Ok, let's present the alternative then.
Walmart automates its massive retail operations substantially over the next 10-15 years, using robotics and AI.
Their goal is to match the dollar of sales to employee ratio of, say, beloved Costco. And or try to keep up with Amazon's ability to highly automate its operations (competitive pricing advantages etc).
There went one million jobs, consisting largely of employees with few skills and very limited alternative employment prospects.
Now what? Here's the reality when you remove the one-sided emotion-driven premise in the article: it's better to subsidize those employees to a modest extent, than to a 100% extent (which is the scenario coming for tax payers; the harder / more expensive you make it on Walmart to employ those people, the faster they're going to automate their jobs away).
Also, if you want to reduce the subsidy, raise the minimum wage (or the earned income tax credit, a vastly superior approach; which then begets a further debate on the subsidy), which Walmart has routinely supported. They support it because it harms their weaker competitors, whereas Walmart can deal with it. When you raise the minimum wage, Walmart will either optimize by firing workers / using automation further, or they'll push increased costs onto customers, which gets you right back to the same place in terms of tax payers paying for Walmart's employees (hint: every employed adult in America pays for their employees, they have $485 billion in sales).