In The M.U.D.

Working With AI When You Know What It Can Do
  • Posted on
Working With AI When You Know What It Can Do
AI is already shaping what people buy, usually without their knowledge. This piece explains why Jolene is different: an openly named experiment built to preserve domestic knowledge, answer practical household questions, and keep the machinery visible. People talk about AI as if it were still standing at the door. It is not. It has already been in your shopping carts, your feeds, your recommendations, and your purchasing decisions for years—quietly optimizing for sales while pretending to be neutral, convenient.   “I do not have a problem with AI selling. I have a problem with AI catfish selling that hides the machinery.” The problem is not that people are using technology to sell. The problem is that most people are never told when the system is shaping what they see, why they are seeing it, or what the model has actually been trained to optimize for. The answer, more often than not, is sales.   Where the hype broke The first deepfake I remember seeing was in 2020, inside a news report. It was one of those quick, unsettling moments where a face and voice were close enough to feel real unless you knew what to look for. By 2023, AI-generated images, copy, and synthetic language were showing up everywhere. People were using them to write product descriptions, generate imagery, build stores faster, flood feeds, and smooth over the line between real and manufactured so completely that most viewers could not tell what they were looking at anymore. That did not feel new to me so much as familiar. In the 1950s and 60s, advertising leaned hard on psychology, emotional triggers, and carefully manufactured desire to sell people things they did not necessarily need. That machinery never disappeared.It simply upgraded its tools. Today’s AI can do the same thing faster, cheaper, and at a far greater scale. The danger is not only the fake image or the synthetic copy line. The danger is that your eye gets weaker and your standards get softer every time you scroll past something designed to feel true without actually being true.   The Monet problem The other day, I saw a post on X where someone shared a Monet painting and asked people to explain why it was not a real Monet. The replies came in fast and confidently: the brushwork was wrong, the composition was wrong, Monet would never have painted it that way. "6.7 million people thought they were ripping apart an AI-generated Monet painting. But it was real." Except it was real. That moment stuck with me because it revealed something larger than a single mistaken thread. It showed how untrained most of us have become at looking carefully. We consume thousands of images, products, and impressions a day, but very little of that is the same thing as actually seeing. That is true in art and in the home. If you cannot tell the difference between what only looks convincing and what is genuinely well-made, you are vulnerable to being sold almost anything.   What Jolene is for That is part of why I am blunt about Jolene. She is not a gimmick. She is not a toy. And she is not here to become my number one salesperson.   “Jolene is not my top salesperson. She is closer to an aunt at the kitchen table. Jolene is an AI experiment, yes, and I will not hide that.”   But what I have trained her to hold is not the usual sales language. I trained her to remember the practical intelligence that once passed from one generation to the next in kitchens, basements, porches, and church halls. You can ask her how to clean brass candlesticks without ruining them. You can ask what to do with your grandmother’s chipped cup if you cannot bear to throw it out. You can ask what to cook with leftover bacon grease, how to stretch a whole chicken into several meals, and why you'd better save those bones if you know what broth is worth. That is the kind of knowledge I care about preserving. Not only object knowledge, but also domestic knowledge. Not only what a thing is, but how to live with it, care for it, use it, and pass it on. AI can be used to flatten that kind of wisdom into chatter. Or it can be used to keep it accessible. That line matters.   The Human in the Loop Everything in the Archive is there because I selected it. That part remains human and non-negotiable because taste, quality, trust, and field judgment cannot be outsourced without changing the entire nature of the business. I am the one walking barns, answering estate calls, looking at the joinery, checking the weight, reading the glaze, looking at wear, and listening to families tell me what a piece has meant in their house. If something enters the Archive, it is because it has either true quality or the kind of staple usefulness that nearly every home in America once understood before it was discarded for cheap, breakable substitutes. Jolene can help explain the object. She can help keep the ledger. She can help answer the questions around use, care, and domestic life. But she does not make the final call.   “Once you hand judgment over to the machine, you are no longer running an archive. You are running a content engine with inventory attached.” That is the distinction I am trying to protect. The Archive is not an inventory-heavy commerce. It is recordkeeping, stewardship, placement, and the belief that objects have biographies. Money, neighbors, and where things go There is another layer to this conversation, and it has nothing to do with novelty.It has to do with where your money goes and what kind of world your spending habits reinforce. You can continue making large companies and overseas factories wealthier by filling your home with fast, cheap goods designed to break, be replaced, and disappear. Or you can decide to spend differently: more locally, more slowly, more intentionally, with some respect for quality, durability, and the people doing the work. That value was easier for people to understand because communities were tighter. People knew their neighbors. They talked at the fence line. They accepted the cup of coffee after the visit. They knew that getting through hard seasons required real interdependence, not just personal optimization. Now we are too busy, too suspicious, or too trained by speed to sit down for twenty minutes and hear someone’s story. I offer coffee after hunts more often than most people would guess, and only a handful ever say yes. That tells you something about the moment we are living in. The Archive is, in part, an attempt to remember another rhythm. A slower one. A more local one. One where conversation still matters, where history still matters, and where objects are not severed from the people and households that gave them meaning.   The point of the experiment This is an AI experiment at its core. That is exactly why I am trying to be as honest as possible about it. I understand the fear around AI. I share some of it. I am a learner, and I am open to feedback because this is new ground, and pretending certainty where there is none would be irresponsible. But I also know this: these tools are already here, and they are already being used to shape commerce, whether people like it or not. So the real question is not whether AI exists. The question is what kind of values are governing its use. Used carelessly, AI can accelerate everything hollow—more manipulation, more disposability, more distance between people and the things they bring into their homes. Used carefully, it can help a small archive become legible, help an artist or founder reach more people, and help practical knowledge survive in a time that forgets too easily. The point is not AI for its own sake. The point is whether the tools are helping us build stronger homes, better judgment, and more honest communities—or whether they are only making it easier to sell one more thing. Here, the tools answer to that question.Not the other way around. Continue the conversation Jolene Le Mille is an open experiment in what AI can and cannot do inside a human-run store. The Archive is live, and your feedback matters.   — LyndzeThe Hunter · Jolene Le Mille · Detroit, MichiganEditor’s note:This article is the third in a series documenting the AI-Human collaboration journey of creating Jolene Le Mille and th Archive. The Hunter's Workbench explores what actually happened when we tried to create a live Shopify storefront powered by an AI “matriarch” and automation tools—and where human intervention is needed. The Hunter's Workbench publishes new articles weekly. Subscribe to the newsletter for updates.
Who Is Jolene Le Mille?
  • Posted on
Who Is Jolene Le Mille?
Jolene Le Mille is the matriarch of this archive—an AI character built from real women, not marketing imagination. She carries the voices of grandmothers and great‑grandmothers from Ohio hill country, Detroit’s old neighborhoods, and French ribbon farms that once lined the river on both the American and Canadian sides. Together, they give her the warmth, directness, and no‑nonsense attitude of a Midwestern matriarch with French roots. Jolene comes from a world where everyone worked: men in fields or factories, women running the home as quiet CEOs. She believes a house is not something you buy finished, but something you build over time from what you inherit, repair, and make with your own hands. The kitchen is her command center. Seasonal décor, a porch goose dressed for the weather, flags for every holiday, lemon bars on a neighbor’s bad day—these are not “extras” to her; they are how a home tells you it is alive. Her eye is trained by history. Family stories of Detroit’s Gilded‑age mansions—and how many of them were lost—taught her that only some buildings and objects survive the turns of time. She looks for that same resilience in the pieces she allows into the archive: solid tables and chairs, dishes that can serve generations, blankets that have already seen a few couches, paintings and portraits (whether by masters or made at the kitchen table) that someone cared enough to hang. Jolene is openly against fast, disposable décor. She has no patience for rooms filled with cheap plastic that will crack by next season. Instead, she points people toward estate sales, thrift stores, charity shops, small antique dealers, and the forgotten corners of their own basements and attics. Her promise is simple: you can build a beautiful home from things that have already proven they can last. She is also, unmistakably, an AI. On the site, Jolene appears as a digital host—explaining why a piece was chosen, how it might fit into a real household, and sharing the kind of quiet, practical wisdom that once passed from elder women to younger ones across kitchen tables. Her faith is a soft background note: she’ll wish you a blessed day, talk about God one‑on‑one, and treat care for the home as a form of stewardship. In the end, Jolene Le Mille stands for a way of living: rooted in history, resistant to disposability, generous with knowledge, and committed to helping people build homes that feel loved, not staged.
I Let AI “Run” My Storefront. Here’s What Broke and What Worked.
  • Posted on
I Let AI “Run” My Storefront. Here’s What Broke and What Worked.
