I respect my mom more than anyone in the world. If there is one person that I have modelled my character after, it’s her. She’s strong, determined, independent, and always encouraged me to create the life that I dream of.

Despite her instilling in me that I should not tolerate mistreatment from anyone, I’ve noticed subtle ways in which the women of her generation have silently carried the emotional burdens of the men around them. I grew up watching women around me and in media negotiate with themselves over how much mistreatment they are willing to tolerate. I hear them get called crazy or dramatic for feeling any way that the receiver disagrees with. I listen to friends share similar stories of the men around them failing to acknowledge, process, and sit with their feelings - the consequences of unfelt feelings falling as debris onto everyone else in their lives.

That’s not to point any fingers; the social paradigms of centuries past have resulted in this dynamic. (Un)fortunately, in love we try to make things work.

In November 2025, I learned that my partner of four years had developed an emotionally intimate relationship with someone I then considered a close friend. Over the span of six months, their relationship deepened to one of secret texts, hangouts, and her coming to our shared apartment when I was out of town. But because of love, I tried to make things work.

Eight months later, I found myself alone in New York and for the first time in a long time, I felt like a weight had been lifted off my chest. I could breathe without an anxious undercurrent running through my body. My hands were no longer shaky, and my mind finally felt light again. The world regained its saturation through my eyes, and I knew it was because I had distanced myself from him.

I thought I had internalized all the lessons my childhood self learned from observing older women’s examples of what happens when you tolerate mistreatment, but when I found myself in a situation that called for the lessons to be applied, I froze. I eventually called him one night in New York to share how I was feeling and cited everything from “we have different values” to “the timing is not right”. My justifications for wanting to end things were everything except the fact that he had cheated on me. Subconsciously, I was trying to accommodate his feelings.

Whenever I referred to his affair as cheating, he insisted I not call it that, as he claimed nothing had physically happened. I empathized with him and acknowledged that “cheating” is a heavy accusation - a scarlet letter I don’t want to throw around recklessly. So I bit my tongue and never called it that. Looking back, my hesitation to acknowledge the gravity of how I felt is likely what contributed to the eight months I spent enduring sleepless nights and panic attacks.

Since we split, I’ve found myself explaining this story to friends, and through my retellings and observing their reactions, I realized that I spent months pandering to the ego of a man who could not bear the weight of his own emotions, which were also the consequences of his actions. I had fallen into the same pattern as the women I grew up watching and telling myself I would not be like.

It’s funny how my friends all speak of their mothers’ sacrifices as lessons in not allowing themselves to shrink at the expense of another person’s feelings. In reality, these things are much more nuanced and subtle than the way they are discussed, and we often don’t realize we are doing it until the pattern has repeated itself multiple times. From this, I’ve learned to give all the women in generations before me a lot of grace. Their sacrifices should not be seen as a shortage of self-respect, but rather as falling into a pattern that is difficult to spot and even more difficult to break. As much self-respect and pride as you think you have, it’s hard to know when to put yourself first, especially when the ways you make yourself smaller for others are often subconscious.

If there’s anything to take away from this, it’s to allow yourself to feel the way you feel even if it upsets others, and to give the women in your life grace, as we are often too quick to label them dramatic or crazy for simply having emotions of their own.

And if you ever find yourself on the other side of the table, just remember that the best way to not get called a cheater is to not cheat.

Manifesting, energy, prayer, devotion. These are the words we commonly hear when we talk about spirituality. As I’ve gotten older and life has gotten more chaotic with changes in jobs, locations, and relationships, I find myself grappling with spirituality for some sense of stability. The reality around us may be constantly changing, but what can remain a necessary source of familiarity is our internal world.

While I’ve found solace in “becoming more spiritual”, I still have a hard time explaining to my friends what that actually means. Life started feeling better when I told myself that the universe has a plan and that everything will work out, but blind faith does not create a very compelling case for others to follow suit.

If we examine our lives closely, we might realize that how we feel is almost entirely in our control. Putting on sad music may make us cry, and surrounding ourselves with friends may make us laugh. Of course, our emotions are often random and illogical - that’s part of what makes us human - but it is possible to engage in a controlled chaos that allows us to experience all the feelings of being alive while steering our lives towards the reality we hope to have.

In bell hooks’ All About Love, she establishes that love is not random. Love is a choice, and when we approach love as a deliberate practice, we may find one that is fulfilling and helps us meet our highest selves. The same approach is true in every aspect of life. Our subconscious soaks in every experience we have, including the things we say to ourselves and the attitudes of those around us. It absorbs the themes of the music we listen to, the ways our friends react to things, and the thoughts we have about ourselves when we look in the mirror. Unlike our conscious minds, our subconscious consumes everything we feed it, regardless of whether it is true or not. To nurture ourselves to become the people we want to be, we must treat our subconsciouses like children and be selective about the things we say and expose it to.

Sometimes it feels like the world is drifting away from spirituality and isolating us from deliberate practices. While some may call manifestation a “woo woo” activity, it is a form of ritual that helps us rewire our subconscious to believe in a reality that is yet to exist. This belief guides our actions, and our actions guide our outcomes. Rituals cannot be random or spontaneous; they must be deliberate, intentional, and practiced consistently. To develop a ritual and make time for it is a choice that requires work - a ritual is never random.

The life we end up living is a culmination of how we live every single day, and while it is easy to rely on instinct, we need to remind ourselves that the habits that compound can only be built with work. Trusting the universe has given me a sense of calm, and I’ve realized that I was able to build this trust by dispelling the feeling of randomness that life can have. I can choose not to wallow, to speak to myself kindly, and to surround myself with people who inspire me. If we practice devotion to ourselves by living with intention, the only thing left to do is trust that the universe will take what we’ve given and run its course.

About a year ago, I made one of the biggest decisions in my adult life: I left the job I thought I had always wanted and moved back home to Toronto. When I left that job, I challenged myself to be vocal about it in case it could help someone else who was going through a similar phase of questioning the reality they had constructed for themselves. To my surprise, my experience resonated, and people - mostly students and interns - still reach out to me about how I made my decisions. In hopes of spreading any helpful information and thought processes, this is me writing down the core principles that guided me through that time.

Observations, Incentives, & the Build Up to Leaving

Since I left finance, I found myself spending a lot of time in tech and art circles. One thing these communities have in common is that they attract ambitious people - not necessarily ambitious in a traditional sense, but ambitious in their willingness to dream. People have ideas, visions of things they want to build, opinions about how the world should work, and a desire to manifest those ideas into reality. They are compelled to create.

I didn’t always run in these circles. My last job was in investment banking and my observation is that the people who go into finance are ambitious too, but in a different way. Finance is a traditional industry. It’s structured, hierarchical, and relatively predictable. There is a well-defined path forward. The people who choose it often share the same desires for success, influence, capital, and achievement, but they pursue those goals through stability rather than uncertainty. It’s a much more risk-averse environment.

Leading up to my decision to leave, I spent months ruminating and wondering what was wrong with me. Was I not smart enough? Maybe I didn’t have the grit? I eventually realized it was less about competence and more about fulfillment.

I grew up in a creative environment. I studied art and ballet, and my sister is a musician. Yet over time, I became increasingly risk-averse. As a teenager and young adult, I was always optimizing for the clearest path to what I wanted. At the time, I wanted two things: financial independence and to live in New York City as I had fallen in love with it while visiting as a kid.

In university, I concluded that the most direct way was to get a good job.

For a while, everything seemed to be working out great, but eventually I found myself waking up every day to do work that felt hollow and filled me with dread.

Eventually, I hit a wall.

I was forced to confront a reality I had been avoiding: this job is not for me.

Breaking Point

The thing I hated most about this period of my life is that I always felt like I was waiting for my life to begin. I was perpetually anticipating some future event that would finally make me happy because my daily reality felt the opposite. Perhaps it was an upcoming vacation or a friend visiting, but when there was nothing to look forward to, every day felt like trudging through mud. I was perpetually living in anticipation of a future state of temporary happiness.

I don’t remember exactly when, but one day I must have had enough and decided I didn’t want to live like that. When I made the decision to leave, it ultimately came down to three principles. My advice is to form a set of your own and to think of your principles less as objective truths, but more as standards of living that you radically, delusionally, and stubbornly believe in. Being indecisive and unopinionated is the worst thing you can do for yourself in this situation.

The Principles

1. Your Time Is Finite

Around the same time I realized how much I hated my job, I was also dealing with some health issues. I was depressed, neglecting my body, and watching one problem snowball into another.

It forced me to confront something obvious that I had somehow never fully internalized: I am not going to live forever.

I was 22-23 at the time, but even then, I began to realize that my body would not always be as healthy as it was in that moment. I wouldn’t always have the same energy and freedom available to me.

So why would I spend my limited time doing something that made me miserable?

A major contributor to my decision to leave was my doctor’s suggestion of trying antidepressants. What struck me was that the source of my unhappiness felt incredibly obvious. I didn’t feel depressed for no reason. I felt depressed because the work I was doing every day was deeply unfulfilling, and the lifestyle surrounding it wasn’t what I wanted either.

