Conversation on AI risk with Adam Gleave

post by AI Impacts · 2019-12-27T21:43:37.085Z · EA · GW · 3 comments

This is a link post for



You can see a full transcript of this conversation on our website.



We spoke with Adam Gleave on August 27, 2019. Here is a brief summary of that conversation:


Comments sorted by top scores.

comment by ofer · 2019-12-27T23:49:13.230Z · EA(p) · GW(p)
Gleave thinks discontinuous progress in AI is extremely unlikely:

I'm confused about this point. Did Adam Gleave explicitly say that he thinks discontinuous progress is "extremely unlikely" (or something to this effect)?

From the transcript I get a sense of a less confident estimate being made:

Adam Gleave: [...] I don’t see much reason for AI progress to be discontinuous in particular.

Adam Gleave: [...] I don’t expect there to be a discontinuity, in the sense of, we just see this sudden jump.
comment by Ben_West · 2019-12-31T01:01:50.051Z · EA(p) · GW(p)
GDP will double in 6 months before it doubles in 24 months

Does anyone have the original version of this? The transcript says that Adam is (perhaps incorrectly) paraphrasing Paul Christiano.

I think I have some intuition about what this is getting at, but I don't think this statement is precisely correct (surely GDP has to double in 24 months before it doubles in six months, as long as there is at least 24 months of data).

Replies from: Tetraspace Grouping
comment by Tetraspace Grouping · 2019-12-31T01:35:56.381Z · EA(p) · GW(p)

Christiano operationalises a slow takeoff as

There will be a complete 4 year interval in which world output doubles, before the first 1 year interval in which world output doubles.

in Takeoff speeds, and a fast takeoff as one where there isn't a complete 4 year interval before the first 1 year interval.