I'm a huge fan of Objective-C, but I'll still admit that it loses to C++ in terms of performance. Message passing and the lack of objects on the stack mean it will never be quite as fast.
Still, it's a shame it's not more prevalent, because it's a wonderful language for a lot of real-world tasks.
What is an "industrial strength development tool"?
A programming language which you can use to develop, debug and distribute a real-world, functioning application in, I'm guessing. Something like, say, Firefox. Use native APIs, native GUIs, run at decent speed, behave like a native app.
I'm not forced at gun point to read everybody else's blogs, I get to pick and choose what I read and when I read it. And if you're like most people in the "blogosphere", you will pick and choose those who you already agree with, turning the whole thing into one big echo chamber where everybody's prejudices and preconceived notions are only ever reinforced and never challenged.
Personally, I bought my mum an iMac. Sadly I was unlucky enough to do so only a month or two before Apple switched to x86. Why is that so unlucky? Does it really bother your mom that her computer's running on PPC?
The prevailing knowledge two years ago was that global warming was going to increase the number and severity of hurricanes like Katrina. The scientific data was being used to try to explain the circumstances of the time. The "prevailing knowledge" in the media, maybe. Certainly not among actual climate scientists. The two are actually different, and you would do well to realize that, so that you do not make statements like:
The point is that there is a groupthink that has formed, where the analysis is being made to fit a particular view of what will be happening to the world. The question has become "how can this data show that climate change is dangerous" instead of "what does this data say about what will happen"? Because now you are accusing an entire field of science of being non-scientific. You're going to need some proof to back such a incredibly bold claim.
The assumption is purely that it would be better at designing ultraintelligent machines than us.
No. There is a hidden assumption, and it is as I said.
Consider this case: We design an AI which is 100% more intelligent than us (we ignore the fact that "more intelligent" here is a very loose term, and close to meaningless). This AI can design an even more intelligent AI. However, that AI is only 10% more intelligent than itself. And that AI can only design another AI that is 1% more intelligent than itself.
No singularity, no nerd rapture, just slightly smarter computers, because it turns out designing AIs is not a linear function.
Yeah, this thread was born from a misreading of the original poster, assuming it was the kind of post you so often see on Slashdot: "This is so stupid! Those scientists totally don't understand what they are doing! Look, I know much better!"
I was just replying in the spirit of such a discussion, even though it wasn't really called for in this case.
I like Neuromancer, but really, it's fairly cheap entertainment and I can easily see how somebody could with very good reason call it pseudo-intellectual garbage. Because it is, and it just happens that some of us are entertained by that kind of thing. If you aren't, it's not going to have much to offer you.
Actually I think if you follow the logic course of our current evolution as well as observing network theories, a singularity type event is all but assured.
And if you followed the logic cource of the evolution of the Roman Empire, it would conquer the entire world in a couple centuries.
Assuming a current trend will continue indefinitely is highly naïve.
As for this idea of technology actually achieving this event horizon described by Good or Gibson or Vinge, I don't think that it's achievable. I can't prove it won't happen just like you can't prove it will happen. All I will say is that I don't even know where to begin. Actually, that's not the real problem. The real problem is here:
an ultraintelligent machine could design even better machines
This is an entirely unfounded assumption. Or maybe I should say, there is an unstated assumption: That any successively more intelligent machine can design a machine that is more intelligent than itself by a constant factor, or at least by a factor large enough that the series converges.
That is a very bold claim to make on no evidence. I seriously doubt it would hold true.
Those are all space operas, which, depending on who you're talking to, are either a subgenre of sci-fi or not sci-fi at all. Gibson writes a lot of hard science fiction, along with authors like David Brin, Charles Stross, Vernor Vinge, and (to an extent) Arthur C. Clarke. In hard sci-fi most of the emphasis is on the scientific details/accuracy, with the story often just being a path the author takes you through their scientifically rigorous vision.
You know, from reading that list of authors, I get the feeling that your definition of "hard sci-fi" has less to do with scientific accuracy and more with who you like enough to suspend your disbelief of their glaring inaccuracies and silly fantasies. The only person on that list I'd call even remotely scientifically accurate is Arthur C. Clarke, and then only when he felt like it.
There is nothing wrong with enjoying fantastic fiction, but dressing it up as "scientific accuracy" is pretty silly.
Wow, a slashdotter doesn't understand an article but still feels the need to post about how he's much smarter than everyone else.