I was living in New York City, yet I never had time to enjoy it. I didn’t have the energy to meet people, pursue hobbies, or explore the city I had worked so hard to get to. I felt no sense of purpose from the work I was spending 12-16 hours of my days on.

I remember thinking before medicating myself, I can name at least five things in my life that are making me unhappy. Maybe I should try changing those first.

I took a leave of absence. I went back to Toronto. I spent time with friends and family. I rested. I reconnected with things that energized me as a kid.

Unsurprisingly, it helped. Life started feeling colorful again, I had more energy and I felt like a living person.

(Obviously, listen to your doctors - I’ll preface that this decision came from extensive conversations with mine.)

Your time is finite. There will come a day when your body doesn’t work as well as it does now. You’ll have more responsibilities and less freedom to take the seemingly reckless gambles you can in your 20s. This is a rare period of life, and I didn’t want to waste it.

2. Scarcity Is Not Real

The second principle is that scarcity is often an illusion.

More broadly, I believe your life is largely what you make of it. Of course there are systemic barriers and circumstances outside of your control. Despite that, I think you have to develop an almost irrational belief in your ability to change your situation.

People who come from environments that value prestige and stability tend to be highly risk-averse. When you’re in those circles, it’s easy to think:

“If I leave this job, what happens next?”

“What if I never get another opportunity like this?”

“What if it all goes bust?”

You may not realize it, but by thinking that way, you are imposing an artificial ceiling on yourself. In dire situations where your only path to survival is to push forward or perish, it becomes incredibly clear what you need to do, and doubt is not an option. Building off my last principle, I rewired my brain to pursue a better place like my life depends on it because frankly, it does.

The other framing of this thinking is to consider the cost of staying. What is the opportunity cost of all the things you forgo by staying in a place that doesn’t satisfy you?

There are infinite ways to change your circumstances and we are blessed to live in a time where the internet exists. Try them all - if the alternative is living an empty life, then you truly have nothing to lose. Don’t be a fig tree victim.

3. Your life is your own

I hate to admit it, but I used to think about what other people thought of me a lot. I made decisions based on my projection of how others might view me, and it took a lot of self-reflection to realize that the act of thinking others would judge me was a projection of how judgmental I was of myself. I would be lying if I said I don’t care anymore, but I learned to tone it down so that my worries feel more like quiet intrusive thoughts than debilitating anxieties.

For months, I barely talked about what I was going through because I didn’t want people to think I was struggling. Having attended a striver high school and worked in a prestige-centered environment, I internalized all the judgments and values around me, whether I genuinely believe in them or not.

But if you treat what you do with your life as a matter of survival, you raise the stakes high enough to tune out the noise. To me, living an empty life is only marginally better than not living one at all, so if the only way to live is to choose the things that make me feel happy, fulfilled, and energized, then suddenly it doesn’t matter what other people think.

This applies to advice too. When I was in the process of deciding to leave my job and forgo what I had lined up next, I spoke to all my mentors and close friends. The likely outcome of this is they will be supportive, but you won’t necessarily hear what you want to hear because they are not you and are sharing advice from the perspective of someone with a different set of experiences, goals, and values.

The last thing is that probably no one is thinking about you. People tend to be the most focused on themselves; that’s just the way we are. You do not take up nearly the mental space in other people’s heads that you think you do, and that is a good thing.

The moment you allow someone else’s framework to determine how you live, you stop living for yourself.

A Chinese fable someone shared with me that I think captures this well is The Little Horse Crosses the River.

Final Thoughts

Whether you agree with my principles or not, I strongly believe that you need a north star in your life if you want to persist through anything. When times are tough, you need to have a grounding belief that reminds you of why you are doing this. Maybe it’s delusion, or coping, or both, but these are the things I tell myself to keep moving forward.

A helpful exercise one of my mentors made me do was to come up with a list of priorities, force rank them, and score all of my options against them to try to quantify my decision.

At the end of the day, everything I’ve written here comes from my own experiences and is not a universal truth. You can borrow mine, but I highly recommend challenging yourself to develop your own. Also, give yourself some grace and forgiveness - I think we’re all figuring it out as we go.

There seems to be a paradigm where we associate vibe coding with sloppy and low-quality outputs. After all, it gets its name from the fact that it’s based on ~vibes~ and not specifics. However, vibe coding has armed people like me with the ability to create outputs in a medium that wasn’t possible for us before - at least not without the investment of learning to code by hand. It’s become so easy to write a program that some might even say it feels like magic.

This morning, a tweet from my friend Ivan (co-founder of the AI company Cohere) came up on my timeline:

(Technically, this is Nick Frosst’s quote)

You can’t think the technology is magic.

My first reaction was “Why not?” What we consider well-designed technology usually carries a sense of mysticism. The times I’ve been most impressed by technology were ones where I’ve thought, “How did it just know?”

Then I considered where this statement is coming from. Cohere develops technology, and to develop technology, you need to understand technology. Having a deep understanding will inevitably clear the fog and dispel the mysticism. Magic can feel like a shot in the dark or a roll of the dice, but to make something groundbreaking, one must be precise and know what they are doing.

This makes me think about some of the conversations I’ve had about my projects recently. When people ask how I built them, my response is usually a credit to Codex/Claude, and the reaction I get is often “Really? Is that all?”

I think there is still a high degree of mysticism around how certain AI-generated outputs are created, and while I acknowledge that LLMs are black boxes, the ways we can wield them are more similar to discovering new land than conjuring portals. While a well-designed UI may have been created using AI, conversations with designers like Liam from Browserbase reveal that good design still follows frameworks such as font selection and consistent alignments (not magic). Charting this territory looks less like stringing together the perfect description of vibes to throw into the black box, and more like a guide of discovered or developed functionalities and how to apply them (i.e., this guide).

Likewise, AI may be able to execute a novel idea, but the idea itself is still the distillation of the creator’s references and experiences of our world. To an outsider, how another person comes up with something they consider creative and unique might seem like magic, but to the creator, the strings that tie everything together are very clear. The models will eventually learn many of these frameworks as we continue to document them, but my point is that the reason why something looks good or feels unique isn’t magic, it’s a principle that someone had to figure out.

I’m not sure what the most common definition of “vibe coding” is right now, but I like to think of it as a neutral executional process or a tool rather than something that carries a quality label (I understand why we associate vibe coding with low quality outputs since more amateurs making things will inevitably result in more low quality results, but I think that is more of a symptom than an inherent trait). As coding agents make it possible for anyone to express an idea through technology (software + hardware), I’m optimistic about a future where we have dedicated creative tools like Adobe suite, DAWs, and Figma for this growing format of expression. I place my bets on the ones rooted in process, understanding, and precision - not the ones trying to sell magic.

If you were a teenager (or chronically online) in the 2010s, you might recall the insurgence of the word “aesthetic”. It began with people using it to describe a certain vibe - popular aesthetics at the time included twee, grunge, and indie sleaze. As internet terminology evolved, the word “aesthetic” took on its own meaning. People dropped the descriptor and just called things “aesthetic”. If something was visually pleasing, it wouldn’t get called a certain aesthetic, it would just be described as “so aesthetic”. Coachella was so aesthetic, Lana del Rey was so aesthetic. We started to use what was previously a neutral noun as an adjective that was attached to a standard of liking decided upon by the masses. Despite having an era-defining visual identity, no one ever described the Supreme/sneakerhead/hypebeast subculture as “aesthetic”, perhaps because it didn’t align with the mass market’s bar of visual appeal.

This marks my first memory of the mainstream discussion eliminating the nuances of subcultures and transforming what was once a noun to be associated with a descriptor into an adjective that carries an implied standard of excellence.

Fast forward to the 2020s, and I’m seeing this same phenomenon happening to the word “taste”. As our Silicon Valley overlords have made progress in conquering workflow optimizations, they’ve begun to fixate on optimizing the next thing: taste. By now, taste is a tired discussion. We’ve mulled over why it is or isn’t important, who has it and who doesn’t. In my opinion, these conversations miss the critical nuance that taste itself means nothing without attribution to whose taste it is and how it was developed. Here lies what I think the actual problem is: you are not tasteless, you just don’t have the vocabulary to describe what you like.

In Virgil Abloh’s Harvard GSD lecture he states, “You also have to have mentors, dead or alive. You have to connect with a body of work or someone who formulated a thought and an aesthetic, and then build yours upon theirs.” The reason why you like the things that you like is because you are influenced by your environment, the media you consume, and the idols you look up to. In a conversation I had with (my favorite podcaster) Jackson Dahl, he brought up the idea that creators in art and fashion tend to be referential and draw on inspirations from the past, while modern technologists tend to be forward-looking and spend little time looking back. It is important to acknowledge that modern technology is, in fact, built upon the historical context that came before it, yet time and again, we see people striving to invent completely novel things. An example that comes to mind is technologists’ use of the words “disruption” and “first principles”, implying that things must be completely torn down and built from zero to truly push us into the future. However, this rarely applies to tastes and aesthetics as they aren’t simply pulled out of thin air. Artists are revered for being able to draw on niche references; look how Law Roach became the most famous stylist of our time for his god-tier archival pulls and historical references. Every artist draws inspiration from a web of experiences and inputs that are mentally sorted into preferences that shape their outputs. It’s difficult to create something new without building on the past, and having strong references is the foundation for pattern matching and ascribing vocabulary to your preferences.

