We end where we began — with the future of work.

Driving the news: In 2013, two young Oxford University researchers ignited a provocative debate with a landmark forecast that 47% of U.S. jobs are vulnerable to automation. Since then, experts around the world have relentlessly argued whether the new age of robots will wipe out whole classes of jobs, or create a unique time of machine-human partnership.

Though few appear to have taken notice, leading researchers have reached important conclusions, including that:

• Robots so far are destroying many more jobs than they are creating.
• But it doesn't have to be that way — if humans realize they are in charge, and take control of what the machines are generally aiming to do.
• One way to understand is to look at the recent debate over globalization: Rather than scorched-earth trade, open borders could have been optimized to preserve the living standards of affected U.S. communities. Similarly, humans can decide what role machines will play in society.

"It is a matter of choice how much we get to invest in replacing workers and how much in complementing workers," said Daron Acemoglu, an influential scholar on automation at MIT.

• "Some people say robots will destroy more jobs, others that it will create more, but the agency we have is de-emphasized," he tells Axios. "When you look at the recent and more distant past, you see how important the choices are that firms, governments and people make."

The big picture: In a way, we are in a heyday for employment. In the U.S., the jobless rate has been near a half-century low at or below 4% for more than a year. Almost anyone who wants a job — including the stubbornly unemployed, felony convicts and workers with a drug record — can get one.

Wages, too, have been climbing beyond the rate of inflation: Since 2015, workers have carved out real pay increases every month on average, a change from the prior decade, when they routinely lost money to inflation, according to government data.

But economists say they are surprised that wages are not rising more robustly, given how tight government figures say the job market is. That reveals other conclusive facts about current and perhaps future jobs:

• We may not be in a tight job market at all. Instead, the market is tight in some places, but with considerable flab elsewhere.
• This includes lots of people who are working part-time or in gig jobs when they want full-time work, and others underemployed in full-time jobs, according to a 2018 paper by the University of Stirling's David Bell and Dartmouth's David Blanchflower.

If you remember one thing: There is no "future of work." Instead, we are in the midst of a gigantic and painful economic transition that began four decades ago and could go on for years and decades longer, as we have reported.

What's next:

• Work of all types is at risk — low- and higher-paying, from drivers to cashiers, investment advisers and paralegals, says Al Fitzpayne, executive director of the Future of Work Initiative at the Aspen Institute.
• If you want to tell your kids anything, advise them to get some digital training: From 2002 to 2016, the number of U.S. jobs requiring high levels of digital skills more than quadrupled, from 5% to 23% of total employment. And the number of jobs requiring little digital skills fell from 56% to less than 30%, according to the Brookings Institution.

The bottom line: One of the most wrenching experiences of the current age has been falling out of the middle class. It's what underlies much of the disaffection and anger across the U.S. and Europe.

• While humans are taking charge of robots, they have to make an affirmative decision whether public policy should more or less require a middle-class wage, suggests Benjamin Pring, head of Cognizant's Center for the Future of Work.

Another week or so of summer remains. Here are a couple of new titles that are in my backpack:

The Economists' Hour: Growing up watching Paul Volcker and Alan Greenspan, I assumed that the pronouncements of outsized Fed economists always had the ear of the markets and the President. Not so, writes the NYT's Binyamin Appelbaum in this highly readable biography of big economic personalities.

When Volcker was coming up through the Fed ranks, "the central bank's leadership included bankers, lawyers, and an Iowa hog farmer, but not an single economist." The age we now regard as a fixture began only in 1970, with Richard Nixon's appointment of Arthur Burns, a Columbia University economics professor and expert on business cycles, to chair the Fed.

End Times: There is anxiety coming out of the wazoo at the moment, what with the trade wars, upheaval in Hong Kong, topsy-turvy stock and bond markets, the outburst-of-the-moment from the White House, Brexit and more. But Bryan Walsh, a former senior editor at Time, writes that we ain't seen nothing yet. What happens if a mega-colossal volcano erupts, and lays waste to human civilization? Or if the Earth collides with a big asteroid?

In a way Walsh is playing provocateur: The existential risks that he plumbs are not high probability. But the various ways of possible human extinction are not invention, either — the volcano Toba really did almost wipe out humans 74,000 years ago; and the Chicxulub asteroid killed all the terrestrial dinosaurs on the planet 66 million years ago.

This is a playbook for how to navigate the most anxiety-inducing dangers of all.

You have been buried in work? Never mind — here is the top of this week's Future:

1. The new threat to capitalism: Demographics are another threat to the system

2. Truth and falsehood: Social media contemplates arbitrating the difference

3. Russian interference, 2020: Americans are still sitting ducks

4. Reality check for AI hubris: AI is still in its infancy, and here is why

The creation of randomness (Amanda Shendruk - Quartz)

Why self-driving carmakers are coming clean (Joann Muller - Axios)

The painstakingly human task of training AI (Cade Metz - NYT)

The hottest food fad: Peas, soy, mung beans (Deena Shanker - Bloomberg)

The future of the toilet (Angela Lashbrook - Medium)

Cockroaches don't look like the most elegant of runners as they scurry out from under the refrigerator, but they could be a model for how to make robots move really fast without tripping or falling over, according to a new paper.

• In their research, Izaak Neveln, Amoolya Tirumalai and Simon Sponberg studied 17 cockroaches and 2,982 of their strides to come up with mathematical equations and principles for how they manage to scurry as they do.
• The objective was to accomplish one of the most difficult feats in ambulatory robots — perfect how they walk, the school said in a press release.