Menu
Sign In Search Podcasts Charts People & Topics Add Podcast API Blog Pricing

David Duvenaud

๐Ÿ‘ค Speaker
1059 total appearances

Appearances Over Time

Podcast Appearances

80,000 Hours Podcast
Why 'Aligned AI' Could Still Kill Democracy | David Duvenaud, ex-Anthropic team lead

I mean, yes.

80,000 Hours Podcast
Why 'Aligned AI' Could Still Kill Democracy | David Duvenaud, ex-Anthropic team lead

I think I've mentioned it in passing, but that absolutely, I think, is a big part of what we can do now.

80,000 Hours Podcast
Why 'Aligned AI' Could Still Kill Democracy | David Duvenaud, ex-Anthropic team lead

And it's maybe not even all that under-incentivized, though.

80,000 Hours Podcast
Why 'Aligned AI' Could Still Kill Democracy | David Duvenaud, ex-Anthropic team lead

And yeah, AIs are just going to...

80,000 Hours Podcast
Why 'Aligned AI' Could Still Kill Democracy | David Duvenaud, ex-Anthropic team lead

Even things like crypto or whatever, there's just all these experimental things that will let us have much more powerful coordination mechanisms than we have in the future.

80,000 Hours Podcast
Why 'Aligned AI' Could Still Kill Democracy | David Duvenaud, ex-Anthropic team lead

And I think that's a big part of avoiding doom, is developing those before we are disempowered in more serious ways.

80,000 Hours Podcast
Why 'Aligned AI' Could Still Kill Democracy | David Duvenaud, ex-Anthropic team lead

But all of this is dual use because

80,000 Hours Podcast
Why 'Aligned AI' Could Still Kill Democracy | David Duvenaud, ex-Anthropic team lead

It's harder for humans to participate in these coordination mechanisms than AIs, so it's not really clear which way speeding this up cuts.

80,000 Hours Podcast
Why 'Aligned AI' Could Still Kill Democracy | David Duvenaud, ex-Anthropic team lead

But in general, I'd probably be in favor of developing it faster.

80,000 Hours Podcast
Why 'Aligned AI' Could Still Kill Democracy | David Duvenaud, ex-Anthropic team lead

Well, the way it backfires is like, you know, think of like people are always trying to build like decentralized autonomous organizations.

80,000 Hours Podcast
Why 'Aligned AI' Could Still Kill Democracy | David Duvenaud, ex-Anthropic team lead

And it's like kind of doesn't work for like a number of reasons.

80,000 Hours Podcast
Why 'Aligned AI' Could Still Kill Democracy | David Duvenaud, ex-Anthropic team lead

But that's the sort of thing where people could accidentally create these sort of like self-replicating like beings or polities that get to live in some sort of margins and be like really hard for us to shut down.

80,000 Hours Podcast
Why 'Aligned AI' Could Still Kill Democracy | David Duvenaud, ex-Anthropic team lead

And I think people are going to be constantly like seeding the world with these attempts at like self-replication.

80,000 Hours Podcast
Why 'Aligned AI' Could Still Kill Democracy | David Duvenaud, ex-Anthropic team lead

sustaining machine life and civilization, just for various reasons.

80,000 Hours Podcast
Why 'Aligned AI' Could Still Kill Democracy | David Duvenaud, ex-Anthropic team lead

Like I'm saying, I want to be on the right side of history.

80,000 Hours Podcast
Why 'Aligned AI' Could Still Kill Democracy | David Duvenaud, ex-Anthropic team lead

I think it's like every month for the foreseeable future, there's going to be at least one attempt where someone's like, I gave this AI a Bitcoin wallet and asked it to go make money or something like that and try to start these self-sustaining little societies.

80,000 Hours Podcast
Why 'Aligned AI' Could Still Kill Democracy | David Duvenaud, ex-Anthropic team lead

And to the extent that those AIs are able to just coordinate amongst themselves, that's another source of danger.

80,000 Hours Podcast
Why 'Aligned AI' Could Still Kill Democracy | David Duvenaud, ex-Anthropic team lead

Yeah, so I'm not doing much of this.

80,000 Hours Podcast
Why 'Aligned AI' Could Still Kill Democracy | David Duvenaud, ex-Anthropic team lead

This is like Deir, Jan, some other people.

80,000 Hours Podcast
Why 'Aligned AI' Could Still Kill Democracy | David Duvenaud, ex-Anthropic team lead

The idea is to try to operationalize some concrete questions that help us at least settle the questions of, you know, do we think that this kind of loss of control that I'm talking about is going to happen?