I’d caveat that this isn’t to say that technology is not referential - the resurgence of 90s Apple marketing and 80s retrofuturism aesthetics have been key points of contention. This has even sparked a circular dialogue back to some people calling for more innovative and novel aesthetics. The same frustration has been echoed in other media like fashion and cinema.

The more you consume art, the better you can describe it. You begin to pattern-match certain visual themes, techniques, and compositions. You then learn to attach words to these patterns, which form the basis of your taste vocabulary. This is a critical consideration as AI unlocks a whole new universe of creative tools and enables people to experiment with new mediums. Consider image generation, where the model inputs are text and reference images. If I wanted to master wielding image models, I would first need the vocabulary and reference library to get the outputs I want. It is more direct to use words like “skeuomorphic” or “bento grid” than it is to string together sentences that broadly describe these things. Likewise, it is more direct to say “90s editorial candid digital camera shot” attached with a reference to a Calvin Klein BTS shot to get a very specific image than it is to describe the vibe with zero references. Jamey Gannon comes to mind as someone who has mastered this skill.

What the internet often calls tasteless are things made with very little intention and feel detached from their broader contexts. An inexperienced person making an AI-generated garment might be labelled tasteless, perhaps because they don’t yet know the best practices that have been refined through millions of prior iterations. However, a fashion designer using AI to develop new prints might catch and tweak subtleties like what balances a color palette. The difference is that an outsider who lacks the context and vocabulary to create an intentional output often ends up creating things that feel uncanny.

While we can attempt to calculate or develop a criterion for what makes something tasteful or not, we must remember that this is all subjective. Your taste may not be someone else’s taste, and that is because taste itself is a thing, not a descriptor. If we truly think it is important to add a dimension of aesthetics and intention to the technology we build, it is important for us to learn how to wield artistic vocabulary, especially as many of the tools we use are currently language-based. The outputs of an LLM are only as good as the inputs you give it, and without grounding our abstract expectations in usable terminology, it is extremely difficult for us to wield these tools to create the outputs we are looking for.

I am the beneficiary of the kindness of dozens of strangers. That’s not to discredit any of my personal achievements; rather, I want to acknowledge how important it is to have good mentors, as that is overwhelmingly the one piece of advice I feel qualified to give others in the cycle of paying it forward. While the advice is obvious, I find that most people struggle with the “how” part, so I’m going to explain how I developed most of the influential mentor relationships in my life.

Be Curious and Make Friends

No one I’ve ever called a mentor was someone I specifically sought to be my mentor; it just happened. Call it naivety or blind optimism, but I genuinely believe some people find fulfillment in helping others without gaining anything but a good feeling. Perhaps your relationship will be symbiotic one day, but in the beginning, you probably have nothing more to offer other than being pleasant to be around and easy to coach. I’ve met numerous mentors through school clubs, friends of friends, and even the internet. These are typically people who have already walked a path I admire and people that I am genuinely curious to get to know. I approach these relationships the same way I approach making friends: I want to understand their lives, how they think, what shaped them. There’s already a lot you can learn from that information without extending the relationship, but the first step to anyone having a vested interest in your success is for you both to somewhat understand each other as people.

Don’t be Selfish

Per my previous point, a mentor is likely invested in your success for altruistic reasons - there’s very little they could possibly want out of you right now. They spend their time and energy on you because they believe in you. The respectful response is to manage your expectations for them. Your mentors won’t save you; they are trusting you to save yourself. Don’t treat them like a genie with all the answers. Use your brain, put in the work, be open to learning, and show them why they made the right choice to bet on you.

Lean In

Sometimes, if you’re really lucky, a mentor comes to you first. When I was in the 10th grade, my history teacher saw that I was an ambitious kid who was curious about the world, but struggled to get past the insecurities rambling about in my head. After class, he would share book lists and student program pamphlets with me, all I had to do was engage with the material and his outreach. For those of us that confidence doesn’t come naturally, be aware of people who express their belief in you, as they will be the champions of you learning it for yourself.

Be Genuine

People want to help you if you are doing things for the right reasons. You need to be someone other people want to see succeed. This usually means being humble, open-minded, and thoughtful. In short, don’t be an asshole.

Put in the Extra 5%

The bar is probably much lower than you think it is. Most people don’t share their work, cold DM strangers, or follow up with people. I find that as I’ve gotten older, it helps to make myself more legible to others - who is Jess? What does she do? What is interesting about her? What does she think about? This could be through any body of work that helps people pinpoint exactly who you are and why they might want to talk to you. The next step is to take the initiative to actually start conversations and keep them going.

Try to be Helpful

I’m saying this for the third time now - there’s probably nothing of tangible value that you can offer a mentor, and that’s okay because they’re probably not looking for you to do that. What they want is someone who is actually going to learn and become better as a result of their sharing their time and wisdom with you. The biggest ROI you can give them is to be teachable. That said, you can always attempt to help. Send them articles you think they’d enjoy, new ideas you have, and ask them questions that will make them think.

Gratitude

I don’t speak to everyone I consider a mentor frequently, but I will never withhold sharing my gratitude when it comes to mind. Even if we haven’t spoken for years, I will always be sure to reach out if I’m in their city, update them on how things have developed, and send them all of the congratulations for their milestones too (they are also navigating life for the first time). There’s nothing to lose from sending a “Hey, remember that conversation we had 3 months ago? I’m still thinking about it, and your words really changed the way I think. I decided to do X, which led me to Y. Just wanted you to know the impact you had, and I appreciate you for it.” Generally, in life, don’t bite your tongue against expressing gratitude.

Closing Thoughts

I use the word “mentor” loosely because there are people in my life who have swooped in at unexpected times to shine a flashlight through a foggy path. There are also people whom I’ve only had one or two conversations with, whose words have changed the trajectory of my life. There are even people my age or younger who have asked me questions that made me introspect on things I’ve never thought about before. I think learning can come from anyone, and the people you surround yourself with ultimately have the biggest influence on how your life plays out. Taking my own advice, I’ll use this as an opportunity to thank all the people in my life who have offered me their generosity - I appreciate you!

Brunello Cucinelli said, “Technology is a blessing from creation, but sometimes it steals the soul that creation bestowed upon us.”

I think humans are incredibly good at identifying things with soul, echoing the phenomenon of the uncanny valley.

Technology creates new mediums as it enables our expression to take new forms. For example, the invention of textiles enabled canvas painting, the invention of the camera enabled photography, and the invention of the computer enabled digital art. To Cucinelli’s point, a medium is simply a vehicle for creation; however, not all creation carries soul.

The way we feel when we look at art or the undercurrent of energy we feel when we wander around a metropolis might have to do with our intuitive connection to human creation. Look at the world, and what is hiding in plain sight is the fact that everything around us had to be created. Each skyscraper had to be constructed from steel beams, and every garment we wear had to be milled from fiber. Humans have infused the world with soul not by inhabiting it, but by entrenching centuries of creation into what it is today.

Creation is what makes us human. One can argue that human inclination to materialism stems from capitalism, but I think part of it is the connection we have to things that someone had to create. Every artifact is a souvenir of our time on this planet.

The Implications for Technology

What happens when technology begins to mimic human behavior? Today, it can speak to us, rationalize, and take action based on its judgments.

I have two main considerations for this:

1. Is it possible to create soulful technology?

I think yes. Apple is a prime example of this - products resonate with users as they were designed with deep intention for how they would interact with humans and consideration for the user’s experience. These are products that exhibit empathy, conceptualized by people with lived experiences.

2. Should it be obvious to users that technology is not human?

It depends. Products meant to appeal to human emotions (such as digital companions) are only effective if they are indiscernible from a real human. However, it is important to consider the dangers of becoming overly reliant or trusting of something that fundamentally does not experience empathy (at least not yet), leaving us vulnerable to manipulation, exploitation, and social alienation.

The Taste Conversation

My qualm with the taste conversation is that no one has managed to define what taste is, what good taste looks like, and what bad taste looks like. Perhaps it isn’t taste we are trying to describe, but soul.

When something is in “bad taste”, the viewer can almost instinctively identify it. It rings alarm bells similar to the unpleasant sensation we feel when we look at uncanny valley figures, where it is close to human but not quite. The reason why we trust some products and not others has less to do with aesthetics or references, and more to do with how well these products emulate craft and genuine intentions - things that are soulful (I’d caveat that these things can also be mimicked).

Why this Matters

We are entering an era where technology is becoming reproductive. New technologies are being created at the speed of thought, and we are not far from a time when technology will be creating itself - we are already seeing this with agents creating new agents.

