I think the post implies that the most important defining factor of AGI (as defined by this post) is RSI capability in one form or another—the ability for the AI to adapt itself to different economically useful tasks by itself.
(This is technically broader than RSI, but it includes RSI because the task of “making itself better” is among the list of tasks that an AGI can adapt itself to improve at. Also, achieving RSI probably also achieves AGI, given sufficient RSI capability ceiling and a modest amount of capability generalization into diverse tasks.)
Hmm, from my perspective, this post has nothing to do with RSI.
Like if I get a new job, and learn on the job for the first few months, am I engaging in RSI? I think most people would say “no, that’s not what RSI means”. So would I. It’s not merely “technically broader than RSI”, rather it’s a different thing entirely.
I agree that AGI will be able to figure out how to program ever better AI, just like it will be able to figure out how to design wind tunnels and grow tomatoes and everything else.
I think it may be worth mentioning basic time dilation from the point of view of the agent. It may not be the right audience, but for me, this is the most salient fact from the human point of view.
I like this categorization of AGI.
I think the post implies that the most important defining factor of AGI (as defined by this post) is RSI capability in one form or another—the ability for the AI to adapt itself to different economically useful tasks by itself.
(This is technically broader than RSI, but it includes RSI because the task of “making itself better” is among the list of tasks that an AGI can adapt itself to improve at. Also, achieving RSI probably also achieves AGI, given sufficient RSI capability ceiling and a modest amount of capability generalization into diverse tasks.)
Hmm, from my perspective, this post has nothing to do with RSI.
Like if I get a new job, and learn on the job for the first few months, am I engaging in RSI? I think most people would say “no, that’s not what RSI means”. So would I. It’s not merely “technically broader than RSI”, rather it’s a different thing entirely.
I agree that AGI will be able to figure out how to program ever better AI, just like it will be able to figure out how to design wind tunnels and grow tomatoes and everything else.
I was trying to make a separate observation, apologies if I made it sound like I think your post is about RSI. I agree it’s not.
I just wanted to point out that I thought that achieving your definition of AGI leads to RSI, and achieving RSI leads to AGI fairly soon.
I think it may be worth mentioning basic time dilation from the point of view of the agent. It may not be the right audience, but for me, this is the most salient fact from the human point of view.