Malcolm Peralty

Worker dignity under capitalism needs rules that can survive impatient investors, hungry competitors, and outright business failure.

I’ve been working in and around tech companies for over twenty years. I’ve watched cultures go sour after acquisitions, survived a round of layoffs while good friends packed boxes, and once asked an employer that was trying to keep me for an employment guarantee or real severance. The answer was a polite version of “we don’t do that.”

So when a former tech employee’s essay made the rounds recently, comparing corporate employment to slavery and concluding “The machine matters, the human does not,” I understood why it resonated even though the comparison doesn’t hold up. Ordinary employment, even exploitative employment, is not forced labour. But rejecting the comparison leaves the real question standing: what protection can workers actually expect inside a system built on ownership, competition, and investment returns?

(Quick disclaimer: I’m not an economist or a labour lawyer. This is me thinking out loud based on what I’ve lived, not what I’ve studied.)

The tempting answers are employee ownership, enlightened founders, or “stakeholder capitalism.” All three can help, and none is complete.

Adam Smith saw the core problem back in 1776: “It is not from the benevolence of the butcher, the brewer, or the baker that we expect our dinner, but from their regard to their own interest.” If your dinner shouldn’t depend on the butcher’s kindness, your livelihood shouldn’t depend on your founder’s.

The person whose name is on the website may not be in charge

Once a company takes venture capital, it joins a machine with a clock on it. Venture funds typically live about ten years, and the fund needs to sell its winners before time runs out. The math only works when one or two enormous wins cover a pile of failures.

Charlie Munger put it best: “Show me the incentive and I will show you the outcome.” The incentives here produce a familiar playbook:

• Grow much faster than a normal business would
• Spend heavily while the money is available
• Chase a sale or public offering big enough to pay back the whole fund
• Quietly walk away from anything that can’t get there

None of this makes venture capital evil. It funds projects no bank would touch. But a founder who has taken a few rounds of that money no longer has unlimited freedom to prioritize steady employment or the original culture. Counting on a founder’s kindness is weak protection, because you’re asking one person to hold a door open inside a building designed to close it.

Being the nice company is expensive

Imagine a company that keeps extra staff so workloads stay sane, avoids layoffs, shares profits, and skips the surveillance software. Some of that pays for itself through loyalty and trust. But if competitors ship a similar product with fewer people and harsher workloads, the nice company ends up with thinner margins and nervous investors right when a downturn hits.

How much generosity a company can afford depends on how much pricing power it has. OECD researchers have found that wage gains can cost jobs in cutthroat markets, but far less so where companies can set their own prices. In plain terms: there’s a lot more room to be generous at Microsoft than at a ten-person agency bidding against dozens of nearly identical competitors.

Markets don’t reliably reward virtue. They reward whatever wins under the current rules, which is exactly why the rules matter.

Automation lets companies diversify their labour, too

I use AI tools regularly and I’m equally unimpressed by the hype and the dismissal, so treat this as a competition argument rather than a doom prediction, and one that holds whether AI turns out cheaper or more expensive than a human on a per hour basis.

A human employee is a commitment: months to hire, longer to get up to speed, and notice, severance, and morale damage on the way out. An AI system is more like a utility bill. Use more during a big launch, use less the week after, and nobody has to run a painful all-hands about it. When the tools improve, every copy improves at once, while human skill grows one person at a time and can quit, retire, or get poached.

Put another way: automation lets a company do to its labour what investors already do with their money. Split it up, scale it up or down, move it around. The company’s dependence on any one worker shrinks toward zero, while the worker’s dependence on the company doesn’t move an inch. Their side of the ledger gets more diversified. Your side stays exactly as concentrated as it was, even at companies run by people who feel bad about it.

That’s why the arguing over price per hour mostly misses the point. Even if AI costs more per hour of useful output, the company is buying flexibility, and businesses have always paid extra for the option to change their mind. The cost of running these tools could keep falling, or today’s prices could turn out to be propped up by investor money and get expensive later. The flexibility advantage survives either way.

Where does this all land for jobs? I honestly don’t know, and I side-eye anyone who claims to. What I’m confident about is where the transition costs land while we find out: on specific people, in specific roles, with the least savings to absorb the hit.

Okay, so what would actually work?

Companies need to stay free to adapt and fail. People need to stay secure enough to survive it. That takes a few layers that don’t exist in most capitalist markets.

