After posting the previous article on social media, the comments section was fascinating.

Top comments focused on the distribution issue: "Ford's worker benefits aren't mentioned — just thinking about having everyone take on more responsibility?" "So the front-end dev does sales and marketing too — does the salary go up?"

Many people felt that if efficiency improved, why should the company get all the benefits?

This tension is quite real. It points to a question I didn't expand on in the previous article: how should the efficiency dividend from AI be distributed?

Back to the Ford story. A year after the assembly line launched, on January 5, 1914, Ford announced raising workers' daily wages from $2.34 to $5 — more than doubling them.

The entire business world thought he was crazy. The Wall Street Journal called it injecting "criminal stupidity" into the enterprise. Competitors felt he was disrupting the labor market.

But Ford later said: "This was the best cost-cutting decision I ever made."

Cost-cutting? How is a raise cost-cutting?

01. Who Gets the Efficiency Gains

02. The Math Behind $5

Before Ford's raise, the Highland Park factory's annual worker turnover rate was 370%. What does that mean? They had to hire nearly 50,000 people a year just to maintain a workforce of 14,000.

Assembly line work was monotonous and exhausting — one worker repeating the same motion all day, the line never stopping. At $2-something a day for that kind of work, nobody lasted long.

Recruitment costs, training costs, quality fluctuations, production line stoppages — all money.

After the raise, turnover dropped to 16%. Workers stayed, skill levels improved, the line stabilized. Ford estimated that labor productivity improved by 40% to 70%. The costs saved far exceeded the wages added.

But there's a detail often overlooked: Ford's $5 wasn't unconditional.

He simultaneously established a "Sociological Department" that conducted home inspections of workers. No gambling, no drinking, stable family relationships, maintaining "good living habits" — only then could you get the full $5.

By today's standards, this practice is highly controversial. But Ford's logic was clear: I'm not doing charity. I'm buying a stable, high-output workforce. When your life is stable, you can consistently deliver high-quality work on the line.

Ford's $5 wasn't profit sharing — it was an investment in productivity.

When new technology brings massive efficiency gains, how should the surplus value be distributed?

03. The Default Destination of Efficiency Gains

There's a historical phenomenon called "Engels' Pause."

From 1790 to 1840, during Britain's most intense 50 years of Industrial Revolution, GDP grew 46%. How much did workers' real wages grow? 12%.

Productive efficiency soared, but workers' living standards barely moved. Nearly all the surplus value was captured by capital. This persisted for about 50 to 60 years, until workers organized, labor laws emerged, and social pressure reached a tipping point before distribution began to change.

Oxford economic historian Robert C. Allen studied this period extensively. His conclusion was blunt: the dividends of technological progress never automatically flow to workers.

The default is: whoever has bargaining power takes it.

Without organization, without negotiating power, without alternatives, efficiency gains are simply capital's efficiency gains. Profits rise; workers still get the old wages.

The AI era is replaying this story.

04. Two Extremes

In 2025, fintech company Block drastically cut its customer service team, claiming AI could handle most tickets. After the earnings report, the stock price rose. Wall Street loves this narrative: fewer people doing more work, profits up, shareholders benefit.

That same year, Klarna did something similar. The CEO announced in the earnings report that AI was doing the equivalent of 700 full-time employees' work, saving massive labor costs.

The result? Customer satisfaction plummeted, complaint handling quality dropped dramatically, and Klarna had to rehire.

One "succeeded," one crashed. But the logic was the same: all efficiency gains went to the company, savings became profit or stock price, employees were either laid off or had their workload increased.

This isn't an isolated phenomenon.

The Centre for Economic Policy Research (CEPR) found in 2024 that employees using AI tools worked an average of 3.15 more hours per week. An EY survey that same year showed 64% of employees felt AI led to increased workload.

The time saved by AI didn't become employees' free time. Most of it became more tasks.

This is the AI version of Engels' Pause: technology advanced, but the people doing the work didn't benefit.

05. The Choice Is in Everyone's Own Hands

Distribution is actually a huge topic. Macroeconomics, labor relations, social systems — each layer has its own complexity. As a small entrepreneur, I can hardly cover it all.

But there's one angle I've been thinking about: starting from each individual.

EY's 2024 survey asked employees: AI saved you time — what do you want most? Number one wasn't a raise — it was flexibility, at 31%. Being able to manage their own time and have more autonomy.

Everyone wants different things. Some want money, some want freedom, some want to excel in one direction, some want a stable salary and to clock out on time.

One major change in the AI era is that individual choice has expanded. One person plus AI can do what used to take a small team. Your boundaries have been opened up.

My adjustment is: adding performance-based pay on top of each technical role's existing salary — essentially a company-wide raise, tied to company growth.

This isn't a mature solution, and I'm not sure if new problems will emerge. But directionally, I want to give everyone more room to choose.

Being yourself, living your own life, might be the most important thing in this era.

06. Ford Workers Could Afford Ford Cars

In 1914, at $5 a day, annual income was about $1,200. The Model T had already dropped below $500. Ford's workers became Ford's consumers.

A few days ago, a report called "The 2028 Global Intelligence Crisis" went viral. Citrini Research's thought experiment: AI massively replaces white-collar workers → consumer spending drops → company revenue decreases → companies buy more AI to cut costs → more workers replaced. A downward spiral.

The author said it wasn't a prediction — it was a stress test. But the mechanism it describes makes sense: if all efficiency gains flow to capital and workers' income keeps shrinking, the consumer side will eventually break. You can produce more, but the number of people who can buy hasn't grown.

Ford faced this problem 112 years ago. He chose to share some of the dividends — and the result was a positive cycle.

Will history repeat? Unknown. How this round will play out, nobody has the answer.

But I find this era fascinating. Old rules are loosening, new rules haven't solidified. Full of uncertainty, full of adjustment, and full of opportunity. Every person, every company is figuring out their own answers.

Isn't that pretty cool?