So just in very broad terms, where do you think we are in the AI cycle and why has it become such a dominant feature of investment market thinking at the moment?

So let me, just for those who aren't familiar with the cycle that we talked about, and it's not a mini cycle, there's bigger things going on, but let me just talk to it.

If you've got some kind of transformative technology, you know, almost by definition, you start off with the concept and the feasibility.

And if you're going to raise capital for it, you have to get into the promotional phase.

So you have to make all sorts of claims about what it can do.

And these are technologies that are underpinned by science.

It's not made up, but it's some distance from commerciality.

Is that a fair summary of what's going on there?

So what should they be concluding from the Besson Binder

data? So one of the issues in Besson Binder is that If you think of that 4% as a concentration number, which I don't think you should, but if you do think of it that way, it will almost inexorably decline because the denominator inexorably increases as companies fail and the mortality rate goes up.

So if the denominator is consistently going up and the number of the numerator is relatively stable, which it seems to be, that number will fall, but there's nothing happening in concentration.

If you take the S&P, you take out some of the denominator effect.

So it's much more realistic.

You've still got to be very careful because there are other issues in this because it's not time-weighted returns.

It's dollar aggregate.

This may be what the Depp S&W data is actually telling us, as opposed to keep looking for the 4%.

Please give him a round of applause.

Excuse me.

Thanks, David.

So I'm going to make two arguments simultaneously.

Andrew kind of set me up and, if you like, put some words into my mouth that I haven't actually spoken.

So I do feel the need to correct some of those.

I think the starting point was wood for the trees.

Sam?