Shopify is a platform for online storefronts. Shopify’s very popular, so it has tons of money to spend on stupid ideas.
Shopify CEO Tobias Lütke fell hard for the chatbot. He uses it for everything. Last year, Lütke sent Shopify staff a memo: “AI usage is now a baseline expectation”. [Twitter, archive]
Everyone at every level of Shopify must use AI. Your AI use will be in your performance review. You can’t hire new staff unless you try an AI agent in the position first.
How did this go? Lütke did a podcast last week about how great AI was at Shopify. [YouTube]
Lütke still thinks the chatbot is the best thing ever. But he’s set up the incentives such that the staff use the slop machine to make slop:
Interviewer: Has AI made anything worse internally?
Lütke: Yes. So the concept of responsibility is, it’s easy to skip past, right? It’s one thing that’s definitely worse is, the failure case now of lazy work is not lack of output, it’s actually over-output now. Internally, we have come to call these things that people are lobbing slop grenades at each other, which I think is a really fun term that we should push into industry because it’s fun to say.
The AI is awesome, the problem is the humans in the loop. They’re prompting it wrong.
The press, starting with Fortune, picked up on this one quote and headlined it as if Lütke was walking back the Shopify AI push. [Fortune, archive; Futurism]
But if you listen to the podcast — which I don’t advise, it’s a neutron bomb of inanity — he’s not walking back a thing. This is not a redemption story, it’s a doubling down story.
Lütke thinks some of the results are bad — but the causes of those results? They’re very good.
Now, you might think handling money requires precision and certainty, two things AI does neither of. But Lütke still thinks AI’s the best and nobody at Shopify codes any more:
Lütke: Inside of Shopify, the amount of people I know who really write code is vanishingly small now. It still exists at the limits of complexity for sure, and obviously in the reviews and so on, and then state management of all things, it seems to remain to be the thing that’s really the hardest to get right, which people do by hand, and then sort of, uh, vibe the rest around it.
How cooked is Lütke? He’s convinced the following is normal and reasonable:
Lütke: I have a an AI chief of staff, which I think is pretty common amongst the techie nerds at this point, OpenClaw-like systems that just have all my notes and all my access to a lot of company systems, and just can go, and I can send text messages to, and they’ll go and research something. Very often what I require is like, hey, I need five different positions on something from different backgrounds. Then my agent will orchestrate subagents that are tasked to play different roles, look at the same thing, come back, synthesize, and then send me that. I usually have them send to to me as an audio message and queue it up, and then in the morning in the gym I can listen to the entire stack of things that I wanted to get through.
The AI cannot fail, it can only be failed. With luck, Lütke will never have to speak to a human minion ever again.




