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

Key points

• **'AI as Girardian scapegoat for layoffs' — direct counter to the AppLovin/Coinbase narrative.** Lütke: 'What you see right now is not AI layoffs. Those are companies that are really slow that overhired during Covid.' AI is 'the perfect scapegoat' because 'it can't fight back.' But — and this is the crucial twist — his own operating model produces the same outcome via different reasoning: **Shopify aims to stay at 7,500-8,000 people for 5 years at 100x productivity per head.** So the *narrative* is different but the *destination* matches AppLovin / Kalshi / Baseten. Adds nuance to the Issue 03 lean-ops thesis.
• **>50% of Shopify code is AI-generated** and 'converting to much higher numbers.' Many of their best engineers haven't written code this year. **'December changed everything. Opus changed everything.'** Direct timeline marker for when the operator class internally flipped on agent-led coding. Cross-references AppLovin's 80-90% from Issue 03.
• **River — Shopify's agentic engineering system that lives in Slack.** Named itself ('we built her, then asked what she wanted to be called'). Engineers steer it via public Slack channels so the whole org learns from each other. Does 'a ludicrous amount' of Shopify engineering. This is the most concrete production-deployed agentic-engineering example we've seen — a step beyond Cursor/Claude Code as discrete IDE tools.
• **Toby changed his mind on a major prior: senior engineers > junior engineers for AI coding.** Initial thesis: 'no priors' meant AI-native juniors would have the advantage. Reality: 'all engineers are massively underestimating how important the steering is.' Years of reps + experience = better steering = better AI output. **This is the direct contradiction of the early-2025 'AI flattens the experience curve' narrative** and a meaningful operator-side data point.
• **Context engineering as a new career role** that 'will become recognisable and subsume more specialised roles.' Engineer managers are the best AI programmers because they've prompted intelligent agents (humans) for years. **Cross-reference: Issue 02's PM-debate resolution** — this is the new 'chef-mode' role description from a different angle. Aligns with last week's Schoening/Notion preview thread on 'agency over titles.'
• **Token-intensity reality at Shopify scale.** Toby uses **Ghosty (terminal) + Claude Code** himself. 'I have 15 Ghosty terminals on my desktop right now.' He describes himself as a CEO who actively codes via agents while in meetings. The pattern Lemkin described last issue ($254/mo for an autonomous AI VP of Marketing) is consistent with what Shopify is doing internally — most operator tasks below the data-structure layer are now agent-handled.
• **'You can just do things' — Toby's best-advice frame.** 'At any point of time the system exists to try to get to good outcomes. If you know what good outcomes are, you can step out of the system and give it a go. The cost is lower than ever, and action causes information.' Direct alignment with the lean-ops + Coinbase 'no managers of managers' theme this week.
• **Anti-charity framing — most provocative position.** 'Giving money is not virtuous unless it causes the right things even if it sounds good.' 'The world has too much charity dollars. The problem with charity dollars is they can't be given to anything that has a self-healing fitness function — i.e., for-profit / markets.' Praises Bezos's ex-wife's chastised giving as 'the good part.' Carnegie libraries as the rare counter-example.
• **The Enneagram-8 founder framing.** Toby self-identifies as Type 8 (challenger). 'Eights are conspired against in companies — they say things are bullshit when they're bullshit. They're dangerous to everyone else's careers, usually don't get promoted, usually leave, often start companies.' Says Apple bringing Steve back created the boundary condition for boards to give founders longer runs — which proved 'really good for companies because now you have diversity with executives.' Useful frame for founder-led outperformance.
• **The IPO pricing anecdote.** 2015 Shopify IPO at ~$1.67B. Bankers pushed for conservative pricing 'so allocations go to long-term holders.' Toby raised price above the range. First trade was ~$10 higher than the offering price = effectively gave away ~10% of the raise. 'The market told me what it really was.' The banker who ran the IPO is now Shopify's CFO. Sub-text: bankers' incentives misalign with founders' even when relationships are good — verify the comp structure before the road show.
• **'Real democracy is capital allocation.'** Toby's most-pro-markets quote. Every dollar spent is a vote for that product, that supply chain, that company. 'Capitalism causes markets, and markets cause prosperity because capital allocation is distributed.' Direct philosophical opposition to bucket-2 'system of record' SaaS framing — this is the underpinning of bucket-3 'agent-using' growth-reacceleration thesis from Issue 03.
• **Tesla / Elon paean.** Drives Model Y, Starlink in the back for road trips, 'didn't touch the steering wheel of my car this morning.' 'He captured some small percentage of the value he created and used it to create more.' Cross-references this week's All-In on Elon's Colossus 1 → Anthropic deal — the 'no one is better than Elon at converting electrons to tokens' framing has a Canadian-CEO endorsement.
• **Friedrich List / Prussian economics prescription for Europe.** Government's role: 'define games that have societal-thriving as positive externalities, then completely get out of the game.' Anti-climate-cult ('we can't build because some frog breeds in a creek on the perimeter'). UK leaving EU = opportunity for the EU to double down on internal-market discipline. Useful framing for the Hilton/California episode from Issue 03 — same anti-permitting-paralysis logic.
• **Chinese open-source AI in kids' hands.** 'Kids will not stop using AI. What they will do is download Chinese models, and you will never get another high-school essay about Tiananmen Square as a consequence.' Direct argument that AI-age-restrictions will produce a monoculture of collectivist RLHF outputs across a generation of users.
• **Carney critique.** Toby: 'Greatest speech I ever heard, but not a credible witness to the reality on the ground. Canada's only winning strategy historically is win by helping America win.' Strong critique of the Trump-derangement that has 60%+ of Canadians naming the US as the largest risk to Canada. Build pipelines, refine resources domestically, prioritise the highly-educated workforce.
• **Counter to wealth-demonisation.** 'You can't get to a billion dollars by stealing. The people who get their wealth by building companies have not stolen anything — they have created a product that people voted for. It's the most democratic thing that exists.' Direct frame against the Mamdani-style political narrative and the Ken Griffin / Sam Altman house-targeting story All-In flagged this week.
• **On living as a public-company CEO.** Public companies should chase 'trusted public company' status (best) rather than be stuck as 'untrusted public company' (worst). 'It's much worse to be an untrusted public company than an untrusted private company.' Reframes why founders avoid IPOs — but argues they should still do it because trusted-public is the real terminal node.

Notable quotes

Themes

• Counter-narrative: AI as Girardian scapegoat for COVID-era overhiring
• Same lean-ops destination via different reasoning — Shopify flat at 8k people, 100x productivity
• Senior engineers + steering ability beat juniors + no-priors for AI coding
• Context engineering as the new dominant role description
• Markets as real democracy: anti-charity, anti-government-overreach, pro-builder framing