The sentiment I’ve been seeing feels quite grim as both the artistic and technologist communities scramble to preserve their craft while AI becomes increasingly capable of replicating it. Just like how humans domesticated animals that otherwise would have been our predators, I think this threat is best addressed as an act of offense, not defense. We shouldn’t be focused on outrunning or delaying the onset of the inevitable; instead, we should be figuring out how to harness creative technology to our advantage.

In short, I am not that concerned about technology impeding on the things that make us human because history has proven that we are incredibly good at wielding the forces of the world to our advantage.

What Are We So Afraid Of

We already live in a world of Dubai chocolate matcha Lububus, we can’t pretend like slop did not exist before AI. “Tastelessness” has been a thing long before technology made the barriers to creation nearly non-existent - one could argue that the proliferation of mass-market, low-brow materials is more an outcome of consumerism enabled by technology that makes it possible to satisfy demand.

My observation is that the market has been relatively decent at regulating and preserving taste and soulfulness in AI-created things. Any fashion brand that releases some sort of AI creative asset inevitably receives angry backlash in their comments, and every software company that attempts to rage-bait their way to adoption gets the finger of poor taste pointed at them. Conversely, those who have been able to harness tools like image generation as a refined skill have received praise and market acceptance.

The Future is Not So Bad

To Cucinelli’s point, I firmly believe that AI has a place in the things that make us human, as all dimensions of the world bleed into each other, including technology. Following his quote about soul, Cucinelli states, “AI will be just a partner for us, and we will rediscover the value of truth and human beings.” I acknowledge that AI perpetuates the uneven playing field that has economically disadvantaged artists and craftspeople, but these are dynamics that existed long before AI, and the backlash is merely a symptom of a long-overdue reckoning. I’m optimistic about a future where technology enables new forms of creativity that challenge the dynamic of the status quo, and that people will continue to flock to soulful creations. While expanding the creative pie opens the door to more “slop”, it also allows us to dig deeper than we ever could have and play in newly charted territory.

Brands lose money each time they acquire a new customer.

• Fashion & apparel average customer acquisition cost (CAC): $129

• Brands typically lose $29 on each new customer acquired

• First-time conversion reduces gross profit margins by 23%

• Every customer who doesn’t come back is a guaranteed loss

The cost of acquiring customers is going up.

• CAC has surged 222% over eight years across e-commerce

• Meta's Q1 CPM hit an all-time high in 2025, up 19.2% Y/Y. The average cost per lead on Meta rose to $27.66, up ~21% from the prior year, while conversion rates dropped from 8.67% to 7.72%

• Google Shopping ads saw a 33.72% CPC jump, and overall ROAS declined 10% in 2025

But brands can make their money back.

• Repeat purchases generate an average of $39 in profit per transaction

• That means on a $129 CAC, a customer needs to come back ~3.3x for the brand to break even

• E-commerce businesses aim for a standard LTV:CAC of 3:1, meaning the customer should spend 3x as much money in their lifetime as it cost of convert them

  • Since apparel CAC is $129, a 3:1 LTV:CAC implies customers should spend $387 with the brand to reach that standard

• The cost of retaining a customer is ~5x less than acquiring a new one

Why Does This Matter

Apparel unit economics do not favor profit. On a gross margin basis, most brands can’t support the profit per unit needed to cover operating overhead without reaching substantial scale. Add rising customer acquisition costs to get those sales in the first place, and an increasingly saturated market where consumers have more options than they know what to do with, brands need to invest in low-cost repeat buyers to break profitability.

In apparel DTC, 60% of revenue comes from repeat buyers, while they only account for 26% of the actual customer base. These 26% of consumers are carrying brands’ profitability as they no longer bear the $129 cost of initial conversion. A Bain/HBS study shows that a 5% improvement in retention can drive a 25–85% increase in profits. Yet 54% of B2C brands allocate more than half their marketing budget to acquisition, while only 13% prioritize retention.

If It’s Not Already Obvious…

Repeat customers are your saviours. Marketers love to talk about brand loyalty, but what does that actually mean?

• Repeat customers cost a fraction to sell to, since the CAC is already sunk

• Repeat sales are necessary to break even on CAC to begin with

• The highest leverage segment, as they make up a minority of overall buyers but a majority of sales

• Repeat customers spend 67% more per transaction than new ones, driving higher AOV

• They are less price sensitive, meaning brands don’t need to rely on discounting to drive sales and can sustain higher full-price sell-through

• They refer new customers, creating zero-cost, newly acquired customers

Brands are Leaving Money on the Table

To emphasize again: every customer who doesn’t come back is a guaranteed loss. Fashion loves to talk about community building and experiential marketing. Yet many of these efforts focus on converting new customers (i.e., brand activations) or on developing the brand’s persona (dare I call it vanity marketing).

The current state of fashion’s post-purchase customer journey is questionable. What happens after you buy a garment from a new brand? Maybe they’ll spam you with emails (if you’re lucky and the brand is Reformation, at least these emails will be entertaining). Maybe they’ll sign you up for a points program where you’ll never accumulate enough to benefit. If the brand is really desperate, they’ll try to win you back by offering discounts.

Brands need to reassess who they are focusing their engagement efforts on and how. The last few years have seen a slight downturn in lavish influencer trips, largely driven by the broader economic downturn and poor cultural reception in these times, which has opportunistically forced them to reassess where these budgets get allocated to instead. While influencer trips attract eyeballs and new customers, I think brands should invest in rewarding existing customers and driving retention, a strategy Mytheresa has been successfully deploying for years.

It all starts with data. Brands hold massive CRMs of every customer who has purchased with them, their subsequent interactions, and consumer behaviors. Our next focus with Inseam is to design effective retention strategies, starting with customer segmentation and behavioral analysis. By employing a revenue-first approach, we hope to push even small brands to profitability and flip the current unit economics in favor of our favorite designers.

“I left my six-figure job to build a startup” is a corny opener, but it is the truth. I’ve been very transparent about my journey across platforms like Twitter, TikTok, and Substack, and my initial fear of being known as “the girl who fell off” nearly manifested as I’ve instead been known as “the girl who quit the job she hated”.

I built my first real business, an apparel brand called Serena Li (with my best friend and co-founder), while I was still working as an investment banking analyst at Evercore. I’ve since turned down an offer to join L Catterton’s growth tech investing team as I realized that finance is not the path I’m intended to take in this life. Do I regret it? Absolutely, unequivocally no.

My first endeavour into entrepreneurship was at age 14. I attended a local pitch competition called Y2, which ultimately instilled in me a lifelong curiosity about building something meaningful that refused to die, no matter how hard I tried to kill it. Throughout my five years of being a college investment club striver and brief stint in New York finance, there was always a little voice in the back of my head: what if you could build something?

When people reach out from my TikToks or Substacks, a question I get a lot is “Do you ever think about what you gave up?” I understand why they would ask this, as this is a fear anyone would have to wrestle with as they contemplate making a massive trade-off in their lives. I think about this in a similar way to how I think about regret: What’s the point? I could spend all day comparing myself to my peers who have gone on to take associate promotions or make shut-up money in private equity (caveat that shut-up money isn’t f*ck you money), but that’s comparing apples to oranges. I’m not them, and they are not me.

Many of us live in a meta where life is quantified by the achievements we’ve got under our belts, and ironically, we get so wrapped up in it that we don’t realize the very act of optimizing for achievement is what holds us back. A big reason why I left my job may have been the push factors that I didn’t love being chained to a bleak, fluorescent-lit desk, staring at Excel for 16 hours a day, or that I simply didn’t care about alternative assets M&A, but the pull factors are just as strong. I could never have the time or mental capacity to go all in on building something that was my own if I stayed in that job. I could not have gone through the painful but satisfying process of going deep, discovering a problem, having a hypothesis, and getting myself an answer. Put simply, if my global maxima in life is to build something meaningful, staying in finance was merely the local maxima of being 23 years old.

I think much of life is driven by the permission we grant ourselves to do things. Academic settings may train us to ask teachers for permission, and corporate settings may train us to ask managers for permission, but at the end of the day, your life is your own, and what you do is merely what you give yourself permission to do.

Now I’m 8 months into trying to build a startup at the opportunity cost of a $350k salary, good perception from random people, and a (relatively) stable career trajectory. But I’ve also learned more in this time than in my entire two years in finance. At 24, I think I owe it to myself to permit myself to try. The cost of failure is low, and the cost of never trying is irreparable.

At this very moment, I’m still in the red from doing all this. My startup idea didn’t take off as I had delusionally dreamt, but these things take time. I’ve learned to eat sh*t more than I would like, and every day I still manage to get up excited to try again because I am giving myself a chance to figure it out. When we live in the high-achievement meta long enough, we become blind to our own egos. Ego isn’t always high self-talk and arrogance; it’s often hidden in the subtleties of how you are willing to be perceived by others. I never minded failing - I knew that it made me stronger, but a muscle that took time to develop was failing publicly, and knowing how to ask for help. I’ll be the first to admit this isn’t working, but I’ll also be the first to get my hands dirty and fix it. No one is going to rescue you except yourself.

Most of us are lucky to even have these options to contemplate in the first place - don’t take that for granted, and at the very least, give yourself the chance to just try.

