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
Because the stated argument was, well, we think the novices will get uplifted first.
The PhDs, they're too good.
It's going to be the people who have a way to go.
They'll be easiest to uplift.
But the work of CLTR in 2004 had said, no, that actually might not be.
Just like software engineers are uplifted by called code much more than people who have never studied coding because even getting to grips with it is a bit of a thing.
That argument doesn't necessarily hold.
And more than that, I've often been a bit suspicious because, like you're saying, it's much easier to run an uplift study with undergraduates.
They are really cheap.
Right.
There's a lot of them.
They're everywhere.
They don't work for very much money.
But just because they are easier to measure than PhD students and postdocs and professors in some sense doesn't mean that they were the right thing to measure first.
It may turn out that is still the case that novices are not meaningfully uplifted with frontier AI.
In one sense, it seems like that's true because lots of companies have activated their so-called ASL 3 mitigations because they are worried they have breached that threshold.
On the other hand, not all the companies have.
And there's never been sort of definitive evidence one way or another.
It's always phrased as we can't rule out the possibility that.
But I really want to dispel this idea that it will be impossible for PhD students and expert groups to be uplifted before novices, that it's going to go in some sort of sequential order.
Showing 421–440 of 1,784 · page 22 of 90
← Previous
Next →