Exactly. Yudkowsky probably should not be doing a literature slog on machine ethics, especially since 30% of the papers are about how deontological ethics will save the day. :)
This is a great use for a volunteer network. Those with university access can get on a university PC and have access to JSTOR and so on and just download every single paper that matches certain key terms and email the PDFs to whoever is working on the project. You still need someone who knows the fields, and which keywords to look up, and you need someone to read all the papers and understand them and synthesize them and judge which papers are the most important, but many of these little steps can be done by anybody who knows how to use a search engine.
Part of the research for my friendly AI people was to simply download every single paper that exists on about 10 different keywords, skim-read them all, make a list of all the ones that mattered, and then read those ones in more detail.
It’s not that researchers just “know” all this stuff already when they write a paper. It’s systematic hard work.
Could this work be more distributed? Could different people read the papers related to different concepts and contribute to the presentation that concept within a paper?
On that part, I’m skeptical. Being able to think like a mainstream philosopher about concepts and write (and plan the writing) in that style is a very particular skill that almost nobody is trained in, just like almost nobody is trained in deontic logic or Presidential speechwriting.
I used to think the singularity stuff was kinda silly and too speculative, but that’s because I’d only read little bits of Kurzweil. Then I encountered the Good / Yudkowsky / Chalmers arguments and was persuaded. My interests in philosophy of religion have almost completely disappeared, and working on something other than Friendly AI seems kinda pointless to me now.
So I already have the motivation to do the things I’ve described above. The problem is that it takes a lot of time, and I’m not independently wealthy. I’m only able to make serious progress on my Friendly AI paper and book because I recently quit my IT job and I get to work on them 8 hours per day. But I can’t coast for long. Lots of other people, I suspect, also have this problem of not being independently wealthy. :)
I used to think the singularity stuff was kinda silly and too speculative, but that’s because I’d only read little bits of Kurzweil. Then I encountered the Good / Yudkowsky / Chalmers arguments and was persuaded.
This sounds familiar to me, somehow.
So I already have the motivation to do the things I’ve described above. The problem is that it takes a lot of time, and I’m not independently wealthy. I’m only able to make serious progress on my Friendly AI paper and book because I recently quit my IT job and I get to work on them 8 hours per day. But I can’t coast for long. Lots of other people, I suspect, also have this problem of not being independently wealthy. :)
I know I’ve suggested this before, but the SIAI Visiting Fellows program will take care of living expenses short term, and in your case would likely lead to a long term position as a Research Fellow. It seems that you and they have a lot to offer each other.
My interests in philosophy of religion have almost completely disappeared, and working on something other than Friendly AI seems kinda pointless to me now.
Abrahamic religion seems to be a rather useless time sink to me. Rescuing people from it seems to be worth something—but it usually seems like a lot of effort for few results. A gutter outreach program is messy work as well.
commonsenseatheism.com is starting to seem like a misnomer. Your blog is now mostly about intelligent machines. Time for a new domain?
Exactly. Yudkowsky probably should not be doing a literature slog on machine ethics, especially since 30% of the papers are about how deontological ethics will save the day. :)
So, who is willing to collaborate with SIAI research fellows and do litterature slogs for them?
I vote for Luke.
This is a great use for a volunteer network. Those with university access can get on a university PC and have access to JSTOR and so on and just download every single paper that matches certain key terms and email the PDFs to whoever is working on the project. You still need someone who knows the fields, and which keywords to look up, and you need someone to read all the papers and understand them and synthesize them and judge which papers are the most important, but many of these little steps can be done by anybody who knows how to use a search engine.
Part of the research for my friendly AI people was to simply download every single paper that exists on about 10 different keywords, skim-read them all, make a list of all the ones that mattered, and then read those ones in more detail.
It’s not that researchers just “know” all this stuff already when they write a paper. It’s systematic hard work.
Could this work be more distributed? Could different people read the papers related to different concepts and contribute to the presentation that concept within a paper?
On that part, I’m skeptical. Being able to think like a mainstream philosopher about concepts and write (and plan the writing) in that style is a very particular skill that almost nobody is trained in, just like almost nobody is trained in deontic logic or Presidential speechwriting.
*collaborate
Yeah, that word felt wrong as I typed it, but I couldn’t recall the right one. Fixed.
Though I was hoping for a different sort of response from you. As was Louie.
I used to think the singularity stuff was kinda silly and too speculative, but that’s because I’d only read little bits of Kurzweil. Then I encountered the Good / Yudkowsky / Chalmers arguments and was persuaded. My interests in philosophy of religion have almost completely disappeared, and working on something other than Friendly AI seems kinda pointless to me now.
So I already have the motivation to do the things I’ve described above. The problem is that it takes a lot of time, and I’m not independently wealthy. I’m only able to make serious progress on my Friendly AI paper and book because I recently quit my IT job and I get to work on them 8 hours per day. But I can’t coast for long. Lots of other people, I suspect, also have this problem of not being independently wealthy. :)
This sounds familiar to me, somehow.
I know I’ve suggested this before, but the SIAI Visiting Fellows program will take care of living expenses short term, and in your case would likely lead to a long term position as a Research Fellow. It seems that you and they have a lot to offer each other.
I’m not so sure about the Research Fellow thing, but I was accepted into the SIAI Visiting Fellows program a while back, and that’s why I quit my job.
Curious, why unsure about the Research Fellow thing?
Abrahamic religion seems to be a rather useless time sink to me. Rescuing people from it seems to be worth something—but it usually seems like a lot of effort for few results. A gutter outreach program is messy work as well.
commonsenseatheism.com is starting to seem like a misnomer. Your blog is now mostly about intelligent machines. Time for a new domain?
Yes, if I find the time to do that.