You’re right that owners will automate whenver it raises margins, and that 'perfect employee' machines don’t buy groceries. That's the demand side problem economists have flagged for a century. Youre also right that public companies optimize for growth and quarterly optics in ways a family hardware store does not.
However, automation has been eating jobs for 200 years. Weavers, elevator operators, bank tellers, travel agents. Employment and living standards still rose because new work appeared (software, healthcare, logistics, content, maintenance of the machines). AI may be faster and broader. That is a real risk, not proof the pattern must break.
Replacement is rarely total. Most deployments so far are partial. We have draft, search, code assist, scheduling. Headcount often stays while output rises, or shifts to oversight, exception handling, sales, and care work machines still do poorly. Fully automated restaurants exist as prototypes, most fail on taste, edge cases, and customers who still want a human.
UBI is one option. Others are shorter workweeks, wage subsidies, robot taxes, broader equity ownership, and antitrust that stops a handful of firms from capturing the whole surplus. Tax incidence can change as it has before. The ultra wealthy fighting taxes is politics, not a law of physics.
Infinite growth on a finite planet is a genuine tension. So is the fact that most people still want more energy, medicine, and housing, etc... which historically came from productivity growth. Local firms without shareholders are a useful counter-model as they don’t scale the same way and don’t fund most R&D.
On the last point: LLMs are not minds and they are trained on other people’s work. Calling them 'plagiarism software' is half true as rhetoric and incomplete as description. They compress patterns and generate novel combinations. It doesn’t make the ownership and labor questions disappear.
The contradiction is real. The collapse story is not automatic. Policy and who captures the gains decide which way it goes.
We're actually seeing a major increase in small businesses being created since ChatGPT (2022 or so) helped pave the way for AI.