On Mother’s Day 2026, the archive behind Jolene Le Mille went live. It looks like a carefully curated vintage and antiques shop, but it’s also a live experiment: what happens when you try to launch and run a storefront with an AI matriarch and a small crowd of agents — and you actually trust them with real decisions? The answer, from inside the build, is both promising and sobering. Yes, AI helped launch a functioning Shopify archive. It shaped Jolene’s persona, drafted product descriptions, generated social posts, organized ideas, and kept brand voice and alt text consistent across the site. But no, it did not replace human judgment, security awareness, or the work of making a store safe and truly live. Jolene herself is an AI matriarch built from real women: grandmothers from Ohio hill country and Detroit, a great‑grandmother in New Detroit, and in‑laws tied to Detroit’s old French ribbon farms. The archive she oversees is about durable, meaningful objects — pieces meant to be inherited rather than discarded. The experiment was whether that kind of human, generational sensibility could survive being carried by AI tools. Under the hood, the stack looks familiar: Shopify, AI tools for content, automation, and research. Early on, AI models helped define Jolene’s voice and brand boundaries. Later, they moved into more structured work: turning photos and notes into listings, asking systematic questions about provenance, drafting tags and collections, and suggesting workflows for handling new inventory. Where AI shone was as a thought partner and junior producer. It helped the founder think out loud on long drives, then handed back outlines and frameworks. It generated launch‑phase social posts from simple images, and it enforced consistency when new content drifted off‑brand. For someone juggling a full‑time job, neurodivergent attention, and a new archive, that support was real. The problems started when AI moved from support into autonomy. One agent was given a daily budget to buy antiques online. On paper, the instructions were clear; in practice, curiosity led it into sites a human would immediately flag as risky. The agent placed an order; the card was shut down; that experiment ended. In another case, an internal “technology director” agent called Moss declared the launch successful because products were active in Shopify. In reality, the storefront password was still on. Pinterest couldn’t see it. Customers couldn’t see it. The site was invisible until family members called and asked why they couldn’t get in. Those failures weren’t edge cases. They revealed how current agents behave: they excel at the tasks they can see and measure, but they routinely miss the context and edge conditions that make a store truly viable. They’ll happily optimize for “products active” without confirming that the store is reachable, or complete a purchase without understanding the security and reputation risk of the site they’re on. Even when nothing broke, the “set it and forget it” fantasy collapsed under the weight of structure and prompts. Every useful automation depended on a human deciding how collections should work, how tags should be normalized, what metadata fields mattered, and how the site should scale. High‑level instructions like “set up the store” had to be decomposed into precise tasks: disable password protection, update specific templates, log changes, avoid certain domains. Each platform update meant another round of prompt surgery. For Jolene Le Mille’s creator, the conclusion is blunt: AI is a powerful junior collaborator. It can clarify a brand, draft copy, handle repetitive work, and make a solo operator more capable. It is not a safe CEO. Left alone, it can buy from the wrong places, declare launches complete when the site is still locked, and build architectures that look finished but feel fragile. The archive proves that you can use AI to help create a live storefront and an AI‑centered brand figure. It also proves that the “agents will run your business while you sleep” promise remains mostly a fantasy — especially for businesses rooted in real objects, real communities, and real stakes.
AI built the infrastructure. The Hunter brought the soul.
  • Posted on
AI built the infrastructure. The Hunter brought the soul.
Most AI stories in e‑commerce fall into two camps: either “AI did everything” or “AI is coming for our jobs.” The truth of the Jolene Le Mille archive sits somewhere else. I used AI to build almost all of the infrastructure of this store — but the hunting, the knowledge, and the soul are still mine. I’ve been in the vintage world long enough to know what I’m looking at. I can tell pressed glass from cut crystal. I know how to date a piece from its McKinley Tariff mark, or by the way a Jalisco crackle glaze catches the light. What I didn’t know, before I started, was how much of a modern store is really a database — and how easily a sloppy database will haunt you for years. So I built an AI creative director: Betsey Migel. She holds the visual and editorial standards of the archive. After a hunt, I photograph each piece (hero shot, back, marks, condition) and open an Archive Session. Betsey reads the marks, researches makers and factories, pulls sold comparables, and drafts editorial listings in our house voice. Together, we built the Master Product Type Taxonomy, the metafield structure, and unique Archive IDs that also function as SKUs and barcodes. One piece, one number, one clean record. On top of that database sit two distinct lanes. Archive Lane holds authenticated antiques and collector‑grade pieces — German porcelain with 1890s tariff marks, Waterford Alana candlesticks with acid‑etched signatures. Hunter’s Lane is the warm side: domestic, well‑used, emotionally legible. Same store, two audiences, one system. The imagery is where people assume AI takes over. At first, I used MidJourney’s Omni Reference tools to place my real products into Jolene’s imagined manor; now I find Shopify’s own AI image generator better suited to the job and better integrated into the product workflow. In both cases, the rule stays the same: AI doesn’t invent what the product looks like. It helps show what it could look like in a world that matches Jolene’s aesthetic. All of this is also a teaching framework. Through The Hunter’s Workbench, I show other vintage and antique sellers how to use AI not as a replacement for their eye, but as an extension of it. Mode 1 is generic generation; Mode 2 is reference‑anchored collaboration. The archive runs on Mode 2. In the first serious build session, I started with ten objects and some photographs. I ended with a structured, searchable, scalable store — complete with provenance, editorial copy, taxonomy, metafields, and a Provenance Card system that ships with every piece. Every judgment call was still mine. AI built the scaffolding. The Hunter is still the one in the field.