one directly integrated into the human brain. The demonstration of this is somewhat problematic, because Kurzweil keeps what he means by direct augmentation of the human brain quite vague, using terms such as "smarter" that could mean a great variety of things. However, let's pick just one, specific example - calculation of the square root of 2 to 1,000 decimal places. A stand-alone device with little or no AI could do this for an effective cost of a few pennies or less, and is available with even today's computer technology. However, to augment the power of a human brain to allow the calculation of this "in their head" as it were, is an infinitely more arduous and expensive undertaking. It's unclear what the best way to achieve this is, in any case - via a silicon chip implant? This smacks of "unnecessary surgery", and even if established to be feasible, seems like a highly optional procedure. This leads to the next concern:

7. Brain augmentation - and for that matter, transhuman augmentation of any kind - even if proven safe and feasible, seems to fall into the class of "optional therapy" - ie, not life-saving in the sense of treating a chronic ailment that afflict the disease-stricken. Depending on the nature of the enhancement, it also seems like these kinds of augmentations could also be quite expensive. Extrapolating the current trend of the stinginess of health insurers, it seems reasonable, and in fact, quite reasonable, to suggest that they will NOT cover the expense of such optional procedures. This means that any transhuman enhancements will be out-of-pocket for those desiring such enhancements. If this quite likely scenario proves to be the case, this will mean that the economics will keep the procedures expensive and out of reach for all but the well-to-do. Kurzweil seems to suggest that these types of enhancements will become widespread and happily covered by insurers and/or demanded by the marketplace, but there is absolutely no evidence for this. That doesn't mean it's impossible, not saying that, but it is deeply improbable, and in any case Kurzweil gives no detailed explanation as to why this scenario would come to pass in the first place.

8. A standalone device provides greater scope for offloading of tasks and activities than one directly integrated into the human brain. This ties into the key consumer driver of "convenience" and "making life easier" that is a standard part of successful products, technological and otherwise. If it's directly integrated into one's brain, it may make thinking easier, but it's still the human doing most or all of the work.

9. Inputs and outputs can generally be better standardized and communicated via stand-alone technologies than those directly integrated into the human brain. For example, moving beyond the "square root of 2" scenario to something more involved, say, a detailed weather simulation. Again, this is something that current computers can do, even without appreciable AI as part of the software suite. The inputs for the simulation are input in a standard way, and are output to a screen or printer in a standard way (at least for a given software application). If all this is happening inside someone's head, how do you know the inputs are exactly right, and hence trustworthy? How does that person communicate the results of the simulation anyway - do they drive a monitor or a printer with their thoughts? These and other considerations help build a case that for a given computing task, direct integration into the human brain is in most cases vastly complicates the task, makes the results harder to communicate, less reliable, and much more expensive.

10. All of the above considerations are given even more gravitas with the eventual development of sophisticated AI software that has one of its key objectives successful communication with human beings. This is still forthcoming, to be sure, but will be far ahead of direct brain integration of these technologies. As such, by the time such direct integration is feasible, it will prove unnecessary, as well as suboptimal for the above reasons.

11. People, then as now, will value their privacy, and the privacy of one's most intimate thoughts will be guarded most jealously of all. Therefore, projections of a "global mind" composed of actual, commingling human minds is deeply misguided. A "one-off" global mind consisting of advanced, personalized AI systems that reflect the interests of their primary human interactant(s) but that can also precisely control what they share over such a global network will be far more desirable.

12. Assuming that things such as computer viruses still exist, the idea of an actual "global mind" of commingling human brains that gets infected with a computerized pathogen is a nightmare scenario. Far better to have a "one-off", non-invasive system that gets infected with such a virus, simply because it will be simpler to cleanse such a technology-only platform than to cleanse a composite technological-organic brain of such a virus.

No comments: