As long as the SIAI continues to not publish, or otherwise make available, credible documents indicating rigorous progress it is going to be perceived as lacking in rigour. And those of us who aren’t privy to what is actually going on in there may worry that this indicates an actual lack of rigour.
I couldn’t agree more.
This is why I talk almost non-stop within Singularity Institute about how we need to be publishing the research that we’re doing. It’s why I’ve been trying to squeeze in hours (around helping with the Summit and now being Executive Director) that allow me to author and co-author papers that summarize the current state of research, like ‘The Singularity and Machine Ethics’ and many others that are in progress: ‘Intelligence Explosion: Evidence and Import’, ‘How to Do Research That Contributes Toward a Positive Singularity’, and Open Problems in Friendly Artificial Intelligence. Granted, only the last one could constitute significant research progress, but one reason it’s hard to make research progress is that not even the basics have been summarized with good form and clarity anywhere, so I’m first working on these kinds of “platform” documents as enablers of future research progress.
My concern with showing the research that’s going on is also why, in the video above, I repeatedly asked for people with experience writing up research papers to contact me.
Eliezer once wrote about how our lack of a PhD on staff and other common complaints didn’t seem to be people’s “true rejection” of Singularity Institute, but I think the “you don’t publish enough research” is a pretty decent candidate for being many people’s true rejection.
Believe me, few things would make me happier than having the resources to publish those 30-40 papers I talked about that are sitting in people’s heads but not on paper.
I couldn’t agree more.
This is why I talk almost non-stop within Singularity Institute about how we need to be publishing the research that we’re doing. It’s why I’ve been trying to squeeze in hours (around helping with the Summit and now being Executive Director) that allow me to author and co-author papers that summarize the current state of research, like ‘The Singularity and Machine Ethics’ and many others that are in progress: ‘Intelligence Explosion: Evidence and Import’, ‘How to Do Research That Contributes Toward a Positive Singularity’, and Open Problems in Friendly Artificial Intelligence. Granted, only the last one could constitute significant research progress, but one reason it’s hard to make research progress is that not even the basics have been summarized with good form and clarity anywhere, so I’m first working on these kinds of “platform” documents as enablers of future research progress.
My concern with showing the research that’s going on is also why, in the video above, I repeatedly asked for people with experience writing up research papers to contact me.
Eliezer once wrote about how our lack of a PhD on staff and other common complaints didn’t seem to be people’s “true rejection” of Singularity Institute, but I think the “you don’t publish enough research” is a pretty decent candidate for being many people’s true rejection.
Believe me, few things would make me happier than having the resources to publish those 30-40 papers I talked about that are sitting in people’s heads but not on paper.
So it sounds like your answer is: “Publishing research would help, and we’re working on it.”
That’s great! It’s just good that you’ve got a plan. After all, the question was “How are you going to address the perceived lack of rigour”.
Correct!