> Most (or maybe even all) programmers [...] are worse at their respective field than an AI swarm equipped with a good harness

Is this the consensus view now? I'm not any kind of denier and I do recognise that motivated thinking may cause me to overestimate my own value.

But trying to be brutally neutral I'm still not sure this is true. AI doesn't make certain classes of mistakes that humans do, granted. But I'm still surprised daily even in 2026Q4 at things it does seem to make a mess of. Usually, changes that are locally good but subtly globally poor.