A confused otter’s mumblings and rumblings

Monday, 6th November, 2006

Lack of AI, lack of HI

Filed under: AI, Humans — Otter @ 20:00

This is probably a repeat. That just goes to show how right I might be. I know that might convince you. I know I can’t know that.

To the matter at hand: there is still no AI around. I don’t mean that there aren’t machines capable of solving problems explained to them in a language they can understand — that has been the case for decades with general-purpose computers, and arguably includes even the simplest machines if performing physical tasks through physical manipulation were not excluded — but rather that there is no machine capable of being given any problem to solve in a human language and going ahead and solving it. Only humans come close to achieving this goal.

Why is this? I believe, at the moment, (for all that my beliefs are worth, especially the transient ones), that it is due entirely to the arrogance of those humans. All the examples of supposed AI research that I’ve come across fail to realise that the algorithms the intelligent machine must follow should be flexible enough to deal with unforseen circumstances in a “rational” way, defined as relative to its goals (which I assume would be, given our needs, to solve the given problem — or, equivalently in my mind, to reach the desired defined goal — as “resourcefully” as possible, where, again, we define the preciousness of resources as we wish).

Unforeseen circumstances come about because of uncertainty about “the world”. This uncertainty, and how to deal with it, must therefore be considered when designing the algorithm to endow the machine with. This is not hard: Bayesian statistics, anyone?

That’s it. Comments welcome. Read this too, which I link to if only so that there’ll be a link to another blog in this post. Isn’t that one

Thursday, 9th February, 2006

No title yet, but it might possibly be this

Filed under: AI, Humans — Otter @ 01:56

I think now that I got that point or so across, I can get on with some more “normal” writing. Just don’t say you haven’t been warned that you are at best reading something that will become very famous, at worst something that will destroy the world (or at least humanity) [the event with possibly the largest discrepancy between deserved and expected fame], and in some maximum/expected probability way, something that will not achieve anything of “significance”, which while being merely defined as its effect on humanity, subjectively defined, almost completely subjectively determined, is by definition “important”, “crucial”, influential… Without it, you’re masturbating. But I warned you of that, so you shouldn’t be surprised.

[Start the profiling now, if you like. He (and of course it’s a “he”) mentioned probability — so probably trained in probability. Obviously “intelligent”, though you probably don’t realise how little, because it’s relative to the average human, who isn’t that bright in absolute terms. Writes a blog: what a goldmine! Anonymously, yet tries to help his identification. Who is this freak? This self-deprecating freak. Why can’t he just shut up and do something “normal”? He already does; you’ll find it described by many as “living his life”. Who are you to judge me? Don’t you dare, you failure. I can manipulate you, theoretically, although my practice still leaves much to be desired. And I’m glad for it. I’m a nice person. I do care about other people, otherwise why would I have written all that aggressive crap? I’m trying to believe in something simpler, easier, easier. But I just can’t. I don’t believe any of it.]

And so on to my idea: artificial intelligence. Heard of it before? Yes, I decided to use a term you’ve heard of, though who knows if I mean it in the same way as that and those who said it to you before. Now you might be a total idiot, but that’s not your fault.. You’re just built that way. But given that, you’re quite smart. You can certainly read this, although you will make all sorts of assumptions whilst doing so, and you’re not really reading every word, are you? You’re likely to miss it if I put the same word at the end of a line and at the start of the next one. I mean, you’ll see the word, but only one instance of it. But you knew what I mean. See? Smart, but necessarily stupid.

Which somehow, according to me (and you’ll probably agree, you schmuck, you nebbish), brings us to artificial intelligence. Defining “intelligence”, without the quotes, as the ability to perform a task, and letting “artificial” mean created much more directly than usual by humans [Don’t forget it’s bollocks…], you’ll respond, were you both logical and outspoken about your logical conclusion, that we already have artificial intelligence. You’ll perhaps cite examples, examples which for example might include a vacuum cleaner. You’re lucky to be able to think of it and not need an actual, physical example of a vacuum cleaner at hand to make the point to some humans who matter in your scheme of thing. Appreciate it, and marvel.

So if we already have artificial intelligence, you are likely to continue in that cute human way of yours, what is my point? My point is, for the sake of argument, that current aritificial intelligence research focuses too much on creating very unintelligent artifices. What is needed is something much more intelligent (invoking the naive definition, not that you are likely to care, given that you’ve let so many others get away with worse, if it really is worse), which will be able to learn how to do things. Is that so hard? Why can’t we just get a program to sense things, make connections, and just make more connections, while also sensing new things, making yet more connections? We can, and I will work on it. It’s the only project worth working on, except, I argue, making sure I stay alive, which encompasses a wider range of activities on my and your part than you I believe you were likely to realise at first.

Wasn’t this all terribly interesting? I bet you feel better for having read it, if you did read it. If you didn’t, you really should. If there are any ways you can make me richer, you should do those too. Excellent.

Blog at WordPress.com.