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

Summary

• Today’s layoffs are COVID overhiring, not AI, per Shopify’s Tobi Lütke — “AI is going to be blamed for absolutely everything… it’s the perfect Girardian scapegoat. It can’t fight back.” His own plan makes the real call explicit: Shopify sits at ~7,500–8,000 people today, and in five years “my real hope is 7,500 to 8,000 — at 100x productivity level.”
• The agentic inflection is dated to December and Opus: “a fair deal over 50%” of Shopify’s code is AI-generated and “converting to much higher numbers… many of our best engineers have not written code this year. December changed everything. Opus changed everything.” An AI named River lives in Slack and “does some ludicrous amount of Shopify’s engineering,” steered in public channels.
• Lütke publicly reversed a call: he thought no-prior, AI-native grads would dominate; instead senior engineers win because “all engineers are massively underestimating how important the steering is” — steering is programming at a very high level. The role to watch: context engineering (a term he helped popularize), an ascension path he thinks will subsume specialized product/design/engineering roles.
• The bull case for Shopify’s own market: a “golden age of entrepreneurship,” which he calls “by far the most AI-safe job, by far the most AI-benefiting job — think about that combination.” Priors stop mattering when “your handy sidekick AI who acts as your co-founder will just tell you.”
• On wealth and philanthropy he’s deliberately contrarian: “the more wealth an entity has, the more deserving of scrutiny” — and that includes givers. “Giving money is not virtuous unless it causes the right things… we have too much charity dollars.” Non-profits opt out of “the best fitness-function mechanism planet Earth has ever invented.” Stebbings pushed back hard that this scrutiny kills giving; Lütke didn’t concede.
• Governments “are extremely bad at what they do” — anything they take over “will cost 10 times as much.” Europe’s fix starts with “get rid of the climate cult”; the state’s job is to define games with good externalities, then get out — infrastructure and outsourced violence being the exceptions.
• Canada has “Trump derangement syndrome that is just stunning” — 60%+ see the US as a bigger risk than Russia or China, which he calls flatly wrong. One winning strategy in Canadian history is “win by helping America win”; meanwhile “build the fuck out of pipelines,” mines, and refining, because Canada “ought to be the richest country on planet Earth.”
• The China vector: age-gating AI for kids won’t stop usage — “they will download Chinese models and you will never ever get another high school essay about likely Tiananmen Square.” Chinese models carry a collectivist worldview, and “all politics is just collectivism versus individualism.”

Deep dive

1. Founders are “fundamentally crazy people” — and long-term framing is the whole edge

• Stebbings opens with his two-bucket theory (fear of losing vs hunger to win). Lütke’s answer is “unsatisfactory to myself” — he’s mid-figuring-it-out — but his real point: people who fear losing and people who fear winning “are clearly going to look at things much more short term.” A longer perspective changes “a base framing you take to almost every challenge”, because more of the compounding from developing people and handing them the hardest tasks accrues to what you’re building.
• His demystification of leadership: “the world really needs to understand that the people who build companies are fundamentally crazy people… all the aesthetics around what people believe to be good leadership comes from movies.” Read a biography of IBM’s Watson: “you think there’s tyrants in business right now? No. You don’t know anything about how these things went back in the day.”
• “I have this job so other people can have the jobs I wish I could have.” He didn’t want to be CEO — he learned he had to be, because product and company needs are “extremely divergent” until you take a three-year view. Running a product company is a daily marshmallow test, and long-term means “willing to have bad numbers for a while.”

