Dr. Richard Moulange

speaker

1,784 appearances 1 recordings 1 series first heard Mar 2026 last heard 31 Mar

Dr. Richard Moulange’s voice in public audio — every appearance, attributed to the second.

Trend

recordings per month · last 12 months

Recordings per month over the last 12 months — 1 in all, peaking in Mar 2026 with 1.

Appearances

newest first · ▶ plays the moment

80,000 Hours Podcast · #241 – Richard Moulange on how now AI codes viable genomes from scratch and outperforms virologists at lab work — what could go wrong? · 31 Mar 2026 podcast

Is that the optimal way to share access?

Do I want my defenders, my random members of the public, and my malicious actors all to be able to use it at the same time?

No, all things considered, I would rather have the most trusted people who are building defensive technologies to access it first.

If I had to choose who would get it first, I would definitely pick them.

And more than that, I think there are opportunities for sort of virtuous feedback circles.

The frontier companies have said, we want our tools to be used by national security to deal with pressing challenges.

We've seen multiple companies, you know, offering their tools to government departments for famously $1 a department, this sort of thing.

Well, if that's true, you should be thinking about sharing those tools in advance because it is those people you said, you say you're trying to empower, you are trying to uplift.

We don't just want to avoid malicious uplift.

We want to enable defensive uplift.

To do that, we have to generate data about what it is those activities look like and how those people use frontier AI and how they could be using it better.

This is a really good point.

There's actually a great paper from RAND Europe last year that looked at the future bottlenecks to more misuse-relevant biodesign and said that data was the one governance bottleneck that probably halts.

It's about not running the experiments that we haven't even conceived of yet that could be used to train the most dangerous model.

So I quite agree there.

How do we bring the scientific community on board?

Slowly and carefully and with much sweat, I suppose.

I want to draw attention to good efforts that are already taking place.

So David Baker, who we mentioned earlier, Nobel Prize winner and George Church and a few others, luminaries in the field.

have already set up what's called sort of the Responsible AI-Enabled Protein Design Consortium.

Showing 861–880 of 1,784 · page 44 of 90