Shopify CEO on How AI is a Scapegoat for Mass Layoffs & Trump Derangement Syndrome in Canada

Why does it matter? Because the CEO of a $160B company just called BS on AI layoffs, Canadian foreign policy, and the charity industrial complex — all in one conversation.

Tobi Lütke doesn't do talking points. In a wide-ranging conversation, the Shopify founder dismantles three narratives the media has largely left untouched: that AI is causing mass layoffs, that experienced engineers are becoming obsolete, and that Canada's confrontational stance toward the US is principled rather than delusional. What comes through isn't pessimism — it's the frustration of someone watching the wrong diagnoses lead to the wrong cures.

• Companies blaming AI for layoffs are hiding pandemic-era overhiring behind a scapegoat that can't fight back.
• Shopify's best engineers haven't written a line of code since December 2024 — and that's a feature, not a crisis.
• Senior engineers outperform AI-native juniors because steering AI requires hard-won pattern recognition, not just fluency with tools.
• 60%+ of Canadians believe the US is a greater threat than Russia or China — and Lütke thinks that's a catastrophic strategic miscalculation dressed up as patriotism.

'AI layoffs' is a fraud — pandemic overhiring is getting a technological alibi

What you see right now is not AI layoffs. Lütke is blunt about it: "Those are just like the companies that are really slow that like overhire just like everyone else." The companies shedding workers in 2024 and 2025 are the same ones that bloated their headcounts during the zero-interest-rate frenzy of 2020–2022. AI didn't cause the problem. AI is just the story being told about it afterward.

The mechanism is almost too convenient. AI carries no PR risk, can't sue for defamation, and arrives pre-loaded with public fear that makes the explanation feel plausible. "AI is going to be blamed for absolutely everything," Lütke says. "It's the perfect Girardian scapegoat." The tech industry, he argues, has spent years gaslighting the public about AI's dangers — and now those same fears are being weaponized by slow-moving companies to avoid accountability for their own hiring mistakes.

The practical consequence is real: misattributing these layoffs to AI distorts public understanding of what's actually happening in labor markets, inflates fear about automation, and lets executives off the hook for decisions that had nothing to do with technology. When a company says "AI made us do it," the right response is to demand specifics — which teams, which functions, which decisions during which quarters made this inevitable. The answer almost always points back to 2021.

Shopify's best engineers stopped writing code in December — and Lütke loves it

The number is already striking: over 50% of Shopify's code is now AI-generated, and Lütke says it's "converting to much higher numbers." But the real signal isn't the percentage — it's who stopped writing.

"Many of our best engineers have not written code this year. Like December changed everything. Opus changed everything."

Shopify built an internal AI system called River, which lives in Slack and handles enormous swaths of the company's engineering. The name is intentional: Shopify has a massive monorepo called World, and rivers shape worlds. River named herself — the team built her, then asked what she wanted to be called. Engineers now steer River through public Slack channels, which has an additional effect: everyone learns from everyone else's prompts in real time.

Lütke's own workflow has adapted accordingly. He still writes code — specifically the data structure and data access layer, which he believes North American computer science culture systematically undervalues — and then hands everything else to AI. The craft isn't gone. His view is the opposite of elegy: "Growth is not subtraction of things or replacement. It is just adding." Mechanical watches sell more than ever after quartz. The question isn't whether coding survives — it's whether engineers understand what their job actually is now.

Junior engineers with AI fluency are losing to senior engineers with pattern recognition

Lütke expected the opposite. His prior — reasonable at first glance — was that engineers coming straight out of school with no habits to unlearn would have a structural advantage in an AI-native world. Clean slate, no legacy thinking. He changed his mind.

"Initially I thought just having no prior would be better because we had to reinvent everything. Turns out the way I think idiomatic programming actually works... senior engineers just steer these things in such a way that they can accomplish incredible feats in very very little time."

The reason is what he calls steering — the continuous, high-frequency process of pushing AI in the right direction, catching its drift, redirecting it, asking the precise question that unlocks the next step. That skill isn't about knowing the syntax of a prompt. It's about having seen enough systems fail to know which assumptions to challenge before the AI confidently builds on top of a bad one. "All engineers are massively underestimating how important the steering is."

His observation about engineering managers is equally sharp: people who have managed engineering teams are already excellent AI programmers, because they've spent years prompting intelligent agents — humans — to do complex work. The translation is more natural than it looks. What's changing isn't the value of experience. It's where that experience gets applied.

Canada has Trump Derangement Syndrome so severe it's sabotaging its own economy

Sixty percent of Canadians — and Lütke says the number has climbed since — believe the United States is the largest risk to Canada. Larger than Russia. Larger than China. Lütke's reaction is unambiguous: "Of course it's wrong."

His diagnosis isn't about Trump specifically. It's structural. Canada is, in his framing, a small economy physically attached to the largest economic engine on earth. "There's been one winning strategy in Canada's history, which is win by helping America win." That's not subservience — it's geography and pragmatism colliding with a century of trade data.