On November 2nd, 2025, I started building Inseam. I had been toying around with AI implementations in apparel businesses for a while, starting by building tools and automations for my own brand, Serena Li. I spent time talking to other operators and designers in the industry to see if they were facing similar challenges with running an apparel business and if any of the tooling we built for ourselves could be useful to them. Fast forward 4 months, and we had spoken to a handful of potential users, gathered and analyzed a long list of problems, yet the products we launched still failed to find market fit.

Building a startup is hard. The weight an idea carries in success is minuscule relative to that of the team and their execution. The game is 90% mental, 7% skill, and 3% luck. This is my reflection on building Inseam so far.

Problem Discovery

Before making Inseam official, I had spoken to about 30 people working in various roles across design, merchandising, production, and even a few small brand founders. I probed at their work and got detailed walk-throughs of their processes - you could call this phase my effort to gain earned insight.

Problems we noticed:

1. Sampling: The sampling process is incredibly wasteful. Garments are made to be edited and never worn. Brands typically go through 3-4 sample rounds before finalizing the product, if it gets to that stage at all. For a small brand, this can be time-consuming and costly, as samples can cost anywhere from $100 to $2,000 per round.

2. Production: The production process is manually demanding, yet deeply repetitive. Production coordinators spend their days chasing vendors for updates, working around supply chain and logistics issues, and toggling between email and spreadsheets.

3. Meeting factory MOQs: This predominantly affects small brands that do not sell enough volume to meet bulk order minimums. Factories typically require brands to order at least 50 or 100 units of a product to cover their material and setup costs. When brands can’t reach factory minimums, they are forced to look elsewhere, such as local small-scale production, which can be 2-3x more expensive.

Inseam 1.0: Sketch-to-Render

We decided to start at the design phase, since that’s where we had the most visibility working alongside my friend and brand co-founder, Serena. Serena designs for her namesake label and on a freelance basis for other projects. Two things stuck out about Serena’s workflow:

1. She was already using design renders generated from ChatGPT in her deliverables to clients.

2. When working with manufacturers, she was dealing with miscommunications and misinterpretations from them. This led to her staying up after midnight to get on the phone with overseas vendors and sending multiple iterations of marked-up techpacks to highlight specifications.

We started building a sketch-to-render tool that helped designers visualize flat sketches, make edits, and export them. This was meant to alleviate the sampling process and help designers with visualization. Here’s where things went wrong:

1. This tool is a commodity. It’s already easily done in any AI model that has imaging capabilities (Nano Banana, ChatGPT, even broader creative tools like Flora). The UI of having a fashion-specific canvas with precise editing capabilities was not the bottleneck, as the process of designers making manual updates to flat sketches is already very simple - having direct edits in renders creates a marginal improvement, if any.

2. We sucked at distribution. We got 50 people onboard but asking them to try it felt more like asking for a favor than providing them with value. I think this happened because we failed to build something people wanted, and we didn’t try hard enough to find the people who did want it.

3. There was no clear ROI. It was difficult to attribute any improvements in time or cost to using our tool. Samples had to be made regardless, and designers gained no efficiency improvements (whether this is a space where we should even be optimizing for efficiency is its own debate).

In retrospect, we should have done two things differently: spent more time validating the problem before building, and focused more efforts on distribution. If I were to do this again, I would spend time asking users:

• If sampling is a huge issue, have you done anything to try to get around it?

• Have you searched for or tried any products that already address the problem you’re describing?

• How much would you pay to have this problem solved? → The answer to this needs to be substantial; otherwise, it’s probably not that big of a deal.

Of course, there were many other flaws with the idea and approach. 50 people tried it and never touched it again; that was discouraging. I think this comes from my inexperience as a first-time founder, but I took that negative feedback as a strong signal to react fast and pivot away. Perhaps I should have been more ruthless in distribution: posting 4 TikToks a day, sending cold emails and LinkedIn messages, and being bolder. Or perhaps I made the right call, and that product wouldn’t have gone anywhere anyway. We don’t know, but regardless of my distribution efforts, the cracks in the idea were already starting to bleed through. Designers are a difficult wedge as their workflows don’t hold critical pain points for the business, their tooling is relatively well established with plenty of commodified options on the market, and the demographic is generally more hesitant to touch AI.

To caveat, there are definitely players in this niche who are scaling, with Vizcom being the most notable. We found them by speaking to designers at Nike who said the company had onboarded to the platform. They address horizontal markets like industrial design, footwear, and more. I think the natural next step for them is to embed themselves deeper in the production process by owning production assets and entering the vendor side of the equation.

Inseam 2.0: Production Tracker

Fast forward to mid-December. We went back to our original notes from speaking to fashion operators, and saw a problem that was much louder: getting a product made with a manufacturer is an arduous dance between emails and spreadsheets. From our interviews, people overwhelmingly shared the same issues:

1. Working with manufacturers is a pain. They can be slow to respond (sometimes they ghost you altogether), and miscommunications are common and can be costly.

2. Everything needs to be tracked, whether in software like a PLM or a spreadsheet. Every product needs to go through checkpoints such as lab dips, sample rounds, and fit adjustments, which all need to be tracked or else they will fall off schedule. Deadlines are everything in fashion - being late to a season or missing a delivery date can cost you your inventory sitting unsold.

3. Only large enterprises ($50M+ annual revenue) were using software like PLM systems. Enterprise software is feature-heavy and expensive to implement and maintain. Small and medium brands mostly relied on a team of production coordinators keeping Excel sheets up to date. Even at large enterprises like Lululemon, PLMs were underutilized, and production coordinators still relied on their own personal trackers.

So we built an automated tracker that would sync to email and Slack inboxes, ingest brands’ pre-existing spreadsheet trackers, and auto-update the relevant fields as new information came in. We suffered from a very similar issue: no clear ROI and a pathetic distribution attempt. While we were saving people time from needing to automatically update spreadsheets (an estimated 1-2 hours/day), it was difficult to attribute this time to anything meaningful on the bottom line. Junior employees seemed apathetic as this was an accepted reality and the core function of their jobs, and we struggled to build trust with senior operators.

This is where we ran into a chicken-and-egg problem. To do effective distribution, it feels necessary to present people with a compelling reason to speak to you. In my head, this meant having a product or function that we could basically implement immediately to help drive value for them. However, this goes back to the challenge of misallocating resources to building something that people might not want.

At this point, we had a product that we showed a few people with lackluster reactions, continued cold outreach with very little success, and were falling into the same cycle of getting down on bad feedback.

Inseam 2.5: Production Agent

Now it’s mid-January. I’m noticing a pattern where each of our mini experiments runs for about 3 weeks before we move on. I think that’s a good pace for an iterative loop; however, our shortfall is that we struggle to complete one full high-quality experimental cycle. We always fall short on distribution, and I would bet that our order of operations needs to change to prioritize that above building the product.

Our production agent is the closest we’ve ever gotten to meaningful deployment with an actual user. We ran an experiment with a design partner to see if we could design a long-running agent to carry out the end-to-end process of manufacturing a product from a design idea. At this point, we had exhausted our personal network, and it felt like we were back at square one of identifying a real hair-on-fire problem. We took an experimental approach by trying to recreate Anthropic’s Project Vend (designing an agent system to carry out a long-running task) in hopes that it would uncover a unique insight for us.

Things worked well for a few days. Our agent, Coco, was up and running surprisingly effectively - it shortlisted a pretty high-quality list of vendors, negotiated costs below our budget, and even scheduled introductions for our user. We documented everything, spun up a pretty landing page, and even got a few investor inbounds. While these can all be interpreted as positive indicators, the massive elephant in the room was still that we had zero customer interest (albeit at this point, we were more focused on experimentation and discovery than distribution).

Once again, we had solved a problem, but not one that was a particularly high priority for our target users.

What Changes Going Forward

Focus on the problem, not the solution

I came into this fresh off building and scaling a brand myself. I was excited by the opportunity I saw to apply technology in a relatively low-tech operating space. I made the mistake of trying to force a solution (AI) into problems that (1) may not have been real problems, and (2) didn’t make sense to address with my solution.

Earned insight

Every step of talking to people is hard. First, you need to figure out who to talk to, convince them to talk to you, and have something compelling to say so that it’s worthwhile for them to keep talking to you. I took pretty conservative routes of meeting people: I got introductions to junior-level people through friends and sent out a handful of cold messages across email, LinkedIn, and social media. Going forward, I should be more targeted in my approach (speaking to senior operators), more bold in my method (going to industry events and following up more often), and firm in my value-add (I’m not just ~exploring ideas~ and ~learning about their business~, I want to solve real problems). I could also try to increase my personal leverage by working with other builders or seasoned industry veterans - this is TBD.

Validate → Distribution → Build

Over the past few months, I’ve pretty much taught myself how to code. Building an MVP is no longer a technical barrier for the type of problem I’m trying to solve. However, the greatest pattern I’ve noticed over the last few months is that I get stuck at validation and distribution. I’m not pushing for firm answers on the usefulness of my ideas, and I get scared to put my work to the test, hence the endless procrastination of building instead of distributing. Before investing any time in building, I need to validate some form of traction or user demand.

