Posts by BenWilson
Last ←Newer Page 1 2 3 4 5 Older→ First
-
Well I've never Godwined a thread before, but hell why not:
It's probably a Godwin corollary that if you write enough on internet forums, the chances of eventually Godwining a thread approach 100%.
-
Ben, I'd be interested on your take on the WoW 'corrupted blood' virus.
A self-replicating virus that managed to escape the lab and spread in a way that it's creators never intended it to...
Heh, it was a fascinating event. It's a pity I missed it. Shows a lot about people that griefing became an instant hit, the idea of spreading the virus as fast as they could just to annoy and frustrate other people, and how the natural reaction to that is to abandon population centers. It's also interesting that most of the people who were playing at the time thought it was a pretty damned cool thing that it happened. The most popular bug of all time.
In terms of what it says about GE/GM? Well, considering it was designed to be a self replicating virus that caused characters damage by the Blizzard programmers, that would put it in the biological weaponry category, something that I think should definitely not be allowed in reality. Real people obviously have to put up with a little bit more than a 5 minute run from the graveyard if they die.
But, OTOH, viral outbreaks happen in nature all the time. I spent all of last week dealing with one, as it happens, up to my elbows in vomit and shit. If GE/GM had any solutions, I'd be bloody glad to hear them. If developing antiviral techniques can be sped up hugely by GE, it would save millions upon millions of humans and animals from unnecessary suffering and death. The chances of them somehow accidentally developing a virus are possibly there, I guess, but there's no particular reason to think it would be any more serious than what I had spewed into my face by my 9 month old last Friday, courtesy of Mother Nature.
-
Ben - algorithms and heuristics are all very well within incredibly small, finite parameters,
Don't I know it, Dyan. I've spent a heck of a lot of time trying to model things with massive numbers of variables, and optimize for them, and these variables are still just a tithe of what could be put in there, but are left out because of intractability. But this does not mean these algorithms are of no use, and/or indicate nothing. Usually there are diminishing returns from increasing the number of variables - the ability to solve some problem does not improve significantly. This is exactly the same for non-computational methods of problem solving.
Again, my reference to them is for analogy purposes only. They've never been intended to model nature - they're a problem solving tool. It's just interesting how many emergent macroscopic features they seem to mimic from nature.
-
Perhaps living in a female body makes it easier to be feminist
Now that I can accept. Seems likely, even, given the comparative numbers that even bother to make the claim that they are feminists.
-
This thread has restored my faith that most people have dirty minds. Keep it up, and I do mean....
-
your monkey's paw is in the post, use it wisely...
Heh, the funny thing about monkey's paws is that in reality the only people who do actually use them are the ones advising against using them.
Edit: Oh and they seem to get a lot more than 3 goes out of it.
-
Nature didn't discover the "efficient optimisation algorithm" - that algorithm, aka natural selection, just is. It's an integral part of nature. Discovered to be so by Darwin.
Not really sure if you're disagreeing with me here. Sounds like a quibble along the lines of "Discovery requires a discoverer, so nature can't discover anything, any more than a stone can". I don't really care to argue about that. It goes to how language is used to convey an idea, rather than implying any deep anthropomorphizing of nature by me. Of course I consider my work to be part of nature, in one sense. But not in the sense people usually mean, so I didn't talk that way.
-
be careful what you wish for...
Righto, Grandma.
Bart, I believe you. Some people don't get that the word experiment means some uncertainty about the outcome, and that it's one of those concepts built into science.
-
Well it's almost certain a lot of animals suffer from experimentation. But it's also true a lot of animals suffer from nature, and normal agriculture slaughters them in their millions every year. When it comes to supplying humans with food, the suffering of animals doesn't usually figure highly. I don't think GM experimentation is more inhumane than a lot of other experimentation, and that includes things as simple as changing feed. You can make an animal suffer or die with mistakes like that, and it just goes down in the log book for next time round.
-
Well, the analogy is loose. The population was always stable, a constant number of individuals - that's how the algorithm was designed. The analogy to mass extinction would be when individuals who were diverse ended up becoming all very similar, because the supersurvivors bred the rest out. But this happened by whatever means the genetic changes happened - random or otherwise. Any change that conferred a big difference in the optimality of the candidate solution led to it crushing out the diversity of the rest of the population. I guess it's like they discovered guns or something. Very quickly they all had to have guns or they were finished. There's no going back, no way that not having guns can keep you ahead of the rest of the competitors.
The danger is that the lucky individuals who found the guns first end up owning the gene pool no matter how weak their rest of their structure might be. But, and I have to keep reiterating this, this is how it worked with random mutation too, indeed more so than the other ways. "Survival of the fittest" is a selfish method, the advantages aren't handed around fairly. And that is what nature, left to its own devices, will do. A cornfield, left to nature, will revert to a forest pretty fast, because trees can grab all the light. Which is wonderfully diverse but it can't support very many humans compared to a cornfield.
In answer to your quantification question, the answer can't just be given as a number, it depended on numerous variables, the aggressiveness of the selection algorithm, the range over which you would let the values mutate, the number of individuals in the population, and just exactly what the change (random or guided mutation, or as a result of crossbreeding) led to vis-a-vis the search space. But I think I can fairly safely say it was "polynomial time", which is what computer scientists usually mean by "fast", when they are solving problems that have exponential search spaces. In the end, using settings that seemed to provide the best convergence, if I ever saw a big jump in the optimality of the best candidate individual, I could be pretty sure that within 10 or 20 generations, if I examined the population, it would all be pretty similar to the individual. It took sophisticated tricks to keep the diversity in there, and doing so was not really worth it in terms my my objective, which was finding optimality in the search space. Mostly the suboptimal individuals I insisted on keeping alive were a waste of resources. I mean like 99.99% of the time.
Now I'm not saying this is all perfectly analogous to nature and GM. The algorithm was not identical, and it certainly did not operate on such an enormous scale, with hundreds of billions of individuals, across millions of years. Also, the "environment" was the objective function I was optimizing, and this remained stable during a run of the algorithm. Most importantly, the objective function itself was decided by me, whereas with real biology, we aren't in control of that to the same extent. We know we want the bigger yields, but we don't want our entire populations to have some weakness that was unknown by us at the start. And we can't just "rerun the algorithm", wiping out the entire population of the world, and expect to get back from microorganisms in tidal pools in only a few days or minutes.
It's "suggestive" at best. I don't really know too much about evolutionary biology. In the end, it was not that relevant.
Last ←Newer Page 1 … 743 744 745 746 747 … 1066 Older→ First