The people that are excited by this nonsense have no regard for the wonderful complexity of the world. It’s so terribly reductionist.

The same DARPA Grand Challenge in 2004 inspired the investment in self driving cars. That, checks notes, still aren’t a thing 22 years later.

To be fair Musk also planned a whole load of Tesla factories that either never opened or will soon shut down.

We hit the run rate when measured over a 30 min period after a giant marketing push.

What if I tell you we can pay relatively little to someone to own or lease the car, open the doors, deal with luggage, wash, maintain, store the vehicle and drive the passenger to their destination.

'Purpose' can is second the request for an edit button.

I would extend this to argue that general purpsoe automation is a fools errand and will always be outperformed by specialist automation systems. A dishwasher is a great domestic robot, a conveyor belt is pretty amazing at moving things etc.

Thank you, and good luck with whatever comes next.

I call it demoware. Genuinely amazing on first use, but doesn’t solve a problem anybody has. Consequently easy to raise money, very hard to make money

I don't think I said we are at the end of history. Only that I believe standards can have real value often overlooked by the tech community. Documentation, institutional knowledge, deep understanding of constraints, edge cases etc. Not judging your personal tool preference.

Tech hasn't really learned that the value of standardisation exceeds the value of innovation in mature products. Everything has to be pointlessly reinvented.

Haven't we spent the last decade of the internet fighting bots. If we wanted "agents" to book a flight wouldn't travel sites etc, just give us an API to actually build robust automations with.

Strong Microsoft HoloLens army contract energy.