2. Eights get “conspired against” — the post-Jobs founder experiment

• His Enneagram taxonomy (use a pre-2019 test, he jokes, before it “went woke”): 20 years of executives are mostly achievers plus a few deep-diving fives, often the CFO. He’s an eight — the type that says “this is shit, and no amount of fancy dressing changes what it is.” Eights are “dangerous to everyone else’s careers around them,” don’t get promoted, leave — and start companies.
• The experiment running “ever since Apple brought Steve back”: boards giving founders longer runs installs “someone no one can get rid of who just calls bullshit bullshit — and that just turns out to be really, really good for companies.” Shopify is “remarkably high on eights because I really seek them.”
• His leadership physics: he now says “temperature” instead of chaos — “great leaders must be exothermic and must be a heat source for the company… it’s very hard to forge anything new at room temperature.” He spends his time only on what’s broken, and keeps an Obsidian folder of kind words from people he respects — “motivational plumbing” for the weeks of unrelenting bad news.
• On wrongness, unusually candid: “I’ve been wrong more times than most people will ever be wrong… I make falsifiable statements because I want people to falsify them.” His most public mistake — full-on logistics and warehouses just before AI got good. He said the initial decision was right; now: “I probably just actually got it wrong and maybe just admit this. That one sucked because it affected people’s lives.”

3. IPO mechanics: your banker’s customer is the book, not you

• His stack ranking: trusted public company > trusted private > untrusted private > untrusted public. Founders dodge the trust step-down of listing — “which is funny, because it’s actually required to then get to the best spot in the end.” Shopify listed in 2015 at roughly $200M revenue precisely to “make the careers of public market investors who bet on us… and now they trust us.” The Collisons at $150B are “a really bad proxy for what you should be doing, because most people aren’t” them.
• The 80th-floor pricing meeting: the bankers’ customers “are actually the people in the book… their job is to get everyone in the book the cheapest possible deal — beautifully obscured.” Having studied banker comp beforehand, he told them “my fiduciary duty starts T-minus 10.” He priced past the range against the whole room; the first trade still popped ~$10 — “markets are excellent distributed brains and they tell you what it really was.” Out at ~$1.67B, “100 times higher” at points since — and the banker who took him public is his CFO today.
• The ticker: he hasn’t looked “in the last 23 days,” probably far longer. “The ticker is just other people’s game to make trades… I work on the fair market value of a company.” Most wrong: at the beginning of COVID, he thinks the ticker went up “a lot—100% something more” — “we did not get 100% smarter in that moment.” Contrast his gold-mine-CEO anecdote: with an undifferentiated product, “technology is actually investor relations” — charisma nudges the multiple so you can buy the other mines.

4. The layoffs are COVID overhiring — “AI is the perfect Girardian scapegoat”

• Shopify today: about 7,500–8,000 people. In five years: “my real hope is 7,500 to 8,000 — at 100x productivity level.” Staying flat isn’t shrinkage; being “opportunity limited” would be the failure, and Shopify’s mission — make entrepreneurship more common — leaves room to do a lot.
• “What you see right now is not AI layoffs. Those are just the companies that are really slow doing the overhiring correction now. AI is going to be blamed for absolutely everything — it’s the perfect Girardian scapegoat.” Meanwhile, he says, his own industry “has been gaslighting everyone into AI fear, and sci-fi has done it for the 60 years before that.”
• Stebbings’s pushback on the utopian lens — Claude already handles execution-oriented tasks well. Lütke’s reply: “none of the people who are doing jobs that are just tasks for other people have good jobs. Being an automated task queue is not a great job.” Purchasing power “is going to go up like crazy,” products get enormously cheap, and people get options.
• The consequence he’d trade on: a “golden age of entrepreneurship — by far the most AI-safe job, by far the most AI-benefiting job. Think about that combination.” Priors stop mattering — you no longer need a family business background when “your handy sidekick AI who acts as your co-founder will just tell you.”

