Interesting article. There are twwo main points I would raise.

Firarly, combining evolution with social sciences has a very long and largely unpleasant history. This is not to say you shouldn't do it, but every alarm should be set on "highest sensitivity". It is easily ised to justify all sorts of things that are repugnant.

The second thing is that "Greed 1" is essentially a hill climbing algorithm and so inherently produces sub-optimal solutions. I would like to hear his views on that.

But let's stick to Brexit!