Note that while this is an effective method to steer an asteroid, the difficulty scales with its mass and delay before impact. This would not be practical in a (very unlikely) "don't look up" scenario, where the asteroid is huge and discovered only a couple of months from impact.

>It’s a mistake to view this as just a threat to unskilled labor.

Absolutely. In many cases it's the complete opposite: intellectual jobs are often easier to automate than manual ones.

Building a bot that can do *everything* a good plumber can do pretty much means cracking AGI. On the other hand, some very smart people do some difficult, but ultimately very narrowly-bounded tasks that a computer will do better.

As an extreme/forced example, look at chess: it's a "smart person" activity where computers utterly dominate.