> it somehow produces a lot more jobs.

Recall a few days ago some people were trying to reconstruct the genome of the first black man in Iceland. There's no way anyone could have a job doing that in anything but a very wealthy society.

In other words, AI won't create jobs in manufacturing, but it will enable unexpected jobs to be created that would otherwise not be cost-effective.

> Recall a few days ago some people were trying to reconstruct the genome of the first black man in Iceland.

Recall who funded that work.

Nobody would have funded or volunteered for that work if they were concerned with getting enough food put away to survive the winter.

Automation reduces costs. As long as those costs savings are passed on to the consumer that's a good thing.

#1 Seems more likely to destroy jobs than create them. It could be an engine for growth, but self driving for example have taken a long time and vast investment. It's useful, but seems to have limited direct benefits.

#2 is a broken window fallacy. It may create some up months, but you need increased productivity for sustained growth.

#3 Biotech has been around for a long time. It's a world of hard problems, and again you need increased effecency not novelty to drive sustained growth.

4) Unmanned mining in space - as automation and machine learning progress capturing the right asteroid could bring in a few Billion or Trillion worth of nickle or whatever.

Honestly, I'm bearish on space mining. The Earth has a lot of unexplored volume left. The oceans' floors are relatively unexplored in terms of mineral wealth, even just 100 miles offshore. Antarctica is also a giant mystery, with many potentially rich mountain ranges sitting under 1/2 mile of ice. Heck, in Australia people will buy land, walk around, and just find nuggets of gold sitting on the dirt, it's something of a 'thing' there [0] So, 10k/lbs flight costs aren't cheap enough yet to justify the very hazardous environment and delta-V costs. Maybe in a century, but for now it's a ways off.

[0] http://www.mining.com/this-australian-man-just-found-a-massi...

Well, there is significant investment in this space (har har), with timelines much closer than a century making investors open their wallets today. And costs aren't really relevant in isolation, it's the ROI that matters.

Space mining is not the answer to things we can find otherwise on the ground. It is the answer to how we build things in space, and get things we can't get in bulk down here. The right rock could reshape our relationship to some of the rare earth metals, for example. Mining en masse in remote regions is no picnic, so if hundreds of millions invested in going up can bring many billions down on demand, the economics will make sense. Mining Antarctica for hundreds of millions of tonnes for a not-so-interesting minerals offers an upside that pales in comparison to the amazing nuggets we have floating around the solar system...

And if the economics are sound, technologically: if you combine a UAV and a brain about as good as we get in a Tesla you're just a few rockets away from some insane riches and the power to crush terrestrial markets on a whim... No soup for you <mineral producer>... that's the kind of thing Billionaires dream of.

The promises of low and micro g manufacturing, the decreasing weight requirements and increase intelligence systems, and holding the economic 'high ground' for humanities next step offer the potential to move the needle more than a little. Launch tech is the big bottleneck, so keep an eye on SpaceX :)

> it's the ROI that matters.

So true!

Your points about some unknown material/process in space being the key are correct. However, I don't think there is anything up there that we can't figure out how to make down here. There are some long, single grain, high weight elemental crystals that we can't make here, but they also takes millions of years to cool down in a vacuum. Space is really empty, and there really aren't any magic rocks up there. It's mostly just feldspar.