For 1),

I think the difference is the mechanism of automation. The cotton gin meant the millions cotton pickers were out of cotton picking work, but this was being somewhat offset by increased low-skill, low-intelligence factory work elsewhere. Inventions disrupted and enabled greater individual productivity in single verticals of work: cotton picking, logistics (steam engines), circular saw, etc. These productivity innovations enabled a single person to do the work of dozens or hundreds.

Today, robots are removing the need for a human at all in not just one but effectively every low skill, low intelligence job. We're not far from a world where the great automated farms that feed the majority of the US are run by a few people making sure a swarm of end to end farming machines aren't broken down. This will happen in the next 10 years. Slightly different robots are already automating logistics in warehouses, and soon they'll be driving our commercial trucks. They'll fly our planes, build our buildings, our cars, our electronics. Oh, you say, but someone needs to build the robots. Robots build the robots. They build the robots that build the robots. There will need to be people who design and program and repair the robots, but none of that is low skill or low education; the cotton picker of today couldn't build or design or repair a robot. It's impossible to predict the future, but the nature of accelerating returns suggests that in the next 50 years we will replace almost every low skill job worth automating. Notable exceptions are human facing jobs in the service industry, though I expect innovation will change things pretty wildly there as well. Looking at you, accountants.

The TL;DR: the difference between today and the industrial revolution is that robotics is capable of automating almost any low skill job, rather than enabling greater productivity for humans doing the job.

My main point is why will the automation from 2014-2024 be so qualitatively different than the automation from 2004-2014 such that society will be forced to phase in unconditional basic incomes or else be in substantial trouble? (And to be clear this is the question for nkoren. I agree with everything that you said. :D)

3) Growth curves are curves, not linear. As per 2) changes in efficiencies across the board (end use, production, transmission) along with the point of use for solar means that the grid will be less relevant in the future, not more.

1.) Yes, automation has been replacing jobs (and creating new ones) since the beginning of the industrial revolution. The difference is that when it destroyed agricultural jobs, for example, it was relatively easy for workers to transfer to new jobs in factories with relatively little training. If you had reasonable manual dexterity and half a brain, you could make the shift in a few weeks. Today, the new jobs being created require VASTLY more education to enter, making the mobility of labour type much more difficult than in the past. This has actually been a growing problem since the mid 1970s, when industrial robotics began eliminating most factory labour and microcomputers began eliminating most secretarial labour. Not coincidentally, that's the point where wages and productivity gains became completely decoupled, as human labour became less valuable. This trend has been accelerating continuously since then, and will continue to accelerate into the next decade -- at the same time as the difficulty of entering new jobs is moving in the other direction. This has now led to pervasive structural unemployment and underemployment; those problems will continue to become worse. Left unchecked, they are fully capable of destabilising societies.

2.) Yes, the apparatus of the oil industry has a long lifespan and will have tremendous inertia, and I think we're in agreement that the balance of power production will be nowhere close to having shifted by then. However by the end of this decade, the writing will very much be on the wall as far as the petrochemicals are concerned, and energy companies and governments will be much more proactive about scrambling for a position (territory and IP-wise) in the post-oil world than they are today -- even though the actual balance of production will probably take a further 20 years to shift to solar.

3.) What's different about the next 10, relative to the last 50? 50 years ago, the Western Democracies hadn't dropped taxes on the rich to historically low levels, leading to the most extreme wealth inequalities in history; hadn't built a domestic surveillance apparatus surpassing the wildest dreams of the East German Stasi; hadn't militarised their police forces; and (fairly brief Red Scares notwithstanding) were more outwardly-directed in their paranoia, rather than declaring perpetual "wars" (largely upon their own populations) in the names drugs and terrorism and such. All of these are corrosive to democracy, and it's pretty obvious to me that toxicity has reached very unhealthy and ultimately unsustainable levels.

[Edit: Just to be clear, I didn't say that Western Democracy will collapse in this decade; I said that its legitimacy will collapse (arguably is collapsing right now), which is something else. This has both internal repercussions in the nature of the political discourse/conflict, and external repercussions in that Western-style Democracy will less and less be what the rest of world aspires to. Also to be clear: I'm a big fan of Western Democracy -- more than any other form of government around -- I just think that in practice, they've been buggering things up pretty badly for the last decade or two, and consequently are going to have an extremely challenging decade relative to much of the rest of the world.]