The Curious Companion: Ep. 17 – Will ChatGPT Get Old Navy’d?
|
Curious Reader! Welcome to this week’s Curious Companion newsletter. What you came for is below, and you can CLICK HERE to listen to the episode if you decide you prefer earbuds to eyeballs. Happy reading! In this episode I dig into whether ChatGPT might get “Old Navy’d”, a concept that ties together capitalism, business strategy, and the evolving AI landscape. Using a Scott Galloway story about Gap and Old Navy as the jumping-off point, I explore what happens when a cheaper, “good-enough” alternative comes for the market leader. From Alibaba’s Qwen to DeepSeek’s leaner, lower-cost models, this episode unpacks the global economics of AI efficiency, U.S. protectionism, and what it would actually take for consumers, or companies, to switch away from OpenAI. The InspoThis episode was 100% inspired by a recent episode of Prof G Markets called How China’s AI Efficiency Could Gut the U.S. Economy. It’s brilliant (despite the clickbait title), and I highly recommend giving it a watch. You’ll find the link below along with my own Curious Finds page on ChatGPTCurious.com, where I share the most useful stuff I watch and read about AI. The Old Navy AnalogySo, WTF do I mean by ChatGPT getting Old Navy’d? In the podcast episode, one of the hosts, Scott Galloway, tells a story from his early career, when his firm worked with Gap. Their research showed that a rapidly growing demographic at the time was single mothers. These women, being the gotdamn superheroes that moms are, wanted their kids to be able to dress in a way that made them feel good about themselves, but they didn’t have as much disposable income. So the team positioned Old Navy as 80% of the Gap for 50% of the price, and by doing so got Old Navy to a billion dollars faster than any other retailer in history. This is the same formula we’ve seen from companies like Southwest Airlines (I don’t ever check a bag but they fucking up with that bag policy change) and IKEA. So the question becomes: Will someone do to OpenAI what Old Navy did to Gap? Enter: ChinaGalloway’s argument is that China might be the one to do it. ChatGPT is massive and expensive AF, while China is out here building leaner, cheaper, and more efficient models, most notably, DeepSeek and Alibaba’s Qwen. Yes, the Alibaba, the same company known for online shopping. Alibaba realized commerce required massive cloud infrastructure so they built Alibaba Cloud, and the rest is history. I can’t speak firsthand to how good these models are, but in the episode, Ed (the other host) noted that Qwen cut GPU use by 82%, and DeepSeek’s models cost 10x less per million tokens compared to GPT-5. That’s 80–90% of the quality for 10% of the cost. Hello, Old Navy. Again, I haven’t used either of those models so I can’t speak to how good they are, but I’ve got a friend whose fiancé works in AI, and he says these Chinese models are good but still don’t come close to the top US commercial ones. You and I both know it’s only a matter of time until they catch up. Communism and authoritarian regimes aren’t great for individual rights, but they’re great at getting shit done, as is evidenced by China’s significant EV progress. The Economic Model: Three LinesBefore I share my two pennies on what I think could happen, I want to run through a mental model Scott shared that is a phenomenal way of conceptualizing economics. He credits Bruce Buchanan, a Stern School of Business economist, with teaching him this. Imagine three lines stacked vertically:
Companies make money by doing one of two things:
My summary of this model is that if the goal is to make money, then the focus becomes increasing the distance between the price (middle line) and perceived value (top line) as much as possible without it becoming sketchy. Stretch it too far and the model collapses as perceived value plummets. If someone tried to sell a Lamborghini for $5 you’d immediately assume it was a scam. Clearly OpenAI is focused on pushing that top line higher, leaning into very American principles of “bigger is better”, with bigger models, bigger data centers, and insane amounts of compute. China on the other hand seems to be focusing on lowering that bottom line, the cost line. What Do I Think?First off, this isn’t happening tomorrow. Second, it ain’t gonna be China. Why not? Because the U.S. won’t allow it. I’m not basing this prediction on any sort of economic or political expertise…but rather by simply looking at what already is! The Chinese EV market is poppin’ and there’s a company, BYD, that would do numbers in the US market but they can’t because they’re banned here via tariffs (100% tariffs, not even joking) and data-security restrictions. And let’s call a spade a spade, these restrictions