Speaker unverified: prediction

Transcript context

…ps, it becomes… you have a much more powerful position as a tail merchant because Momax is a tiny brand, but they can just get recommended as the best product. So that's step one that I think happens. I think a slight modification about that I’ll just make, I don't think the tail part matters. I think merit matters. It's kind of an important point because often when you talk a lot about commerce and products at some point just because of prior discussions, people are like, “Yeah, but aren't you just feeding consumerism?” So this is sort of the negative take and again, consumerism is a thing that is being thrown around, but where does it come from? It comes from people throwing away a lot of things that they probably shouldn't. But they throw away things because they hate the things they have. These things don't work, they fall apart, they don't do the thing, they're cheap plasticy, this kind of thing. The thing that solves consumerism is quality products which you want to keep using. You want to be excited about this and you want this travel adapter to last forever and solve…. you literally never have to Google, “What plug is used in Singapore?” You just know that the thing that you have can deal with everything to everything, so problem solved. And so I think that's an important modification on the point. Yeah, I guess implicit in my view is that there's good stuff in the tail, whereas the head has distributional advantages and so it is more meritocratic to bubble more stuff up from the tail. That is I think the reconciliation of those two world views. So that's kind of step one that happens. I think you're getting to step two, which is it feels like there's a lot of changes potentially that can happen around personalization about agents actually acting for you. And maybe one thing I'm curious about is if you have thoughts on. It just seems like a lot of search and e-commerce is really bad, where it's keyword-based search still and you can't do things like, “Let me know if this becomes available in gray” or, “What other products do you have like this?” And you want some kind of embedding space that thinks about similarity between the products and no one can do this. So, when are we getting better at personalizing and search? I just want really good search of the products that are out there. Totally. I remember obsessing about this now and we should have been— Tell the secret Shopify roadmap. Shopify should have solved this earlier and hasn’t and I am pretty mopey about this. You guys are doing ok. I think we will solve this now. So the thing is really fascinating. I really deeply appreciated as a field, it’s so different. It has its own law. Everything is interesting is my fundamental opinion. And search has been a wonderful area to actually learn a lot more about the history now. And the fundamental thing is there is a generic bias in search and a text is king and frankly the people who worked in search are much more interested, we're fundamentally more interested in searching through papers than through products. a generic bias in search and a text is king and frankly the people who worked in search are much more interested, we're fundamentally more interested in searching through papers than through products. Shouldn’t you do o search across… I mean I find it interesting when Shopify does things broader than the Shopify universe. Because the Shop app can track packages not bought from Shopify merchants, right? Yeah. And so, shouldn’t you do the search thing across the whole web and not just Shopify? I don’t know, should we? I think so, because the thing that feels like the big opportunity to me is clustering was always so hard to do because all the features were developed manually. And so, you had to go say that the dress is red and everything like that. And it feels like, have you ever used LLM as just a recommender engine where I like feeding in, “These are all the books that I've read recently . Give me a narrative nonfiction book that someone who's read all these books might like,” and it does extremely well. But you're kind of using the LLM to access its world model and it feels like now you can have an interesting product world model and access it. But for the people using it, whether a merchant is on Shopify or not is probably an incidental detail they don't care about. Yeah, I totally agree. And you're talking about “how,” but the “what” is the really important thing. I want the LLM to be proactive and just tell me, “Hey, this outfit would look good on you and here's a preview. And t here's the full cost” and you want it. It's the personal shopper outreach kind of component, which I think is just really valuable. And this is not going to be some kind of, you'll get this from everyone because it's really easy to put together. It's hard to get the data and embeddings and the sort of understanding about. We have to figure out… It's funny, the most important thing again in the AI space is also about unquantifiables. It's the vibes, people call them. It seems like Midjourney specifically is one company that managed to actually scale taste because— They're a very distinct aesthetic in the images they produce. And maybe this is because they drove towards a particular aesthetic t they felt they just sort of had ownership over, but it's one that's clearly appreciated. But David is also a deep founder mode type in how he operates the company. Yeah, exactly. And so when something outperforms, it's usually a product of a place that doesn't make decisions by consensus, as we know. So we need to figure out taste in this product sense and we'll figure this out. Someone's going to figure it out. Maybe everyone's going to figure it out, maybe it's already figured out in the neurons and we just don't give enough information for it to reason around taste. And so putting, I think it'll be very, very vague with personal shopper. The thing that I think the aggregators still— to just get back to your point on the aggregation side—there's many reasons why people buy things. Lots of things are input purchases. It's like the things next to the checkout counter in the supermarket. I don't think people are going to use ChatGBT to buy Mars bars or something like candy. ings. Lots of things are input purchases. It's like the things next to the checkout counter in the supermarket. I don't think people are going to use ChatGBT to buy Mars bars or something like candy. They just don't do so online. Right, exactly. But is there an online impulse purchase that is not suited to happening in an AI app? Interesting question. Nothing comes to mind immediately. Yeah, I mean I guess online impulse purchases are things where you're scrolling through Instagram and it's like, all the COVID stuff, it's the sharpest knife ever developed. I wonder if an— Umbrella that changes color in the rain. I wonder if what's a truly impulse purchase? I mean they sort of feel like input purchases, but they tend to be very few people part with money unless they are like, “Holy shit, this is the thing I always wanted.” It's just tapping into demand that's been there for a long time. It's sometimes a little bit the opinion…when you look at log files, you're going to get a very incorrect view of the world. It's very often the longest consideration phases. For years, I wanted a travel adapter that can deal with everything through everything. It goes in a chat from you or in an Instagram ad. Doesn't matter. I'll Insta purchase it at this point. Okay, you were just trying to do that. My other purchase I made recently that I really liked is I was going on a long day hike. And if you read the accident reports, there's often people day hiking who get into trouble because people who are going out on a multi-day hike, they have their tent with them and they're prepared. You know what I mean? And then people go— It's a problem with bicycle helmets, too. Same issue. It's a short journey, to the store that gets people in trouble. So day hikers, they get in trouble because they think they don't need any preparedness and then they end up out after dark and a rescue party and everything like that. And so it was like, it's called bivy, a little sleeping bag for emergency purposes and it's made out of reflective foil. And so again, if you end up stuck, you have an emergency foil sleeping bag just so you don't die of exposure, but it folds up to it's like that size and so you can just stick it in a day pack and now you are somehow more prepared for the elements than you were before. I dunno, do you have a favorite recent purchase? I just really like these… I mean someone going just to ridiculously further and to almost products that just celebrate craftsmanship in a way, even in places where it's just like no one else does. Real example: I take all these vitamins in the morning and I've bought these CVS vitamin pillowcases at some point just for travel. And then I found on Shopify this Japanese perfectly sealed, unbelievably beautiful pillowcases. The entire website is just a love letter to— The most beautiful pillo cases— To precision engineered, tolerances— What's it made of? Metal, which is absolute, it feels so… Whatever. I'm really excited about it. You can put a link in. I forgot the name, sadly. And now I gift those things because they're just beautiful. No, that's cool. How’s Shop Pay doing? We referenced it there.…

Stored transcript either side of the excerpt. The highlighted words are the published quote; the surrounding text is unedited source, never generated.