A setback for AI?

I don't think anyone ever thought AI was going to be easy. Even if you did think of individual brain cells as very simple devices, a human brain has some hundred billion of the little critters. Worse (for AI researchers, not for us) are the connections between cells: each human brain cell connects to a thousand or so other brain cells. Those connections make modeling a human brain pretty daunting. This paper estimates that you'd need to start with "about twenty million billion calculations per second, give or take a couple of orders of magnitude."

From that guesstimate, it might seem that all you'd need is to map the neural connections (get the mechanical organization down pat) and pack in enough electronics to match, and poof, there's your AI brain. Getting to that point is more than difficult now, but Moore's law will help us out and we should be able to do that in a decade or two. So watch out: those bionic brains may be kicking us around before mid-century!

Well, no, I don't think so. A few days ago the science sites reported a new study of rat's brain cells that indicates a lot of power in individual cells. This study was concerned with the sensation of touch, but earlier studies have shown individual cells apparently understanding "two-ness" and even being able to recognize faces. That's a long, long way from the sort of simple on-off decision making role that we envisioned years ago.

Here's a little clip from that "recognizing faces" study:

He noted that in one participant, one brain cell responded both to Aniston and to Lisa Kudrow, her co-star on the TV hit ''Friends.''

So we are not talking about just wiring up a hundred billion logic gates. This is a hundred billion things each with a tremendous amount of individual computing power. Can we even begin to approach that level of complexity and power in a few decades? Probably not..

We will get there. Brains are just machines, there's no magic that drives them. The task may even be a little less difficult than it looks now, and advances in electronics and AI software may point to shortcuts we can take - nothing says that biological brain organization is the most efficient or most simple way to solve the given problem. But it does seem plain that we aren't "close" in any sense. Super intelligent electronic "brains" aren't likely in our near future.

Jump to Comments

Many of the products and books I review are things I purchased for my own use. Some were given to me specifically for the purpose of reviewing them. I resell or can earn commissions from the sale of some of these items. Links within these pages may be affiliate links that pay me for referring you to them. That's mostly insignificant amounts of money; whenever it is not I have made my relationship plain. I also may own stock in companies mentioned here. If you have any question, please do feel free to contact me.

Specific links that take you to pages that allow you to purchase the item I reviewed are very likely to pay me a commission. Many of the books I review were given to me by the publishers specifically for the purpose of writing a review. These gifts and referral fees do not affect my opinions; I often give bad reviews anyway.

We use Google third-party advertising companies to serve ads when you visit our website. These companies may use information (not including your name, address, email address, or telephone number) about your visits to this and other websites in order to provide advertisements about goods and services of interest to you. If you would like more information about this practice and to know your choices about not having this information used by these companies, click here.