Artificial Intelligence Meets Natural Stupidity (1976) [pdf] | Hacker News Reader

Artificial Intelligence Meets Natural Stupidity (1976) [pdf]

homepage.univie.ac.at

74 pointskick8 comments

McDermott's paper is a classic for a reason. Back at the dawn of the first wave of AI hype, he was taking people to task over wishful thinking, and over applying aspirational labels to mundane computations (e.g. "thinking", "solving" for simple graph searches) thus making them sound more powerful than they were. And over building only toy problem solutions, under the assumption that the results would scale to real world problems (spoiler: they never did).

We need this kind of insight for this current wave of AI hype. Once again we're seeing a lot of wishful thinking, and a lot of aspirational labels (e.g. that DNNs are "thinking" up solutions etc). And inevitable extrapolation of toy solutions, assuming they will scale to real world scenarios (e.g. the fully autonomous driving fiasco). And the hype does not correspond to reality.

I have a problem with how we use the word "thinking", usually implying something somewhat magical, or at least something more than the type of computations done in an algorithm or in a DNN. My belief is what we call "thinking" is nothing more than a (vastly more) complicated system doing the same kind of elementary calculation that are done in a neural net. There's nothing magical in our brain, it's entirely mechanical (chemical really), what we do when we "think" is, I think, comparable to what happen in a DNN (if not in scale at least in nature). Consciousness and thinking are illusions created by our "mechanical" brains. People who say machines will never "think" the way we do are, I believe, mistaken.

I don't disagree that the "thinking" process is entirely a physical process (electro-chemical, etc.).

However, we're not even at a point yet where we can articulate specifically what that physical process _is_ - much less reproduce simplified artificial versions of it.

To imply we've somehow captured the essence of thinking in a DNN - and that it just needs to get bigger and more complex - that is exactly the type of thing this guy is mocking (deservedly so).

Not sure how representative I am, but I can say that when I'm trying to solve an abstract problem, there is definitely a sort of dialog process that goes on in my head. I mean that literally, as in, subvocalized English words saying things like "How shall I proceed? Well, what if I assume such-and-such... Oh, wait, that wouldn't help at all, nevermind. Ok, then what about..." There's no such process in any of these AI systems.

Of course, this doesn't apply to knee-jerk reactions. I don't carry on an inner monologue in order to realize "That thing over there is a stop-sign".

A lot of what people call "Artificial Intelligence" these days would be much better described as "Artificial Knee-Jerk Reactions".

Thinking is that which is left when you remove all pattern matching. Pattern matching is the bulk of our mental activity.

This is quoted a lot at the beginning of my favorite LessWrong article, “Truly Part if You”, in particular:

>As McDermott says, “The whole problem is getting the hearer to notice what it has been told. Not ‘understand,’ but ‘notice.’ ” Suppose that instead the physicist told you, “Light is made of little curvy things.” Would you notice any difference of anticipated experience?

>How can you realize that you shouldn’t trust your seeming knowledge that “light is waves”? One test you could apply is asking, “Could I regenerate this knowledge if it were somehow deleted from my mind?”

https://www.lesswrong.com/posts/fg9fXrHpeaDD6pEPL/truly-part...

I haven't seen LessWrong mentioned in quite some time! I remember the first time I heard about them, from this article in Harper's:

https://harpers.org/archive/2015/01/come-with-us-if-you-want...

LessWrong is great as intellectually stimulating entertainment but I don't think they are actually any good at all for making people more rational.

I think the bigger problem there is that they take themselves entirely too serious. It all comes across as bombastic and mostly based on wishful thinking. Nice intro material but you can't build a castle on a thin foundation like that.

Reply on news.ycombinator.com