1. Protect the worker, not the specific job

Making firing extremely difficult protects the people already inside while slowing hiring and pushing work toward contractors. The better trade pairs flexible hiring and firing with a real cushion: notice, severance, benefits that follow you between jobs, strong unemployment insurance, and funded retraining. Denmark’s “flexicurity” model is the usual real-world example. It isn’t perfect, but the principle travels: protect the worker’s ability to survive change instead of promising the employer will never change.

2. Set the rules for whole industries, not one company at a time

One employer raising wages alone just raises its own costs. The floors have to apply to everyone competing: minimum conditions, industry wage floors, benefits that move with the worker, severance rules, and limits on workplace surveillance and being managed by software. OECD research suggests this works best when the industry sets the framework and each company keeps some room to adapt. The floor exists so nobody wins purely by shoving risk onto workers.

3. Share real profit, not imaginary future value

France requires companies with 50 or more employees to hand a legally defined share of profits to workers, and since 2025, certain consistently profitable smaller companies must set up some form of profit-sharing too. No profit, nothing owed. Real profit, workers get a set share, and every competitor faces the same rule. It’s also one of the few ways workers see any of the gains from automation, which will show up as profit long before they show up as anyone’s raise. Yes, the formulas can be gamed, which is why auditing and anti-cheating rules matter.

4. Treat a company sale as a people event, not just a payday

For investors, a sale is a payday. For employees, it can turn life upside down. A sale should automatically trigger a few things:

• Real notice ahead of time, and stock you’ve already earned becomes yours right away
• Better severance for people let go because of the sale
• Extra pay for the people asked to stay and handle the handover
• Retraining money for people whose jobs go away

No employee veto over the sale. But the price a buyer pays reflects systems and knowledge the workforce built, so the deal should acknowledge the human costs it creates.

5. Give workers more voice as the company grows

A five-person startup needs speed, and layers of committees would kill it. Voice should grow with the company:

• Early stage: founder control, transparent pay, honest explanations of what the stock is really worth
• Growth stage: an elected employee council, consultation before big layoffs, enough financial openness to understand major decisions
• Mature: worker seats on the board, formal worker councils, union bargaining

A major review of research on worker seats on boards found mostly neutral or mildly positive effects. Voice is plumbing. It improves information and trust, and it can’t make an uncompetitive business viable.

6. Stop treating venture capital as the default

VC fits companies that must spend heavily before earning anything, with a huge potential payoff at the end. Plenty of businesses are better served by growing on their own sales, regular bank loans, selling to their employees over time, or investors who are happy to wait. The funding source shapes the company. My quarrel is with letting VC’s grow-fast-and-sell model become the default picture of every worthwhile company.

None of this is free

Higher standards could mean fewer jobs at companies that barely scrape by. Strong severance makes companies cautious about hiring. Industry-wide rules can favour big established companies over new ones. Retraining costs real public money. No design maximizes wages, security, investor returns, and growth all at once. The honest question is who carries which risk.

Right now, capital is diversified: portfolios, index funds, executives who negotiate their own exit packages. Workers are concentrated: one employer often controls the salary, the insurance, the schedule, and next month’s mortgage payment. Labour rules exist because those two groups have wildly different abilities to absorb a loss.

So the fair deal is openly lopsided. Founders can get rich, investors can insist on returns, jobs can disappear, and bad companies can fail. In exchange, workers get clear contracts, minimum protections that apply to every competitor, a set share of real profits, a bigger voice as companies grow, and enough of a safety net that one lost job isn’t a catastrophe.

A kind founder is a lucky break, not a plan

A kind founder can change. A company can be sold. A fund expires, a big customer leaves, a new AI release automates a chunk of what your team did last year, the market turns. Any theory of worker dignity that depends on one person’s permanent generosity is fragile by design.

Keep capitalism’s capacity for risk, reward, and failure, because that’s where the adaptation comes from. Then build the rules that decide how the gains and losses get shared. Economists call the churn of old companies dying and new ones rising “creative destruction,” and I’m fine with it continuing to exist. I’d just like us to stop handing most of the destruction to workers while saving most of the creation for capital.

I’ve surely gotten pieces of this wrong, and I’d genuinely like to hear where. If you’ve lived inside one of these models (Danish readers, French readers, co-op folks, I’m looking at you), tell me what the brochure leaves out.

AI Use: I used AI to help polish (re-write) this article and create the featured image, but the opinions are my own.