5. We are incredible at inventing jobs: F1, terminals, gentleman programmers

• Stebbings’s stat, which Lütke calls a brilliant start: of the top 10 best-paid jobs today, eight didn’t exist 20 years ago — “that is going to be true in 10 years too, and 10 years after that.”
• His favorite example: Formula 1 — 800 people building an engine at Mercedes (and at every manufacturer), “living in the perpetual golden age of industrial mechanical engineering,” scorecarded every weekend. “It’s a sport named after a rule book… we just made this up, and all of our lives are richer because it exists.” Stebbings adds the storytelling half: Drive to Survive rescued a decaying sport — rulebook plus spectacle equals value nobody projected.
• The likely Ghostty story: a founder sold his company, then “by hand made a perfectly crafted piece of software he thought needed to exist” — the terminal application Lütke runs 15 instances of, “telling my clankers what needs building at Shopify” while on calls about legal necessities. His larger claim: “the world is very impoverished by how expensive it was to make software.”
• The economics underneath, spelled out because “at least half of the population is utterly confused about how this all works”: “growth doesn’t create wealth; growth brings wealth from resources into the economies” — companies corral infinite hypothetical value into actual products people can vote for with money.

6. Scrutinize wealth by what it causes — and stop misattributing the anger

• On American wealth-hatred: “probably at a high-water mark, but still sufficiently low that it can get into self-healing.” His actual position: “the more wealth and resources an entity or even an individual has, the more deserving of scrutiny they are” — what’s broken is that “the process of figuring out who’s additive and who isn’t is completely broken; it’s owned by people acting in utterly bad faith,” especially state-run and distorting media.
• The Elon riff as exhibit A: he didn’t touch his steering wheel that morning, the Model Y is “incredibly affordable compared to the norms of cars,” Starlink in the back of the car — “he’s the one-man engine” who actualized tremendous value and captured “some very small percentage” of it. “How cool is it that when you go through the list of the richest people on planet Earth, all of them have built something.”
• His answer to the they-must-have-stolen-it narrative: “you can’t get to a billion dollars by stealing”—though he immediately allows “maybe you can—maybe there are some instances.” People who get their wealth by building companies have created a product that people voted for. It’s actually the most democratic thing that exists — vastly more democratic than any of the elections. He owns 6% of Shopify; 94% belongs to others, and roughly 10 million people work day-to-day on Shopify. Inherited or divorce wealth is the different case: “you are custodian of the wealth and you need to use it gainfully for society.”

7. “We have too much charity dollars” — the sharpest clash of the episode

• Stebbings recalls Jeff Bezos’s wife being chastised for giving an extremely large amount away inefficiently — “my god.” Lütke, deadpan: “That’s the good part. This is what I want more of.” His principle: “giving money is not virtuous unless it causes the right things, even if it sounds good. We need to get away from making good-sounding things beyond scrutiny.”
• The mechanism: every dollar spent is a vote — “real democracy actually happens by distributed capital allocation.” Non-profits announce they’re opting out of “the best fitness-function mechanism planet Earth has ever invented — the only thing that’s ever raised anyone out of poverty at scale” — and replace “merit of organization with pull of individuals,” attracting “the smoothesters,” not the builders.
• Stebbings’s pushback, worth keeping: he doesn’t have time to diligence every organization, so this scrutiny “stops all giving, and then good organizations that would have got money don’t get it.” Lütke holds the line: “You show me good organizations — they are getting a lot of money. Trust me, money is being captured by really bad actors in these systems.” He even pre-empts the standard rebuttal: Carnegie’s libraries were “a hundred-something years ago… complete banger” though.
• The intellectual through-line is likely Thomas Sowell, whose regret he quotes: society has “spent the last 50 years replacing things that work with things that sound good.” He stress-tested it on nepotism — searched for its redeeming value, concluded “the answer is none” — yet nepotism “might be better than some of the crazy stuff” like intersectionality-tailored academic job postings, since double-blind merit, the gold standard, is vanishingly rare.

