What do you think of contemporary theoretical physics? That is also mostly “arguing on the Internet”.
Some of it yes. At the end of the day though, some of it does lead to real experiments, which need to pay rent. And some of it does quite well at that. Look for example at the recent discovery of the Higgs boson.
What do you think of contemporary theoretical physics? That is also mostly “arguing on the Internet”.
Some of it yes. At the end of the day though, some of it does lead to real experiments, which need to pay rent. And some of it does quite well at that. Look for example at the recent discovery of the Higgs boson.
These theoretical physicists had to argue for several decades until they managed to argue themselves into enough money to hire the thousands of people to design, build and operate a machine that was capable of refuting, or as it turned out—supporting their well motivated hypothesis. Not to mention that the machine necessitated inventing the world wide web, advancing experimental technologies, data processing, and fields too numerous to mention by orders of magnitude compared to what was available at the time.
Perhaps today’s theoretical programmers working on some form of General Artificial Intelligence find themselves faced with comparable challenges.
I don’t know how things must have looked like at the time, perhaps people were wildly optimistic with respect to expected mass of the scalar boson(s) of the (now) Standard Model of physics, but in hindsight, it seems pretty safe to say that the Higgs boson must have been quite impossible for Humanity to experimentally detect back in 1964. Irrefutable metaphysics. Just like string theory, right?
Well, thousands upon thousands of people, billions of dollars, some directly but mostly indirectly (in semiconductors, superconductors, networking, ultra high vacuum technology, etc.) somehow made the impossible… unimpossible.
And as of last week, we can finally say theysucceeded. It’s pretty impressive, if nothing else.
Perhaps M-theory will be forever irrefutable metaphysics to mere humans, perhaps GAI. As Brian Greene put it: “You can’t teach general relativity to a cat.” Yet perhaps we shall see further (now) impossible discoveries made in our lifetimes.
There’s nothing wrong with arguing on the Internet. I’m merely asking whether the belief that “arguing on the Internet is the most important thing anyone can do to help people” is the result of motivated reasoning.
Even if one thought SIAI was the most effective charity one could donate to at the present margin right now, or could realistically locate soon, this would not be true. For instance, if one was extremely smart and effective at CS research, then better to develop one’s skills and take a crack at finding fruitful lines of research that would differentially promote good AI outcomes. Or if one was extremely good at organization and management, especially scholarly management, to create other institutions attacking the problems SIAI is working on more efficiently. A good social scientist or statistician or philosopher could go work at the FHI, or the new Cambridge center on existential risks as an academic. One could make a systematic effort to assess existential risks, GiveWell style, as some folk at the CEA are doing. There are many people whose abilities, temperament, and background differentially suit them to do X better than paying for others to do X.
What do you think of contemporary theoretical physics? That is also mostly “arguing on the Internet”.
Some of it yes. At the end of the day though, some of it does lead to real experiments, which need to pay rent. And some of it does quite well at that. Look for example at the recent discovery of the Higgs boson.
These theoretical physicists had to argue for several decades until they managed to argue themselves into enough money to hire the thousands of people to design, build and operate a machine that was capable of refuting, or as it turned out—supporting their well motivated hypothesis. Not to mention that the machine necessitated inventing the world wide web, advancing experimental technologies, data processing, and fields too numerous to mention by orders of magnitude compared to what was available at the time.
Perhaps today’s theoretical programmers working on some form of General Artificial Intelligence find themselves faced with comparable challenges.
I don’t know how things must have looked like at the time, perhaps people were wildly optimistic with respect to expected mass of the scalar boson(s) of the (now) Standard Model of physics, but in hindsight, it seems pretty safe to say that the Higgs boson must have been quite impossible for Humanity to experimentally detect back in 1964. Irrefutable metaphysics. Just like string theory, right?
Well, thousands upon thousands of people, billions of dollars, some directly but mostly indirectly (in semiconductors, superconductors, networking, ultra high vacuum technology, etc.) somehow made the impossible… unimpossible.
And as of last week, we can finally say they succeeded. It’s pretty impressive, if nothing else.
Perhaps M-theory will be forever irrefutable metaphysics to mere humans, perhaps GAI. As Brian Greene put it: “You can’t teach general relativity to a cat.” Yet perhaps we shall see further (now) impossible discoveries made in our lifetimes.
There’s nothing wrong with arguing on the Internet. I’m merely asking whether the belief that “arguing on the Internet is the most important thing anyone can do to help people” is the result of motivated reasoning.
The argument I see is that donating money to SIAI is the most important thing anyone can do to help people.
Even if one thought SIAI was the most effective charity one could donate to at the present margin right now, or could realistically locate soon, this would not be true. For instance, if one was extremely smart and effective at CS research, then better to develop one’s skills and take a crack at finding fruitful lines of research that would differentially promote good AI outcomes. Or if one was extremely good at organization and management, especially scholarly management, to create other institutions attacking the problems SIAI is working on more efficiently. A good social scientist or statistician or philosopher could go work at the FHI, or the new Cambridge center on existential risks as an academic. One could make a systematic effort to assess existential risks, GiveWell style, as some folk at the CEA are doing. There are many people whose abilities, temperament, and background differentially suit them to do X better than paying for others to do X.