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 monthsRecordings per month over the last 12 months — 1 in all, peaking in Mar 2026 with 1.
Appearances
That is not true.
We can do worse than nature.
This is true in all aspects of science.
There are so many examples where we engineer things
better than the nature has ever provided.
We can make materials that are much stronger than anything in nature.
That is not the ceiling.
And so we should be deeply concerned about the ability for AI to uplift, say, even the Russian Federation to build things worse than we have ever seen on Earth.
You're right.
that it's not just the evals, but a lot of the governance work, the risk management work to date has focused on that particular threat model.
And I think this is most notable with the frontier safety frameworks that the frontier companies use.
A lot of them at their sort of so-called high-risk or AI safety level three threshold, and different companies use different names, center around this idea of AI could meaningfully uplift a novice, a non-expert, to build a known biological threat.
I've always been a bit unsatisfied with this as the thing that the industry centered around.
I think it's really valuable that the industry centered around something at all.
And this, I think, really came through the work of the Frontier Model Forum, which is an industry group that several different frontier labs are part of.
And it allows them to come together and share best practice on risk management and risk mitigation.
I think that's an excellent mechanism.
But in 2004 and a bit in 2025, when they centered around novice uplift, I sort of went...
hang on, is this because that's what you're concerned about?
Or is it because that's sort of the easiest thing to measure?
Showing 401–420 of 1,784 · page 21 of 90
← Previous
Next →