fully exist just to protect US automakers and limit Chinese tech influence. And they say raising corporate taxes discourages competition 🙄. The same logic would apply to AI models. U.S. policy exists to protect domestic automakers and domestic AI companies. Plus, OpenAI is too powerful economically and politically. AI money is propping up (or at least very heavily driving) the U.S. economy. The government’s not about to let that fail. Get ready for the AI bailout of 2026. If an Old Navy-ing does happen, I think it’ll be an “inside job”, meaning a U.S. company building a “90%-as-good” model at a fraction of the cost. China has shown that it’s possible, which feels very Roger Bannister to me. He was the first to run a sub-4 minute mile, and within a year, 23 people had done it as well. It Won’t Be OpenAIFun fact, Gap and Old Navy are both owned by the same parent company, Gap Inc. Old Navy was created to compete with Walmart, but ultimately wound up cannibalizing Gap itself. Consumers learned to expect “basically the same thing” for less. All that to say, I do not think that it will be OpenAI that builds the Old Navy. For any of you smart cookies reading this and thinking that OpenAI’s other models (ChatGPT 4, 4o, o4 mini) might be giving Old Navy vibes (those models use less energy in theory), close but no cigar. Those models are more like Gap Outlet—cheaper because they’re older, not because they’re intentionally designed to serve a new market. True “Old Navy-ing” is intentional design, not a byproduct of obsolescence. What Would Make People Switch?Something worth thinking about is what would actually drive a switch to a different model, namely by the average consumer. Realistically, it would make more sense for companies to switch first because they’re using the models at scale (which costs more), but there’s red tape and kick backs and shit I don’t know about, so I’m thinking about this as it relates to you and me, the average user. What would make us switch? Two things:
Worth noting, and a bit of a soapbox business moment for me, I do think keeping things the same (stagnant perceived value) on a quality product can be very ok for a very long time…as long as the price doesn’t go up. The best example of this…COSTCO! They keep their prices fixed, their margins low, and the value (along with their values) obvious. I’m a customer for life. My TakeDo I think chatGPT will get Old Navy’d? I think it’s possible but it’s not happening any time soon, and I definitely don’t see China as the one to do it. Do I hope it happens? Absolutely! We don’t need these big ass models. We don’t need to be out here chasing AGI. The hopeful and optimistic side of me views the possibility of ChatGPT getting Old Navy’d as a way of futureproofing our access to this technology. If AI were to go away all together, that would also be totally fine, but I highly doubt that’s happening. I do however have concerns about companies like OpenAI having so much control over it, so it’s comforting to know that other options might be possible in the near future. How I Used ChatGPT This WeekEach episode I share how I used ChatGPT that week. This week, it helped me make this episode. One of my favorite ways to use ChatGPT is to help me deepen my understanding of things. After listening to the Prof G Markets episode earlier in the week, I knew that I wanted it to be the foundation for this episode, but wanted to make sure that I understood it well (an econ major I was not). I went to YouTube and copied the transcript then pasted it into ChatGPT, and had a discussion with it about the topics presented. From that discussion I made the outline for the episode, and here we are. Yes, you can paste massive amounts of text into ChatGPT, and yes, you can get transcripts of YouTube videos right directly inside of YouTube. Da Wrap-upThis was honestly such a fun episode to make. If you’re picking up what I’m putting down, consider sharing it on the socials or forwarding it to a friend who’d enjoy it. As always, endlessly appreciative for you and your curiosity. Catch you next Thursday. Maestro out. AI Disclaimer: In the spirit of transparency (if only we could get that from these tech companies), this email was generated with a very solid alley-oop from ChatGPT. I write super detailed outlines for every podcast episode (proof here), and then use ChatGPT to turn those into succinct, readable recaps that I lightly edit to produce these Curious Companions. Could I “write” it all by hand? Sure. Do I want to? Absolutely not. So instead, I let the robot do the work, so I can focus on the stuff that I actually enjoy doing and you get the content delivered to your digital doorstep, no AirPods required. High fives all around. Did someone forward you this email? Stay curious. |