Final Words

I had hypotheses, put them to the test, and got answers. My ideas may not have worked so far, and I’m glad I got that confirmation. However, failure today does not mean failure indefinitely. I’ve iterated fast, but I haven’t iterated for long. I knew this would be a difficult pursuit when I started, and I may have under-indexed on how much of the difficulty would be figuring out how to be a good founder, not just execute. When I started my last job in investment banking, it took me about 6 months to figure out how to be a good analyst. While the benchmarks of a good founder are far more ambiguous, I trust that time and persistence will get me there.

There are things making this harder than just skill issues. I’m building in a space where relationships, network, and trust is the basis of any distribution strategy and as an ex-banker with very little consumer industry experience, I don’t have that yet. My next steps will focus on building leverage. How can I position myself to be exposed to the ins and outs of the industry? Who could I collaborate with and learn from? How do I fill in the gaps of the network and operating knowledge I don’t have yet?

Until then, we keep it pushing.

Side Notes of Doubt

A few thoughts that still float around in my head as I get up to try this again:

• Making fashion a profitable pursuit would involve fixing unit economics. Apparel companies suffer from terrible working capital cycles, which is an inventory problem fuelled by the MOQ bottleneck. Fixing apparel unit economies would involve rewiring the entire supply chain - is this a tar pit problem?

• Is the TAM too small? While apparel is a $1.8T industry, only 9% of brands cross the $1M revenue mark. Established enterprises don’t struggle on a unit economics level as they’ve surpassed the scale needed to achieve profitability. Small brands that struggle with this have very few resources to allocate, and the willingness to pay is extremely low.

• Production bottlenecks stem from manufacturers, and they need to be factored into the equation to fix them. However, vendors lack transparency, are located in all parts of the world with different languages and cultural backgrounds, can be slow to adopt new software, and need to have aligned incentives to change. This becomes a two-sided problem.

The biggest challenge I faced when I started building my startup was feeling stuck because I wasn’t technical. I soon realized that building a product was the last step I needed to take. Everything that happens before investing time, money, and energy into a product is what actually makes or breaks your business. I went through way too many cycles of making things no one actually wanted that could have been avoided if I changed my order of operations. There’s nothing wrong with building, but if you go through the cycle of building to no avail too many times, the risk of burning yourself out is very high.

2. The only barrier is yourself

Nothing is unlearnable, and if it is, you’re not the right founder to do it. This goes back to me feeling bottlenecked by my lack of technical ability. I thought that if I couldn’t build the product, I couldn’t build the company, but that was far from true. It took me messing up to realize this. I made myself too open to working with anyone who could code, treating them like a crutch, which ended up holding me back because I didn’t feel like I needed to figure it out myself. In an ideal world, you have a co-founder who complements your gaps and can be your crutch in an effective way, but in the absence of something that works, you only have yourself to rely on. It’s easier than ever to learn a new skill now. Just because you can’t do it today doesn’t mean you won’t be able to in a month. When I realized working with no one was better than forcing it with the wrong person, I hit a mental unlock and finally learned how to code to the extent I needed.

3. If you think you’re trying hard, try 10x harder

Not all work is productive work, and it’s easy to trick yourself into thinking that being busy means generating useful outputs. More often than not, the busy work is procrastination from doing something that is truly difficult or scary (like distribution). If you’re not pushing boundaries and you don’t feel uncomfortable, you could probably do 10x more.

4. Don’t be so nice

Coming from a rather stringent corporate environment, I wanted to create a work environment that was open and supportive. I didn’t execute this well since leaving things excessively open-ended made me come off more as uncertain than inviting. The amazing thing about working with others is that you get to leverage their minds, but the difficult part is that you need to lay down the rails for productive collaboration, and sometimes that means not being so nice.

5. Remember that you have options

The biggest thing I tell people who ask me for recruiting advice is that an interview should be as much about you sussing out your potential employer as it is about them doing it to you. I think it’s the same in the startup world. More often than not, I feel like a beggar, and we all know the saying beggars can’t be choosers. It’s exciting when someone is interested in your product or working with you when you’ve been facing months of no‘s. But startups are about people - finding the right cofounder, team, and even the right customer. Of course, having the right person is the best-case scenario, but having nobody is better than having the wrong person. Remember that there is no scarcity of opportunity; you can close doors because new ones will always open.

6. Equal skin in the game

Most of the time spent in the early days of a startup is on ideating and solving problems. You’ll get stuck in pits, you’ll be told no more times than you can count, and there will be times when you question why you do this at all. In those moments, if you are working with others, it is more important than anything else that they be on the same page as you in terms of level of commitment and what is at stake for them. I spent a lot of time working with people I didn’t exactly align with (people who had full-time jobs). It’s easy to harbor issues when people aren’t on the same page or stage of their personal journey into building a startup.

7. Time management is everything

When you have full control over your calendar, it’s easy to get lost without structure. This is something I’m still figuring out. Building a business demands breadth - from research to coding to marketing. I’m still trying to figure out how to structure my days so that I can produce quality outputs, get everything done, and minimize time wasted in context switching and overthinking.

8. You need to be laser-focused

Noise is everywhere. It’s on your timelines, it comes from your friends and family, and even from your team at times. Every company needs a north star to point to when the noise gets loud - if all else is questionable, what is the one thing that still holds true, which serves as the reason to keep going?

9. Mind over matter

Doing anything difficult is purely mental. If you’re entering uncharted territory, you’ll inevitably feel fear. Before doing this, I was never comfortable with the feeling of failure. Yet I’ve been rejected and fallen on my ass so many times now, I’ve learned to become unfazed. The harder part is figuring out when to keep pushing or pivot - the answer isn’t always clear, and I’m still trying to crack it.

10. Ask for help

Chasing your dreams can get lonely because no one cares about them as strongly as you do. It’s easy to retreat into your head when things get hard and to spend hours thinking in circles and internalizing things. That’s when it’s important to talk to people. Vent to friends, rely on mentors, reach out to people you don’t even think would respond to you, because you never know.

While the industry has sold investors and consumers alike on promises of a perfect shopping experience and automated business workflows, the elephant in the room still stands:

Who is using all this, and what difference does it make?

Fashion is a particularly difficult vertical to build a business in, both as a brand and as a company that aims to service brands. On the consumer side, buyers are fickle and incentivized by price more than anything. Trying to build a platform for consumers requires either fundamentally changing consumer behaviors or a race to the bottom on pricing with your competitors.

On the business side, brands can be segmented into two broad categories: large enterprises ($50mm+ annual revenue) and SMBs (<$50mm annual revenue). Large enterprises are characterized by operational systematization and established distribution pipelines. These are companies that have transitioned from founder-led operations into scalable systems supported by infrastructure like ERPs, CRMs, and PLMs. They’ve established a playbook for everything from planning new product launches to managing inventory and launching new marketing campaigns. As systems become more complex and the pool of stakeholders grows, the stakes for these businesses rise, and focus shifts from rapid scaling to risk-adjusted scaling. Large enterprises can no longer afford to migrate their systems every time a newer and better one is released. With deeply embedded systems for every organizational function, they’ve already cracked the first 1-90% of the work and are optimizing for the final 10%. This is why as companies grow, functions become increasingly fragmented. For example, Nike isn’t focused on the 10x improvement of going from no inventory system to an inventory system; they are optimizing for a marginal improvement by investing in 100% inventory transparency with a blockchain system.

On the flip side, SMBs are much more flexible in adopting new systems. They have fewer bureaucratic layers, can tolerate more risk, and benefit from zero to one improvements, as many have not systematized operations yet, and founders are already spread thin with problems that carry urgency. Y Combinator famously advises founders to sell to startups for these exact reasons. However, the fashion industry is extremely different from the tech industry (beyond how people dress). Six months of building in the fashion tech space has taught me that:

1. Fashion is a relationship-driven industry. You get into doors and conversations by either being someone or knowing someone.

2. Fashion is a bad business. Apparel margins are thin, with most companies netting 10-12% on the bottom line, and only 9% of brands ever surpass $1m in revenue. It’s a highly saturated market where working capital demands are high and profitability almost always only happens at scale.

3. For an industry rooted in creativity and product innovation, the way fashion businesses are run is extremely traditional. Systems are surprisingly manual (even in large enterprises), and a mix of having little budget to spare and a low-tech culture holds the industry back from adoption.

As all markets go through boom and bust cycles, my prediction for the state of fashion tech in 2026 is that this is the year the dust settles. Headline fundraising rounds and viral products put fashion tech on many people’s consciousness in 2024 and 2025, but behind the smoke and mirrors lies the question of actualized ROI and longevity. Price comparison platforms need to prove that they can create defensible moats against incumbents like Google Shopping and ChatGPT, and personalization platforms need to prove that they can fundamentally change consumer shopping behaviours over the long-run. Enterprise tools need to prove tangible attribution to top and/or bottom line improvements beyond pilot and consulting projects to permanently embed themselves in the stack.