Re:LLVM / C++ backend is a total showstopper
on
GCC 4.2.1 Released
·
· Score: 1
So you want to throw out a far superior optimizer, which would speed up pretty much all code compiled in gcc, to shave some minutes off bootstrapping, which happens... statistically speaking, never?
Slashdot Mirror
The point here was, you know, that that quote indicated nothing of the sort, and was entirely irrelevant to the original post.
If that's what you wanted to do, what was that whole spiel about?
Then why did you reply in the first place?
You know, while that is true, it in no way refutes the grandparent's argument.
Maybe, you know, just maybe, it's not a good idea to stuff crap that will expand and burn in the cracks between the heat shield tiles?
I'm a huge fan of Objective-C, but I'll still admit that it loses to C++ in terms of performance. Message passing and the lack of objects on the stack mean it will never be quite as fast.
Still, it's a shame it's not more prevalent, because it's a wonderful language for a lot of real-world tasks.
What is an "industrial strength development tool"?
A programming language which you can use to develop, debug and distribute a real-world, functioning application in, I'm guessing. Something like, say, Firefox. Use native APIs, native GUIs, run at decent speed, behave like a native app.
All true, but the machine was good enough when you bought it. Does your mom have such a tech-lust that it's no longer good enough now?
No, thanks.
Hahahah, here I thought I was making a personal argument, but it turns out I was making a general one! Well, a wikipedia-editor one, at least.
The assumption is purely that it would be better at designing ultraintelligent machines than us.
No. There is a hidden assumption, and it is as I said.
Consider this case: We design an AI which is 100% more intelligent than us (we ignore the fact that "more intelligent" here is a very loose term, and close to meaningless). This AI can design an even more intelligent AI. However, that AI is only 10% more intelligent than itself. And that AI can only design another AI that is 1% more intelligent than itself.
No singularity, no nerd rapture, just slightly smarter computers, because it turns out designing AIs is not a linear function.
Yeah, this thread was born from a misreading of the original poster, assuming it was the kind of post you so often see on Slashdot: "This is so stupid! Those scientists totally don't understand what they are doing! Look, I know much better!"
I was just replying in the spirit of such a discussion, even though it wasn't really called for in this case.
Intellectual and honest, yup - a real bitch. Nothing is more threatening to nerds with an over-inflated sense of self-worth.
Nice strawmanning there.
I like Neuromancer, but really, it's fairly cheap entertainment and I can easily see how somebody could with very good reason call it pseudo-intellectual garbage. Because it is, and it just happens that some of us are entertained by that kind of thing. If you aren't, it's not going to have much to offer you.
Actually I think if you follow the logic course of our current evolution as well as observing network theories, a singularity type event is all but assured.
And if you followed the logic cource of the evolution of the Roman Empire, it would conquer the entire world in a couple centuries.
Assuming a current trend will continue indefinitely is highly naïve.
an ultraintelligent machine could design even better machines
This is an entirely unfounded assumption. Or maybe I should say, there is an unstated assumption: That any successively more intelligent machine can design a machine that is more intelligent than itself by a constant factor, or at least by a factor large enough that the series converges.
That is a very bold claim to make on no evidence. I seriously doubt it would hold true.
Those are all space operas, which, depending on who you're talking to, are either a subgenre of sci-fi or not sci-fi at all. Gibson writes a lot of hard science fiction, along with authors like David Brin, Charles Stross, Vernor Vinge, and (to an extent) Arthur C. Clarke. In hard sci-fi most of the emphasis is on the scientific details/accuracy, with the story often just being a path the author takes you through their scientifically rigorous vision.
You know, from reading that list of authors, I get the feeling that your definition of "hard sci-fi" has less to do with scientific accuracy and more with who you like enough to suspend your disbelief of their glaring inaccuracies and silly fantasies. The only person on that list I'd call even remotely scientifically accurate is Arthur C. Clarke, and then only when he felt like it.
There is nothing wrong with enjoying fantastic fiction, but dressing it up as "scientific accuracy" is pretty silly.
Seriously, not believing people to be infallible is not a character flaw or sign of stupidity.
Maybe not, but believing yourself to be smarter than pretty much everyone else definitely is.
How about you go look at the results of an entire fucking field of science, you goddamn fool?
You think Hard Gay is "repressed"?
Wow, a slashdotter doesn't understand an article but still feels the need to post about how he's much smarter than everyone else.
So you want to throw out a far superior optimizer, which would speed up pretty much all code compiled in gcc, to shave some minutes off bootstrapping, which happens... statistically speaking, never?