He's not opposed to Mark Carney's instinct to diversify: "Let's get closer to Europe. Perfect. Asia, let's go. Bang idea." The problem is choosing antagonism as the vehicle. The emotionally satisfying stance — standing up to Trump, building national pride around resistance — is in direct conflict with the economically rational one. Canada has all the resources the world needs for the next 20 years. Pipelines, mining, refining. "The obvious way for prosperity here is to build the shit out of pipelines, build the shit out of our industry, like get resources that everyone needs."

The niceness that defines Canadian culture, which Lütke actually diagnoses as a structural problem rather than a virtue, makes the feedback loop worse. Because Trump is aggressively not-nice, the Canadian reflex is to harden in the opposite direction — and that reflex is now being mistaken for strategy.

'Not-for-profit' is a category everyone should be deeply suspicious of

Giving money is not virtuous unless it causes the right things. That's the whole argument, compressed.

Lütke's critique of charity isn't cynicism — it's epistemological. The market is a fitness function: it tells you whether a product should exist by whether people vote for it with their money. Remove that function and you remove the only self-correcting mechanism that reliably separates value creation from value performance. "Not-for-profit — the world term everyone should be deeply suspicious about because you basically said oh the best fitness function mechanism that planet earth has ever invented... is the thing you're not participating in."

What fills the vacuum? Pull. Smoothness. The people who are good at fundraising, not the people who are good at solving the problem. He's not saying all charities are bad. He's saying the structural incentives are terrible, money gets captured by bad actors systematically, and the Carnegie libraries of the world are a century-old exception being used to justify a rule they don't support.

His alternative isn't abstention — it's demanding a fitness function before giving. What does the organization measure? How does it know if it's working? If it can't answer those questions clearly, the money would do more good deployed into markets.

Great founders are 'eights' — and companies systematically destroy them

Most executives who rise through professionally managed companies share a type: achievers, in Enneagram terms. People who treat every ladder as worth climbing. They are excellent at organizational survival. They are often very bad for organizations.

"Eights are basically conspired against in companies. They are pain in the asses. They just say a thing. They say like this is bullshit." And that bluntness — the direct naming of dysfunction — is dangerous to everyone around them whose career depends on the dysfunction remaining unaddressed. So eights get passed over, exit, or start companies.

Lütke is an eight. Shopify is, by design, unusually dense with them. His point isn't that everyone should tolerate unpleasant truth-tellers as a character-building exercise. It's that the people organizations most aggressively filter out are often the ones with the highest actual signal-to-noise ratio. All the cultural understanding of what good leadership looks like, he argues, comes from movies — which present a curated, dramatically coherent version of something that, in reality, looks considerably messier and more uncomfortable. Read any serious biography of a company builder and the pattern is consistent: these are high-variance people. The variance is the point.

Europe can't build a factory because a frog breeds in the wrong creek — and that's the whole problem

Very few people on earth have the capability to truly build things at scale. Lütke treats this as a factual constraint, not an opinion. And any system that blocks those people destroys disproportionate value — because the bottleneck is so narrow to begin with.

Europe's problem, in his telling, is regulatory accumulation that has made building nearly impossible. "We can't build incredibly important infrastructure, factories, because some fucking frog breeds once in some fucking creek on the parameter." The climate politics layer on top — specifically the founding myths of green parties that treat nuclear power as categorically bad — made the constraint structural rather than merely bureaucratic.

His prescription is Prussian: government's job is to define games with positive externalities, then get completely out of the way and let competition run. Not to manage, optimize, or participate — to set the rules and leave. The builders need to be enabled by the village and held accountable by it. What they don't need is a thousand veto points held by people who will never bear the cost of nothing getting built.

Spending money is more democratic than voting — and most people don't treat it that way

"Every dollar everyone spends is a vote. Real democracy actually happens by capital allocation that's distributed."

This is Lütke's most counterintuitive claim, and he means it precisely. When you buy from a local shop, you're not just buying a product — you're voting for the supply chain that made it, the employment model behind it, and the continued existence of that kind of business. The ethically sourced leather jacket at a slight premium isn't charity. It's a signal that cascades: the supplier becomes viable, production scales, prices fall, the alternative gets outcompeted over time.

The implication he draws from this is pointed: people who believe markets are undemocratic have misidentified where collective decision-making actually happens. Elections happen occasionally, in blunt aggregate, with limited resolution. Purchases happen constantly, at the individual level, with immediate feedback. "The people who get their wealth by building companies have not stolen anything. They have created a product that people voted for. It's actually the most democratic thing that exists."

That framing cuts both ways. It vindicates builders. It also means every consumer choosing Amazon over a local shop is casting a vote — and the outcomes people complain about are, at least partially, the aggregate result of how those votes have been cast.

The real story here is about what happens when the scapegoat breaks down

AI is absorbing blame it doesn't deserve — for layoffs, for job losses, for economic disruption that was already underway. But scapegoats have a shelf life. As AI tools become more legible, as the steering skills of experienced engineers become more visibly decisive, and as the gap between what AI can actually do and what people feared it would do becomes measurable, the alibi will stop working.

What's left when that happens is the harder question: which organizations were actually building, and which were dressing up stagnation as transformation? Lütke's bet is that the builders win that reckoning. You can just do things.

---

Topics: AI and the future of work, software engineering, leadership philosophy, Canadian politics, Trump, economic policy, entrepreneurship, Shopify, charity vs markets, Europe stagnation, code generation