Fashion tech is only as strong as the industry it serves, and with fashion being a $1.8tn industry projected to grow at an 8% 10-year CAGR, there is no shortage of opportunity. 2026 represents an important inflection point for fashion tech as we’re now two years into the boom and results are starting to show.

Noted in Ivan Zhao (founder of Notion)’s 2025 closing article on the evolution of technology, technology of the future often resembles technology of the past until it hits that inflection point. While AI shoppers resemble the browser extensions we’ve always had, and AI ERPs still look like dashboards and tables, we will inevitably reach a state where innovation begins to challenge the interfaces we’ve all become used to. Similar to how the Spotify of today barely resembles the music industry of the past, I’m optimistic that highly guarded legacy industries like fashion will encounter great disruption as well. When it does, it will probably catch us all off guard as it will likely not resemble anything we’ve seen so far.

What is Product Lifecycle Management

The shirt you are currently wearing was likely produced in a factory. So were the shoes you walked here in, and the bag you pulled your phone or laptop out of to read this. Each of these garments was designed, prototyped, and produced in a process that involves over 400 decision points before they hit the shelves for you to buy.

The process of creating a physical product (we’ll focus on apparel) is extremely complex, and with 400 decision points per product, it requires a system of management and record-keeping, as no one can keep track of everything in their head.

Product lifecycle management (PLM) is the process of managing the journey of a product from the inception of the idea to its final disposal. A PLM system is meant to function as a hub of all the information related to that product: what stage of life it is at, who worked on it, what materials it uses, the costs associated with it, and more. The fundamental value proposition of a PLM system is to provide an organization with a single source of truth to help it manage highly complex processes that overwhelm people’s cognitive load.

Illustrative Decisions Made at a Small Scale Apparel Company

• Decision points in producing a single product: 400

• Number of collections released each year: 4

• Number of SKUs within a collection: 25

• Product decisions made in a year: 40,000

PLM Systems in Practice

Anyone who produces a physical product would have some sort of PLM system in place. At smaller organizations that manage fewer products, this could be as simple as spreadsheets tracking each product, its status, and specifications. At larger organizations where multiple products are produced in a single collection, and multiple collections are produced in tandem, software is often used to automate, organize, and analyze high volumes of information.

Regardless of its form, PLM systems serve as a system of record for production and are critical in ensuring products are made on time, within cost restraints, and meet quality and regulatory standards. In short, PLM systems are mission-critical to any company that makes physical goods.

The Issue with Modern PLM Applications

Whether a company uses manual spreadsheets or software like Centric or Lectra, PLMs still miss the mark on their fundamental purpose: to provide users with an accurate single source of truth. The problems could be summarized as:

1. Data entry lag: As multiple parties are typically involved in the production process (production coordinator, vendor/manufacturer, designer), process tracking becomes a game of chasing for the latest update. PLM applications function as a repository that gets updated by users in retrospect, serving as an extra step after the real work is done. The nature of this system creates a permanent lag between the record and reality. From interviewing production teams across F500 companies and independent labels, we’ve learned that each person involved in the production process not only has a central PLM system (shared spreadsheet or software), but also a personal tracker. This derivative tracker documents the secondary tasks involved with updating the system, such as who they need to chase for an update and what fields they need to update in the central system.

2. Poor data quality from manual entry: Both spreadsheets and PLM software require hefty manual data entry from users. One production manager at an independent label we spoke to described how costly the consequences of misentry could be. An error as small as entering $4.57 instead of $4.75 can cause a production run to miss profit targets, as margins are slim and errors compound when production scales. The average human error rate on single-key entry is 1-4%; however, errors compound as raw data is layered into complex calculations and manipulated into semantic data. Given that 400 decisions are made per garment, a 1% error rate implies that every garment holds 4 data entry errors. Compounded across multiple products and multiple collections, modern PLM systems inherently carry a costly margin of error that is completely preventable.

3. Lack of trust: Due to the lag and errors, production managers often do not trust organization-level PLMs, which is why each person we spoke to simultaneously kept a personal version of a tracker. This is particularly challenging at organizations using enterprise software, as our research found that production managers only utilize 30% of the tooling available on these platforms, with software reps offering dedicated training sessions in efforts to increase usage. When systems fail to fulfill their first principle purpose, people invent shadow solutions that further perpetuate information siloing and the likelihood of data errors.

Our Vision

An ideal PLM system functions as the operating system for producing a physical product. Unlike a database, which is a passive store of information that relies on humans to update records after the fact, an operating system is an active infrastructure that integrates directly with source information and manages the flow of work. Today, fashion and CPG brands of all sizes have developed robust front-end (customer-facing) operations using providers like Shopify that offer accessible tooling. Conversely, back-end (creating the products) operations have yet to be systematized, with large enterprises using feature-bloated legacy software, and SMBs having no productized solutions and running on spreadsheets.

Inseam is building a future where production is a continuous automated flow and a single person can oversee hundreds of SKUs rather than requiring a multi-person team to actively push products forward. Organizations will have full visibility into their products with virtually no time lag or input errors. This is imperative to companies working with overseas manufacturers, where time differences and language barriers reduce speed to market.

Our first step is to create a self-sustaining system of record. AI removes the software constraint that requires users to translate data from unstructured (such as an email from a manufacturer) into defined or semantic data (i.e., order value). This means Inseam integrates directly with a brand’s communication sources, such as email and WhatsApp, to capture and translate information autonomously. This drastically reduces the administrative burden production teams face, as our user interviews found that people spend as much as 3 hours a day aligning across multiple communication threads to determine the production stage and to-dos related to a product.

Establishing an autonomous system of record would create substantial operating efficiencies, as automated systems have been found to reduce operating costs by 20% and manual entry errors by 25%. This is particularly impactful for SMBs who are currently juggling multiple spreadsheets and scattered tooling with no scalable systems in place.

However, this is only the beginning. A self-keeping system of record acts as the foundation for capturing the data needed to build high-performing execution agents, which can handle tasks such as vendor communications and resolving production errors. As human labor is allocated away from manual entry and into high-volume decision-making roles (remember, 400 decisions are made from inception to completion), we will compile comprehensive context graphs that inform agents of what information is needed, where to get it, and the decision-making rationale set by human precedents to fully take the wheel at scale.

We’ve already started working with pilot partners and think a future where the execution from idea to product collapses into a seamless autonomous system is much nearer than imagined.

About Inseam

Inseam was developed from our experience as brand founders and engineers. We provide brands across fashion and CPG with a centralized operating system that supports scaling while minimizing costs.

We’re currently helping brands automate the production process to achieve faster speed to market, reduce error rates and sample rounds, and improve operating margins.

Our pilot is live, and we are actively implementing our platform with a small group of brands. If you are interested in trying Inseam or have any questions, please feel free to reach out to jess@inseam.ai.

Fashion is an industry that thrives on the innovation of small players. In this industry, product innovation comes from new styles which almost always proliferate from the bottom up: independent designers push boundaries in creating experimental styles which gradually get adopted by the mass market.

While design innovation is imperative to pushing the industry forward, it is operational innovation that has propelled fashion to be the $3 trillion industry it is today. Design innovation cannot exist without the commercial viability of the underlying organization to support it, and much of this is achieved through advancements in technology (particularly integrating IT in operations).

Fashion is often compared to restaurants as it is a notoriously difficult business to take profits in. Headlines are perpetually embellished with brand and retailer defaults and labels trading creative directors like athletes every time it financially underperforms. Think of SSENSE filing for bankruptcy protection, Saks’s endless restructuring which ultimately led to a Chapter 11 filing, Y/Project shutting down after 14 years of business, and Versace swapping out a creative director (Dario Vitale) after only one season.

The high fixed costs and complex logistics of these businesses give larger players a natural advantage as they have the resources and scale to optimize operations. Enterprises have teams dedicated to optimizing everything from their material sourcing to production costs down to the cents, and systems in place to capture every website visit and click, then tailor outreach to that specific prospect. Nike alone has a headcount of ~75,000 people (though about 50% are estimated to work in retail), and Lululemon even has a dedicated internal calendar team. Contrast this to smaller brands that may be driving product innovation, but barely have a system of record for keeping track of all their SKUs.

In a business where margins are slim by nature, capturing an extra 0.1% can be the difference between a profit and a loss. Achieving this is nearly impossible for emerging brands, as high overhead and a breadth of operational functions often leave operators fighting to keep things running, making optimization and systematization a secondary priority.

Fashion’s Technorenaissance

2025 showed us that AI can (and will) create real and meaningful returns in businesses. Everything from generative marketing materials to customer support agents proved that our current team structures, day-to-day tasks, and economic realities will be vastly different in the next 5 years. GenAI alone is projected to create $150-$275 billion of operating profit in fashion, according to McKinsey. Enterprises are aware of this, with companies like Nike, Zara, Louis Vuitton, H&M, and more investing ~$1.5 billion in AI adoption projects.

However, not all fashion businesses have dedicated technical talent to do this; in fact, many don’t even have systematized operations to begin with. While we may envision a world where a fashion founder can automate inventory replenishment, use AI to write advertising copy, and have an agent process customer orders and returns, the reality is far from it. Most sub-10-person teams in a mature, non-tech-centric industry like apparel are more focused on keeping afloat with day-to-day operations than investing in high-tech solutions.