8. Governments cost 10x — define the game, then get out

• He’s a fan of Friedrich List and the Prussian school over Adam Smith and laissez-faire: government’s role is “to define games that have societal thriving as externalities and then completely get out of the game,” letting competition do the rest. Because “governments are extremely bad at what they do… the moment a government takes over anything, it will cost 10 times as much” — announce state grocery stores and “a couple hundred people like me have to go build companies to make up the wealth that simply evaporates in this moment.”
• Two exceptions he grants: outsourcing violence to the state — “one of the most inspiring things humanity has ever done,” since it makes personal property possible, “the foundation of everything else I’m talking about” — and infrastructure, “probably the most profitable thing that has ever existed” in societal value (old airports are “running trillions of dollars of profits”), which even long-term businesses can’t build because the timescales don’t fit.
• As “president of Europe”: “you have to get rid of the climate cult first” — green parties with anti-nuclear founding myths, factories blocked “because some frog breeds once in some creek on the perimeter.” Then define excellent internal-market games — “there’s no speed limit” on growth — and only afterwards debate how to sculpt the economy with the riches. Bonus mechanism: a dollar entering a local community “cycles seven times” through payrolls, which is why local businesses matter.

9. Trump derangement syndrome in Canada — and the Tiananmen-essay warning

• Canada “has all the resources that planet Earth needs for literally everything we’re going to do for the next 20 years… it’s purely a choice that Canada ought to be the richest country on planet Earth. It just has to be willing to build some mines — which somehow is controversial.” Canadians are “massively overfit to niceness,” which “leads to unkind lies” — lying by omission that things are fine.
• On Carney: “one of the greatest speeches I ever heard… I don’t think it’s a full, credible witness to the reality on the ground.” Over 60% of Canadians now rate the US a larger risk than Russia or China — “of course it’s wrong.” There’s been “one winning strategy in Canada’s history: win by helping America win.” He endorses every diversification move (“Europe, perfect; Asia, let’s go”) but “I just really don’t think we need to choose”: “build the fuck out of pipelines,” refine domestically — because since beaver pelts went to London and came back as hats, the Canadian story has been “we get the resources, other people make the money” (name a Canadian end product beyond Canada Goose, Lululemon, Shopify).
• On Stebbings’s point that CCP-funded open-source models quietly power much of the US startup stack: the Chinese threat is “both underestimated and overestimated.” The bigger danger is governments age-gating AI: kids won’t stop using it — “they will download Chinese models and you will never ever get another high school essay about likely Tiananmen Square.” Chinese models, especially when switched to Chinese, carry “a very collectivist worldview” — and “all politics is just collectivism versus individualism; everything else is people smokescreening.”
• His social-media-and-kids answer is an honest non-answer: “by the craziest coincidence my kids are totally uninterested” — no phones, never asked, a house meme that getting a phone costs you your PC. Also three boys, he notes — a data point he flags because he thinks social media’s effect is “very different” for others.

10. “December changed everything. Opus changed everything.”

• The numbers: “a fair deal over 50%” of Shopify’s code is AI-generated, “converting to much higher numbers… many of our best engineers have not written code this year, ever since December. December changed everything. Opus changed everything.”
• River: an AI that “lives in Slack” and “does some ludicrous amount of Shopify’s engineering,” steered in public channels “so everyone learns from everyone else.” The monorepo is called World — “rivers shape world… we first built her, then asked her what name she wants. River named herself. That’s kind of crazy.”
• His stated change of mind: he expected no-prior, AI-native grads to win by reinventing everything. Wrong — “all engineers are massively underestimating how important the steering is… it’s just programming at a very high level,” and seniors with reps “accomplish incredible feats in very, very little time.” (Even so, a 13-year-old Waterloo intern whose mom attends his classes is “very good — just unbelievable.”) The coming role: context engineering, or a “product builder” ascension path coordinating intelligent actors — people with engineering-management experience excel because “they’ve been prompting intelligent agents for much longer than Claude Code existed.”
• The craft survives the way mechanical watches do (“selling more than ever — growth is not replacement, it’s adding”). His personal edge, from the German school of engineering: design how data persists on disk first, then “everything else can be vibe code on top… maybe that changes in the next couple of months.” And the advice he’d rally around: “you can just do things… action causes information” — with the hard constraint that it be victimless; in zero-sum environments don’t run experiments, “figure out how to get a positive sum.”