![[personal profile]](https://www.dreamwidth.org/img/silk/identity/user.png){kind=small}
Complexity, simplicity, springtime
Last night I went to see Stephen Wolfram, the author of A New Kind of Science, give a talk at BU. He was trying to get across the main points of a 1200 page book in 90 minutes so some of it moved remarkably fast, but I liked the talk a lot more than I thought I might (largely because Wolfram didn't seem quite the egomaniac I'd heard he was). The first half of the talk was about how you can get behavior from simple things that's about as complex as it gets, even when there's no randomness in the starting conditions (cellular automata with certain rules, etc). The more you look for insanely simple rules that give rise to insanely complex behavior, the more of them you find, intimating that even really big things (the universe) might be the result of some set of rules that's a lot smaller than we're currently looking for. The second half focused more on that limit of complexity; that is, various rule sets lead to the "as complex as it gets" limit fairly quickly once you get beyond really simple patterns, and then everything is equally complex (irreducible, like data you can't compress). So an intuitive leap or two later, you can postulate that even the rules giving rise to human beings probably aren't much complex than any others. What we have that's special is our history, which never could have been predicted because the process happening as time passes in the universe is technically irreducible--that is, there is no computation that can tell us what will happen in 1000 years that could ever work faster than simply running the universe for 1000 years. Same for turbulent fluid flow and so forth (which he did cover in his lecture before talking about people--HE was fairly coherent--unfortunately you're now reading my summary of his summary, which is less so).
Leaving aside whether he cited others as often as he should in his book, which seems to be a sore spot with some, I actually liked the results of his research. The first half has some practical applications in brute-force searching for efficient algorithms(need something clever done? Well, we have a Cray and 7 billion pretty simple Turing machines to iterate over; maybe one of them will happen to do it really efficiently. Let's look for it.). He's actually done some of this for certain functions in Mathematica, which he was clever enough to make a fortune off of before going off into all this speculative science. The second half keeps the future happily unknowable (and confirms the whole "might as well act as if we have free will" idea, which I find an especially nice way to live).
All this aside, I was at least as interested in just seeing someone who had closeted himself away for so long kind of redoing things from scratch, rather as A is doing with his new computer language (which is universal as of last week--yay!). I wanted him to come with me to the talk for that reason, but I guess he didn't understand why I wanted him to go, and he didn't. Now he wishes he had, but too late. :( Doesn't matter that much, really... I just hope that what he's doing is more like Mathematica than like a 1200 page book that a lot of people think is crackpot. This whole trust business is so scary for me. I have to let him do his thing though. Someday I may do mine and it may be at least as freakish.
Now, today, a new and long-running seminar on artificial intelligence is starting up at work. It's cool. I bogged the whole discussion down trying to figure out if an agent that gives no output can be considered an agent (I say, if it's changing itself in response to stimulus, yes! but only if you consider the agent a part of the environment, and the book we're reading says the agent is separate from the environment. Errrrrrggggh). I'm kind of excited about the seminar since it's being run by a good friend and I still have almost no background in a lot of this basic stuff. I'm going to try to be good and read the book at home on the weekends when it won't be such a chore.
So, I guess the summary is that I'm still kind of focused on nerdy/philosophical stuff. To me it is artsy as well. There are probably cool poems and screensavers in all these these ideas somewhere, if I cared to dig them out.
I forgot to say that this morning I saw not-so-small green plants growing UNDER a layer of transparent ice in my neighbor's yard. Some were in little melted spots but some were completely under the ice and seemed to be growing just as well as the others. You see the craziest things in spring.
Leaving aside whether he cited others as often as he should in his book, which seems to be a sore spot with some, I actually liked the results of his research. The first half has some practical applications in brute-force searching for efficient algorithms(need something clever done? Well, we have a Cray and 7 billion pretty simple Turing machines to iterate over; maybe one of them will happen to do it really efficiently. Let's look for it.). He's actually done some of this for certain functions in Mathematica, which he was clever enough to make a fortune off of before going off into all this speculative science. The second half keeps the future happily unknowable (and confirms the whole "might as well act as if we have free will" idea, which I find an especially nice way to live).
All this aside, I was at least as interested in just seeing someone who had closeted himself away for so long kind of redoing things from scratch, rather as A is doing with his new computer language (which is universal as of last week--yay!). I wanted him to come with me to the talk for that reason, but I guess he didn't understand why I wanted him to go, and he didn't. Now he wishes he had, but too late. :( Doesn't matter that much, really... I just hope that what he's doing is more like Mathematica than like a 1200 page book that a lot of people think is crackpot. This whole trust business is so scary for me. I have to let him do his thing though. Someday I may do mine and it may be at least as freakish.
Now, today, a new and long-running seminar on artificial intelligence is starting up at work. It's cool. I bogged the whole discussion down trying to figure out if an agent that gives no output can be considered an agent (I say, if it's changing itself in response to stimulus, yes! but only if you consider the agent a part of the environment, and the book we're reading says the agent is separate from the environment. Errrrrrggggh). I'm kind of excited about the seminar since it's being run by a good friend and I still have almost no background in a lot of this basic stuff. I'm going to try to be good and read the book at home on the weekends when it won't be such a chore.
So, I guess the summary is that I'm still kind of focused on nerdy/philosophical stuff. To me it is artsy as well. There are probably cool poems and screensavers in all these these ideas somewhere, if I cared to dig them out.
I forgot to say that this morning I saw not-so-small green plants growing UNDER a layer of transparent ice in my neighbor's yard. Some were in little melted spots but some were completely under the ice and seemed to be growing just as well as the others. You see the craziest things in spring.