Fears that automation will kill more jobs continue to grow. An estimated 5 million U.S. factory jobs have evaporated since 2000 and most of those (88%) were lost to increased productivity due to automation, according to a study by Ball State University.
But opinions about what, if anything, can be done to reverse the trend differ greatly.
Real estate billionaire Jeff Greene, who hosted his second Managing the Disruption conference on the topic of job destruction and what to do about it in Palm Beach, Fla., this week, has some ideas. Last year, he raised a ruckus by saying that robotics and artificial intelligence would kill not just blue-collar factory jobs but also many white-collar careers. Paralegals, journalists, airline pilots, even surgeons could be impacted, for example.
Greene continued the drumbeat last week in a Washington Post article, warning that automation will kill jobs much faster than Steven Mnuchin, President Donald Trump’s Treasury secretary, expects. Last week Mnuchin said, controversially, that he didn’t think major automation-related job losses would kick in for another 50 to 100 years.
In contrast, a recent report by consultancy PricewaterhouseCoopers estimates that 38% of U.S. jobs have a “high risk” of being wiped out by automation by 2030.
Get Data Sheet, Fortune’s technology newsletter
Greene’s take is that automation isn’t entirely bad if it can be perfected and deployed to lower the cost of living for the middle class so that they wouldn’t need to earn as much money to get by. For example, using 3-D printers to build homes could cut the cost of housing to a more manageable level so that families wouldn’t have to devote most of their income to mortgage payments, Greene told Fortune on Wednesday.
If machines could 3-D print homes using high-density resins, a structure that now costs $200,000 might cost $50,000, Greene said. That would take a big chunk of debt off the table for most families, he said.
Another huge drain on family budgets is energy. Part of that problem could be solved by using alternative energy. Families could save money if the U.S. used solar energy to power electric cars and heat homes. Of course that assumes that the cost of alternatives goes lower than the cost of oil.
The net impact, in his opinion, is that a couple would no longer have to work 80 to 90 hours a week to pay the bills. And if they can live on less money, the need for a two-income household is lessened, allowing one parent to stay home with any children. That, in turn, reduces the risk of children turning to drugs or getting pregnant.
Greene acknowledges that his ideas are “possibly Utopian” in that he’s seeing the use of automation that poses risks to workers across the spectrum in the best possible light.
Many who worry about automation also tout universal basic income as a way to make up for lost jobs. In this scenario, all citizens who are too young for Social Security would receive a flat annual payment from the federal government. It’s a trendy idea in Silicon Valley and other tech enclaves, where it’s seen as a way to hedge against automation-induced job losses.
One small problem: It is unaffordable, according to former Treasury Secretary Lawrence Summers, who spoke at this week’s event. It would cost about $5 trillion annually, or about $4 trillion more than the country’s annual income tax revenue, to pay each American adult $25,000 a year.
“It’s almost impossible to make the arithmetic work,” Summers said, according to The Palm Beach Post.
Others see hope in smaller, more tactical steps, like improving tech education for students and re-training current workers. Last week, General Electric contributed $50 million to the Boston Public Schools to improve science technology engineering and math (STEM) education to train students to help fill what the company sees as a gaping skills gap.
That’s a step in the right direction, but retraining people on a much broader scale is needed to address skills gap nation-wide. There does not seem to be much desire in the Republican-controlled Congress to boost funding on education.
Others say technologies like augmented reality, which layers information onto the real world through connected eye glasses or goggles, could help. For example, field repair technicians could get diagrams and instructions, even video, projected into their goggles so that they can work faster and better without having to stop to consult manuals. That’s technology that Upskill, a Herndon, Va.-based tech company, already provides to customers like GE (GE) and Boeing (BA), which just invested in the company.
“This technology can augment the skills of less specialized workers and help expert workers work faster,” Upskill executive chairman Dr. Magid Abraham told Fortune at a GE event last week.
But back to the conference: Greene says he was struck about how optimistic most of the speakers were—the roster also included former British Prime Minister David Cameron and New York Times columnist Thomas Friedman. “If I were optimistic, I wouldn’t spend all this time and energy on this conference,” he said.
The difference between the industrial revolution of the 19th century and the current situation, he said, is that back then machines replaced physical labor but also created many retail, bookkeeping, machine repair, and accounting jobs related to the goods produced. Now, however, the world is dealing not only with robots that do physical labor but with AI that does mental labor as well.
Says Greene: “We can’t compete with both physical machines and thinking machines.”