If we’re focusing on Nick’s four classes, can he write up a more in depth description of them and post them here, so that we can think about them in more detail.
Good point. I’ve contacted him. I suppose we should discuss it at a later date instead.
I suggest as an alternative topic of discussion—identifying cascades, cycles, insights and recursive loops that might be available to altruistic actions. An abstract but important issue.
“Cascades are when one development leads the way to another—for example, once you discover gravity, you might find it easier to understand a coiled spring.
Cycles are feedback loops where a process’s output becomes its input on the next round. As the classic example of a fission chain reaction illustrates, a cycle whose underlying processes are continuous, may show qualitative changes of surface behavior—a threshold of criticality—the difference between each neutron leading to the emission of 0.9994 additional neutrons versus each neutron leading to the emission of 1.0006 additional neutrons. k is the effective neutron multiplication factor and I will use it metaphorically.
Insights are items of knowledge that tremendously decrease the cost of solving a wide range of problems—for example, once you have the calculus insight, a whole range of physics problems become a whole lot easier to solve. Insights let you fly through, or teleport through, the solution space, rather than searching it by hand—that is, “insight” represents knowledge about the structure of the search space itself.
and finally,
Recursion is the sort of thing that happens when you hand the AI the object-level problem of “redesign your own cognitive algorithms”.
What concerns me is whether this will lead to specific, new knowledge or concrete actions. I feel like we talk a lot about big ideas in a vague sense and small ideas in a concrete sense. I’d like to be able to talk about big ideas more concretely (though if need be, breaking them down into smaller chunks).
I don’t actually have a recommendation right now (although I plan on talking with a few local people tonight and hopefully generate ideas). But I think as many people as possible should come to this online meetup with a concrete list of things to talk about - last time we covered a few new ideas but the discussion was sort of meandering. I think pre-planned mini-presentations would help a lot.
I have a recommendation: set the focus of the discussion to be “how do we make near-mode progress on far-mode problems?”
Just to prime you on the sort of things I’d expect to come up:
“discussion of abstract concepts”. Many of us feel we already do too much of this, comparatively speaking.
“identify measurable goods produced by orgs like SI or FHI”. We can’t directly measure xrisk, but we can measure things like papers produced or awareness-generating media coverage.
“get expert opinion on the value of these goods” and see how wide the spread of opinion is.
“build good relationships with experts” , try and find out what’s really going on inside their minds and where the core disagreements come from.
“build a consensus on guidelines for rational debate”
Just to be clear, those points aren’t a suggested agenda for a meeting but rather they’re an example of what I’d expect to come out of a meeting—a list of points that a researcher could conceivably start working on tomorrow, and which (while they don’t directly address the main issues) would seem to be aimed at directly tackling relevant stuff.
If we’re focusing on Nick’s four classes, can he write up a more in depth description of them and post them here, so that we can think about them in more detail.
Good point. I’ve contacted him. I suppose we should discuss it at a later date instead.
I suggest as an alternative topic of discussion—identifying cascades, cycles, insights and recursive loops that might be available to altruistic actions. An abstract but important issue.
“Cascades are when one development leads the way to another—for example, once you discover gravity, you might find it easier to understand a coiled spring.
Cycles are feedback loops where a process’s output becomes its input on the next round. As the classic example of a fission chain reaction illustrates, a cycle whose underlying processes are continuous, may show qualitative changes of surface behavior—a threshold of criticality—the difference between each neutron leading to the emission of 0.9994 additional neutrons versus each neutron leading to the emission of 1.0006 additional neutrons. k is the effective neutron multiplication factor and I will use it metaphorically.
Insights are items of knowledge that tremendously decrease the cost of solving a wide range of problems—for example, once you have the calculus insight, a whole range of physics problems become a whole lot easier to solve. Insights let you fly through, or teleport through, the solution space, rather than searching it by hand—that is, “insight” represents knowledge about the structure of the search space itself. and finally,
Recursion is the sort of thing that happens when you hand the AI the object-level problem of “redesign your own cognitive algorithms”.
What concerns me is whether this will lead to specific, new knowledge or concrete actions. I feel like we talk a lot about big ideas in a vague sense and small ideas in a concrete sense. I’d like to be able to talk about big ideas more concretely (though if need be, breaking them down into smaller chunks).
I don’t actually have a recommendation right now (although I plan on talking with a few local people tonight and hopefully generate ideas). But I think as many people as possible should come to this online meetup with a concrete list of things to talk about - last time we covered a few new ideas but the discussion was sort of meandering. I think pre-planned mini-presentations would help a lot.
I have a recommendation: set the focus of the discussion to be “how do we make near-mode progress on far-mode problems?”
Just to prime you on the sort of things I’d expect to come up:
“discussion of abstract concepts”. Many of us feel we already do too much of this, comparatively speaking.
“identify measurable goods produced by orgs like SI or FHI”. We can’t directly measure xrisk, but we can measure things like papers produced or awareness-generating media coverage.
“get expert opinion on the value of these goods” and see how wide the spread of opinion is.
“build good relationships with experts” , try and find out what’s really going on inside their minds and where the core disagreements come from.
“build a consensus on guidelines for rational debate”
Just to be clear, those points aren’t a suggested agenda for a meeting but rather they’re an example of what I’d expect to come out of a meeting—a list of points that a researcher could conceivably start working on tomorrow, and which (while they don’t directly address the main issues) would seem to be aimed at directly tackling relevant stuff.