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

Daniel Kokotajlo

๐Ÿ‘ค Speaker
See mentions of this person in podcasts
617 total appearances

Appearances Over Time

Podcast Appearances

Dwarkesh Podcast
2027 Intelligence Explosion: Month-by-Month Model โ€” Scott Alexander & Daniel Kokotajlo

And now we have genetic evidence for the Yamnaya, which we believe is this group.

Dwarkesh Podcast
2027 Intelligence Explosion: Month-by-Month Model โ€” Scott Alexander & Daniel Kokotajlo

You have a blog where you do this.

Dwarkesh Podcast
2027 Intelligence Explosion: Month-by-Month Model โ€” Scott Alexander & Daniel Kokotajlo

This is your job, Scott.

Dwarkesh Podcast
2027 Intelligence Explosion: Month-by-Month Model โ€” Scott Alexander & Daniel Kokotajlo

So why shouldn't we hold the fact that language models can't do this more against them?

Dwarkesh Podcast
2027 Intelligence Explosion: Month-by-Month Model โ€” Scott Alexander & Daniel Kokotajlo

I mean, it seems like such an economically valuable... But how would you set up the training environment?

Dwarkesh Podcast
2027 Intelligence Explosion: Month-by-Month Model โ€” Scott Alexander & Daniel Kokotajlo

Well, maybe that's why you should have longer timelines.

Dwarkesh Podcast
2027 Intelligence Explosion: Month-by-Month Model โ€” Scott Alexander & Daniel Kokotajlo

It's a gnarly engineering problem.

Dwarkesh Podcast
2027 Intelligence Explosion: Month-by-Month Model โ€” Scott Alexander & Daniel Kokotajlo

And then let me just address that in a second, but just one final thought on this thread.

Dwarkesh Podcast
2027 Intelligence Explosion: Month-by-Month Model โ€” Scott Alexander & Daniel Kokotajlo

To the extent that there's like a modus ponens, modus tollens thing here, where one thing you could say is like, look, AIs, not just LLMs, but AIs will have this fundamental asymmetric advantage where they know all this shit.

Dwarkesh Podcast
2027 Intelligence Explosion: Month-by-Month Model โ€” Scott Alexander & Daniel Kokotajlo

And why aren't they able to use their general intelligence to use this asymmetric advantage to some enormous capability overhang?

Dwarkesh Podcast
2027 Intelligence Explosion: Month-by-Month Model โ€” Scott Alexander & Daniel Kokotajlo

Now, you could infer that same statement by saying, okay, well, once they do have that general intelligence, they will be able to use their asymmetric advantage to make all these enormous gains that humans are, in principle, less capable of, right?

Dwarkesh Podcast
2027 Intelligence Explosion: Month-by-Month Model โ€” Scott Alexander & Daniel Kokotajlo

So basically, if you do subscribe to this view that AIs could do all these things if only they had general intelligence, you're going to be like, well, once we actually do get the AGI, it's actually going to be totally transformative because they will have all of human knowledge memorized and they can use that to make all these connections.

Dwarkesh Podcast
2027 Intelligence Explosion: Month-by-Month Model โ€” Scott Alexander & Daniel Kokotajlo

You're so conservative, Daniel.

Dwarkesh Podcast
2027 Intelligence Explosion: Month-by-Month Model โ€” Scott Alexander & Daniel Kokotajlo

Look, your AI product works best when it has access to all of your client's information.

Dwarkesh Podcast
2027 Intelligence Explosion: Month-by-Month Model โ€” Scott Alexander & Daniel Kokotajlo

Your copilot needs your customer's entire code base, and your chatbot needs to access all of your client's documents.

Dwarkesh Podcast
2027 Intelligence Explosion: Month-by-Month Model โ€” Scott Alexander & Daniel Kokotajlo

But for your users, this presents a massive security risk.

Dwarkesh Podcast
2027 Intelligence Explosion: Month-by-Month Model โ€” Scott Alexander & Daniel Kokotajlo

Your client's IT people are going to need more assurance to even consider your product.

Dwarkesh Podcast
2027 Intelligence Explosion: Month-by-Month Model โ€” Scott Alexander & Daniel Kokotajlo

Enterprise customers need secure auth, granular access controls, robust logging, and a lot more just to start with.

Dwarkesh Podcast
2027 Intelligence Explosion: Month-by-Month Model โ€” Scott Alexander & Daniel Kokotajlo

Building all these features from scratch is expensive, tough, and time-consuming.

Dwarkesh Podcast
2027 Intelligence Explosion: Month-by-Month Model โ€” Scott Alexander & Daniel Kokotajlo

That's where my new sponsor, WorkOS, comes in.