If I spend up to another 100 hours talking to Timothy B. Lee about AI existential risk, will it be productive?
Standard
28
Ṁ2050
Jan 1
30%
chance

Timothy B. Lee is a reporter who writes about AI, economics, and other things.

https://twitter.com/binarybits

https://substack.com/@timothyblee

https://www.understandingai.org/

https://www.vox.com/authors/timothy-b-lee

https://timothyblee.com/

https://www.fullstackeconomics.com/

Most of his work is high quality and informative, and he seems to have a solid understanding of economics and the tech world in general. The exception is any time he writes about AI risk; his articles on that subject tend to contain a lot of fallacious reasoning, indicating a lack of understanding of probability theory and risk management.

He's very open to discussion on this point, as you can see from e.g. the comments on this post. He even reached out to me via email after one of the comments I made in order to continue the discussion more directly. Unfortunately, neither of us has changed their mind on anything substantive so far.

Timothy has a large audience of 10,000+ people and is pretty well known in the tech world, so changing his mind on this subject would be a big win for AI safety, and would easily be worth the time. Conversely, I'm quite concerned about AI, and 100 hours is a minuscule price to pay for the peace of mind that would accompany having a worldview like Timothy's.

If I decide that it's not worth continuing to argue with him, this market resolves N/A. If I decide that it is worth continuing to argue with him, then I commit to spend at least 100 hours doing so (if he's open to that much discussion; I'm not going to harass him), and this market resolves to whether the discussion was productive. If either of us changes their mind in a serious way, it resolves YES, otherwise it resolves NO.

If one of us changes their mind for a reason that's clearly unrelated to our discussion, that doesn't count and the market still resolves NO. But if it's unclear why the mind was changed and it's at least plausible that it was partially due to the discussion, that's good enough and this will resolve YES. (For example, if Timothy posts an article that says he's concerned about the risks and doesn't mention any particular person that changed his mind, that would resolve YES.)

The time limit for evidence is 3 months after the 100 hours is up. If I haven't changed my mind and no evidence has emerged of him changing his within that time frame, it resolves NO.

Get
Ṁ1,000
and
S1.00
Sort by:
predicts NO

I am a bit concerned that if I keep posting info about this it'll drive the price down and result in the market getting cancelled 😬

(Sorry for the spam, I tried using embeds and it didn't work)

predicts NO

@agucova Hmm, how does this affect the probability? It's possible he intends this tweet as a "dunk" or ridicule of me, in which case it should go down. But he didn't phrase it that way, and he blacked out my name, so it seems like he was just making a goodhearted joke about a market about himself.

bought Ṁ200 NO from 30% to 28%

@IsaacKing I don't consider it a dunk on you. I do take it to mean he is only interested in thinking about it for 10-25 hours at most (which could make the conversation lower cost to you and more likely to happen?).

predicts NO

My take is that Tim has spent a lot of time thinking about this, and has spoken to many people with opinions on this subject.

Simultaneously, from the fact you created this market, you must care deeply as well, and are unlikely to change your mind.

Therefore I’d be quite surprised if either of you changed positions.

Were his email responses more substantive than the two paragraphs I see in the comment section? I feel like you wrote a whole lot more than he responded to there.

predicts NO

@RobertCousineau A bit. He asked me to elaborate on my claim that superintelligence might plausibly be able to design a rocket without any tests first.

@IsaacKing Sounds good. For what it's worth, I have not read him before, was unimpressed by the blog post, and then so much more unimpressed by him saying "hey, writing out a detailed response, not ignoring you" and then responding with 2 remarkably uninteresting paragraphs.

predicts NO

@RobertCousineau They haven't responded at all actually, that was someone else.

bought Ṁ80 NO from 31% to 30%

Related: