Comment author: gsastry 26 March 2018 06:01:11PM 3 points [-]

What key metrics do research analysts pay attention to in the course of their work? More broadly, how do employees know that they're doing a good job?

Comment author: Richard_Batty 03 February 2017 04:17:41PM *  6 points [-]

Is there an equivalent to 'concrete problems in AI' for strategic research? If I was a researcher interested in strategy I'd have three questions: 'What even is AI strategy research?', 'What sort of skills are relevant?', 'What are some specific problems that I could work on?' A 'concrete problems'-like paper would help with all three.

Comment author: gsastry 07 February 2017 02:54:20AM *  2 points [-]

Luke Muehlhauser posted a list of strategic questions here: (originally posted in 2014).

Comment author: TsviBT 12 October 2016 06:03:46PM 7 points [-]

In my current view, MIRI’s main contributions are (1) producing research on highly-capable aligned AI that won’t be produced by default by academia or industry; (2) helping steer academia and industry towards working on aligned AI; and (3) producing strategic knowledge of how to reduce existential risk from highly-capable AI. I think (1) and (3) are MIRI’s current strong suits. This is not easy to verify without technical background and domain knowledge, but at least for my own thinking I’m impressed enough with these points to find MIRI very worthwhile to work with.

If (1) were not strong, and (2) were no stronger than currently, I would trust (3) somewhat less, and I would give up on MIRI. If (1) became difficult or impossible because (2) was done, i.e. if academia and/or industry were already doing all the important safety research, I’d see MIRI as much less crucial, unless there was a pivot to remaining neglected tasks in reducing existential risk from AI. If (2) looked too difficult (though there is already significant success, in part due to MIRI, FHI, and FLI), and (1) were not proceeding fast enough, and my “time until game-changing AI” estimates were small enough, then I’d probably do something different.

Comment author: gsastry 06 December 2016 06:07:58AM *  1 point [-]

By (3), do you mean the publications that are listed under "forecasting" on MIRI's publications page?

Comment author: AnnaSalamon 01 December 2016 08:38:56AM *  15 points [-]

I suspect it’s worth forming an explicit model of how much work “should” be understandable by what kinds of parties at what stage in scientific research.

To summarize my own take:

It seems to me that research moves down a pathway from (1) "totally inarticulate glimmer in the mind of a single researcher" to (2) "half-verbal intuition one can share with a few officemates, or others with very similar prejudices" to (3) "thingy that many in a field bother to read, and most find somewhat interesting, but that there's still no agreement about the value of" to (4) "clear, explicitly statable work whose value is universally recognized valuable within its field". (At each stage, a good chunk of work falls away as a mirage.)

In "The Structure of Scientific Revolutions", Thomas Kuhn argues that fields begin in a "preparadigm" state in which nobody's work gets past (3). (He gives a bunch of historical examples that seem to meet this pattern.)

Kuhn’s claim seems right to me, and AI Safety work seems to me to be in a "preparadigm" state in that there is no work past stage (3) now. (Paul's work is perhaps closest, but there is are still important unknowns / disagreement about foundations, whether it'll work out, etc.)

It seems to me one needs epistemic humility more in a preparadigm state, because, in such states, the correct perspective is in an important sense just not discovered yet. One has guesses, but the guesses cannot be established in common as established knowledge.

It also seems to me that the work of getting from (3) to (4) (or from 1 or 2 to 3, for that matter) is hard, that moving along this spectrum requires technical research (it basically is a core research activity), and one shouldn't be surprised if it sometimes takes years -- even in cases where the research is good. (This seems to me to also be true in e.g. math departments, but to be extra hard in preparadigm fields.)

(Disclaimer: I'm on the MIRI board, and I worked at MIRI from 2008-2012, but I'm speaking only for myself here.)

Comment author: gsastry 04 December 2016 09:56:50PM *  1 point [-]

I agree that this makes sense in the "ideal" world, where potential donors have better mental models of this sort of research pathway, and have found this sort of thinking useful as a potential donor.

From an organizational perspective, I think MIRI should put more effort into producing visible explanations of their work (well, depending on their strategy to get funding). As worries about AI risk become more widely known, there will be a larger pool potential donations to research in the area. MIRI risks becoming out-competed by others who are better at explaining how their work decreases risk from advanced AI (I think this concern applies both to talent and money, but here I'm specifically talking about money).

High-touch, extremely large donors will probably get better explanations, reports on progress, etc from organizations, but the pool of potential $ from donors who just read what's available online may be very large, and very influenced by clear explanations about the work. This pool of donors is also more subject to network effects, cultural norms, and memes. Given that MIRI is running public fundraisers to close funding gaps, it seems that they do rely on these sorts of donors for essential funding. Ideally, they'd just have a bunch of unrestricted funding to keep them secure forever (including allaying the risk of potential geopolitical crises and macroeconomic downturns).

Comment author: gsastry 12 October 2016 11:47:02PM 2 points [-]

Do you share Open Phil's view that there is a > 10% chance of transformative AI (defined as in Open Phil's post) in the next 20 years? What signposts would alert you that transformative AI is near?

Relatedly, suppose that transformative AI will happen within about 20 years (not necessarily a self improving AGI). Can you explain how MIRI's research will be relevant in such a near-term scenario (e.g. if it happens by scaling up deep learning methods)?

Comment author: gsastry 12 October 2016 11:37:09PM *  4 points [-]

The authors of the "Concrete Problems in AI safety" paper distinguish between misuse risks and accident risks. Do you think in these terms, and how does your roadmap address misuse risk?