Rory Stewart

Then we have the threat of Russia to Europe.

What are we doing with Russia to Europe?

Are we trying to be Ukraine, defense in depth, cheap drones, or are we trying to be the US and Israel?

Let's have exquisite planes and take out Russia's air defenses on day one and then be able to use our fancy planes to take out all that stuff.

When's Russia going to attack?

Because if they're attacking in 2030, having a plane delivered in 2040 is not much easier.

If they're attacking in 2040,

Then there's my final point.

Actually, we need to be investing in AI.

I'd actually scrap most of this stuff and put it into building data sensors, frontier models, because actually, what are we learning?

Modern warfare is about new software every six weeks, new hardware every six months.

There's no point pretending.

that we know what's going to happen in 2040.

What we have to do is build the engineers, the systems, the industrial manufacturing capacity that is able to do new software every six weeks, new hardware every six months.

And that's basically leaning into software engineers, industrial production capacity, ability to build munitions, rather than betting on exquisite platforms.

And which is why I think pretentious things like kind of the ecosystem or the environment of British innovation, technology, talent.

infrastructure is probably much more important than trying to guess what the weapon or what the war will be in 15 years' time.

Just a quick plug for a brilliant final six-page Economist article by my friend Shashank Joshi, who tries to lay out what he calls transparency.

He thinks war is all about

transparency, that these devices can see things they were never able to see before.