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

Rob Wiblin

πŸ‘€ Speaker
3881 total appearances
Voice ID

Voice Profile Active

This person's voice can be automatically recognized across podcast episodes using AI voice matching.

Voice samples: 1
Confidence: Medium

Appearances Over Time

Podcast Appearances

80,000 Hours Podcast
AI designs genomes from scratch & outperforms virologists at lab work. What could go wrong? | Dr Richard Moulange, CLTR

The idea there would be, I guess, that you're like engineering proteins that have the right shape.

80,000 Hours Podcast
AI designs genomes from scratch & outperforms virologists at lab work. What could go wrong? | Dr Richard Moulange, CLTR

And again, they target conserved parts of the virus that it's going to have a very hard time changing because I guess they're functional.

80,000 Hours Podcast
AI designs genomes from scratch & outperforms virologists at lab work. What could go wrong? | Dr Richard Moulange, CLTR

So if they change the shape, then the function would break down.

80,000 Hours Podcast
AI designs genomes from scratch & outperforms virologists at lab work. What could go wrong? | Dr Richard Moulange, CLTR

And I guess you could like try to do lots of slightly different molecules all at once, the same idea with the vaccines where you do like many at once.

80,000 Hours Podcast
AI designs genomes from scratch & outperforms virologists at lab work. What could go wrong? | Dr Richard Moulange, CLTR

Okay, just to come back to the concern I had with this whole approach earlier.

80,000 Hours Podcast
AI designs genomes from scratch & outperforms virologists at lab work. What could go wrong? | Dr Richard Moulange, CLTR

We're imagining here that we're in a world with significantly advanced biological tools that are being used to enable the creation of these broad-spectrum antivirals, broad-spectrum vaccines.

80,000 Hours Podcast
AI designs genomes from scratch & outperforms virologists at lab work. What could go wrong? | Dr Richard Moulange, CLTR

So it's a lot easier for people to go from a function to a sequence to engineer specific viruses, I guess, that have particular characteristics.

80,000 Hours Podcast
AI designs genomes from scratch & outperforms virologists at lab work. What could go wrong? | Dr Richard Moulange, CLTR

if you're up against a very capable, like an intelligent adversary, aren't they going to find out what antivirals you've stockpiled, what vaccines you've given your armed forces, and then basically just cherry pick a design for a virus that isn't captured by this one?

80,000 Hours Podcast
AI designs genomes from scratch & outperforms virologists at lab work. What could go wrong? | Dr Richard Moulange, CLTR

I mean, they might have a hard time, but they can just keep searching until they eventually find something.

80,000 Hours Podcast
AI designs genomes from scratch & outperforms virologists at lab work. What could go wrong? | Dr Richard Moulange, CLTR

It's just like defense, offense thing that they can just keep searching until they find a weakness.

80,000 Hours Podcast
AI designs genomes from scratch & outperforms virologists at lab work. What could go wrong? | Dr Richard Moulange, CLTR

And then if, I suppose you would say, initially, they're not gonna potentially have that capability because they just don't have the same resourcing, unless they're a state, I guess, as the defenders, where there's just a lot more money and a lot more people who are interested in preventing disease than causing it.

80,000 Hours Podcast
AI designs genomes from scratch & outperforms virologists at lab work. What could go wrong? | Dr Richard Moulange, CLTR

then eventually I guess they might catch up and they might be able to find the one virus design that evades your various defenses.

80,000 Hours Podcast
AI designs genomes from scratch & outperforms virologists at lab work. What could go wrong? | Dr Richard Moulange, CLTR

I guess by that point, perhaps we might have built up the manufacturing capacity and the scientific capacity to very quickly respond to any new threats, to design a new antiviral and to manufacture it en masse and distribute it within 100 days or whatever.

80,000 Hours Podcast
AI designs genomes from scratch & outperforms virologists at lab work. What could go wrong? | Dr Richard Moulange, CLTR

And so maybe again, we achieve some kind of balance that's tolerable.

80,000 Hours Podcast
AI designs genomes from scratch & outperforms virologists at lab work. What could go wrong? | Dr Richard Moulange, CLTR

Okay, let's talk about another technology we might want to differentially advance.

80,000 Hours Podcast
AI designs genomes from scratch & outperforms virologists at lab work. What could go wrong? | Dr Richard Moulange, CLTR

Yeah, you and your organization, including my wife, I was a co-author on this one, you recently published a cost-benefit analysis suggesting that screening DNA sequences that people have requested to synthesize for scientific purposes...

80,000 Hours Podcast
AI designs genomes from scratch & outperforms virologists at lab work. What could go wrong? | Dr Richard Moulange, CLTR

that screening them to make sure that they're not dangerous one way or another passes a cost benefit test, even if the UK goes out alone and only does it for sequences that are requested in the UK.

80,000 Hours Podcast
AI designs genomes from scratch & outperforms virologists at lab work. What could go wrong? | Dr Richard Moulange, CLTR

Well, I think people won't understand given the threats that we've been talking about, given like how many people could die in a biological catastrophe, why in general it might be a good idea to screen.

80,000 Hours Podcast
AI designs genomes from scratch & outperforms virologists at lab work. What could go wrong? | Dr Richard Moulange, CLTR

But I think a common objection has been, well, if only one country does it, then a bad actor will just request the sequence in the mail from another country that isn't doing the screening.

80,000 Hours Podcast
AI designs genomes from scratch & outperforms virologists at lab work. What could go wrong? | Dr Richard Moulange, CLTR

How then can it be that it passes a cost-benefit test for the UK to go out alone and potentially do it, even if maybe other countries don't follow its example?