There are 70 drivers for 3000 vehicles. Which goal is good enough for you? We’ll make a note, I’ll tell you when we passed it, and you can tell me why it’s not real. I’m willing to wait.

I don’t think we’ll ever stop moving the goal posts. You can still meet people who don’t use computers and have never seen the use in them.

I rode in one last month, down the highway.

Even the most pessimistic reports of human involvement still puts them in the ‘mostly self-driving’ camp, and I’d rather have one with a fallback than one without.

Should I disbelieve my lying eyes?

Except it is, and it won’t be.

People are fucking expensive, if you ran the same uncharitable calculations people do for AI on people they would rapidly conclude that there is almost nothing more expensive then having a whole person do something, needing clean water and air all the time, destroying the environment by inefficiently cramming it into their face and then shitting it out a short time later.

Right now, it’s on the line (our current generation of AI is just a little more efficient then something which spends literally years in diapers and needs over a decade of careful and often misguided education just to punch a clock and read some email), but one of these things is getting more efficient and the other one is definitely not.

You can get emotional, maybe burn a data center to the ground or something, but the idea that, ‘what this stuff actually costs to run’ is going to land anywhere close to cost of the people doing it, you’re out of your mind.

How about figuring out how to use this disruption to create systems and technologies which are better? Imagine if the OSS and maker movements started in 1880 instead of 1980.