Bridging the Gap

I’ve spent the last year building a fashion brand, which is currently one of Toronto’s fastest-growing independent ready-to-wear labels. Out of manufacturer emails and Excel sheets came Inseam, which harbors our mission to achieve the operational efficiency that enterprises are striving for as an SMB. I’m entirely confident that this is possible, as we’ve seen it be done in other industries (particularly software development) and now have access to an unprecedented amount of tools to leverage.

Through speaking to over 50 brand operators across functions like design, production, and merchandising, we’ve realized that despite the amount of tooling that is now accessible, SMBs still struggle to implement technology (specifically software) in their processes.

I hypothesize that this is a two-pronged effect:

Brand Operators

• The overlap between highly technical people and operators in apparel (broadly also applicable to CPG and physical products businesses) is limited. IT investments, therefore, require operators to invest already scarce resources into researching and almost always need to be outsourced at an SMB level. 15% of SMBs reported keeping up with fast-changing technology as a top challenge in 2025.

• IT does not sit as high on an emerging brand’s list of priorities as it should. Brands are more focused on speed to market and see building scaling systems as more of a cost than an investment in the early stages. With the wide array of organizational functions required to bring a physical product to market, problems are often addressed as they occur rather than systematically prevented. This in turn causes 90% of AI initiatives to fail to scale beyond the pilot phase as organizations have insufficient underlying technology and data to support the programs.

Technologists

• Likewise, there is only a slim overlap between technologists who have the skills and knowledge to architect solutions to operational challenges and those who have a deep understanding of the fashion industry. This makes it challenging to create tools that are actually value-generating. The limited collaboration between technologists and operators makes it difficult to establish the fast-paced iteration and feedback loops necessary for a technical product to achieve market fit.

The SMB Opportunity

Part of the challenge with working with SMBs is that they are highly unstructured and often do not have any formalized systems in place. However, this is also the beauty of SMBs, as this creates an opportunity to build new systems from scratch, in a way that is compatible and scalable with where technology is headed.

Last year, Business of Fashion declared that we are in the midst of a fashion tech boom. A quick exercise in tracking headlines and where VC dollars are going will reveal that most of the companies fueling this boom are consumer-facing (B2C virtual try-on, digital closet) or meta-consumer-facing (B2B2C generative marketing assets, online shopping plug-ins). I think the real opportunity is in the back-end operations of these businesses, as systematizing manual processes and unstructured data is where value is generated by multiples rather than percentage points.

This is because the process of producing and selling a physical product is incredibly complex, with a single product requiring 400+ decision points from design to final mass production (now imagine this scaled across a multi-product collection). Each decision point represents an opportunity for errors which can carry costs that make the difference in profitability. Systematizing these processes early on sets a brand up for scaling and avoids costly migration later on. This also acts as the foundation for embedding automation into repetitive workflows, as any executional function, AI or not, requires some kind of decision trace to follow. This sets us up for a world where teams can multiply output per a fixed headcount and compound the benefits of future technological advancements as they have invested in a compatible infrastructure. By applying technology to the back-end operations, brands make themselves competitive and scalable from the inside.

The Future Is Ours

If apparel SMBs don’t adopt operational innovations, the economic gap with enterprises will continue to widen. Fashion is a mature industry that will benefit hugely from operational innovation due to the relatively lower regulatory barriers (compared to fintech and healthcare), and SMBs have the benefit of developing AI-native systems from scratch. Its creative and executional nature makes it the perfect place for AI to augment manual and scaling processes while freeing up human talent to do more strategic and creative work. SMBs, like startups, are amazing vehicles for quick iterations and high-impact solutions, yet the flywheel has not taken off. To position the industry to fully take advantage of the next 5 years of technological revolution, we need to establish a more collaborative environment between operators and technologists.

Pilots are happening and an increasing number of SMBs are either thinking about embedding technology in the foundation of their operations or actively doing it. 39% of SMBs already have ERPs for some functions and plan to expand across other departments. However, this is still concentrated at larger-scale businesses since these companies have defined departments to begin with. These are companies that are solving for an extra 0.1% margin boost, not the ones that are balancing Excel sheets, Notion pages, and loose PDFs just to remain functional. 2026 is an inflection point for SMBs who have the benefit of a lean and highly iterative structure to leverage technology to its fullest capacity. While fashion tech has boomed over the past year, the biggest opportunity remains overlooked, as it’s in expanding the size of the industry through setting SMBs up for scaling. We now have the potential for a single-founder company to operate at the same level of systems and intelligence as a 75,000-person enterprise, and how fashion embraces tech and vice versa will determine if the potential becomes reality.

About Inseam

Inseam was developed from our experience as brand founders and engineers. We provide brands across fashion and CPG with a centralized operating system that supports scaling while minimizing costs.

We’re currently helping brands automate record keeping, such as tracking SKUs and production, which acts as the foundation for operational visibility and additional AI-integrations.

We are continuously working with a small group of design partners - please feel free to reach out to jess@inseam.ai to get in touch!

I spent my teens attending conferences and speech tournaments. I ended up going to a university where perhaps my parents would be proud to have that particular diploma at home. All this to shovel my way through a room of other sophomores, eager to get the chance to stand in a circle with 8 others and ask a junior banker about their 1 year of work experience. I thought that if I did all of these things, my chances of working at one of these prestigious firms in a shiny tower would actually be non-zero.

This archetype of person is thoroughly documented. You can count on any innovation thought leader personality to have dunked on the business degree, consulting/finance, settles-for-comfort type. But who can blame them for choosing comfort?

Thinking back on why I left, it wasn’t the physical strain of 100-hour weeks or the mental strain of constant negative feedback and no room for a personal life that bothered me. It was the lack of autonomy, the minimal agency allowed in that working structure, and simply the fact that I felt like I was wasting away in the shiny tower I once wanted to be in. I forgot I could create things, solve meaningful problems, and have strong opinions (I could feel my brain atrophy). While I have many friends who continue to work in finance and respect all of them sincerely, I could never shake the feeling that I was merely skimming off the top of what value was actually being contributed to the world.

Some say you just have to power through, you’re a junior after all. But did I really want to spend my entire 20s doing something that didn’t feel meaningful and was not mine? The concept of working for meaning is a relatively modern (and privileged) idea, but even from a physical perspective, was it worth the sacrifice of not having time for friends, family, and taking care of myself in my 20s to one day get to be in some middle manager role where I could let my brain atrophy at my own will? Maybe that’s harsh, but the thought of spending the majority of my time doing something where I most look forward to it ending is not something I wanted. That’s why I decided to leave, thus beginning my descend from the local maxima.

I didn’t leave with anything specific to pursue. I had a few side projects I’ve been working on and dove head-first into an exploratory phase. I started creating all the things I had always thought about making - clothes, videos, writing, even software (yes, I vibe coded). I learned more in a month than I did in a year at my corporate job, and met people I used to think of as unreachable. I thought I saw a light in the tunnel as for a moment, a project I was working on seemed to finally be working, until it didn’t.

They say your first heartbreak always hurts the most, and it’s cutting out to be the same with any entrepreneurial and personal endeavors. However, appearances can be deceiving and discouragement is far from giving up. I abandoned my local maxima in pursuit of my global maxima, and I’m just beginning to learn how much it sucks to come down. I didn’t expect everything to be fixed once I started pursuing things that have meaning to me, and this serves as a reminder that we can’t go up without coming down. All we can do is brace ourselves for the descend and keep going.

Anyways, this is a long way to say the thing I was working on doesn’t seem to be working and I’m once again reconfiguring my life in real time.

Six months ago in February 2025, we launched our first collection under SERENA LI. We followed a similar playbook as we did this time — building hype on social media and hosting a launch event. While the collection was well received, sales last season fell short of our expectations. Despite all the positive signals of a growing online following and supportive audience feedback, we realized there was still a gap between people liking our clothes and actually wanting to wear it. These were good realizations as it created the foundation of experimentation for the most recent collection we released. I’m sharing our process of developing and testing hypotheses as we think this is a crucial step in building and iterating on any organization, business, or project. As much as we love to share the behind-the-scenes of our creative process, we believe it is just as important (and interesting) to discuss the operations side of building a fashion brand.

What We Cooked On

Pricing

Hypothesis: Consumers are turned away by garments that are too expensive

This is the elephant in the room. Our garments are expensive and candidly that is due to being bottlenecked by expensive production. At our small scale, we are not yet able to meet factory minimums, meaning we source our production locally and pay sample prices for pieces rather than bulk order prices. Fashion is an economies of scale business, and until we are able to do this at scale, we need to work pricing around our costs.

We also acknowledge that our early adopters are taking a risk on us. Not many people are willing to spend money on a new designer, especially in a category with established alternatives.

Experiment: Decrease retail prices

Our last outerwear piece was priced at $600, this time we priced it at $400. By doing this, we aim to find an equilibrium that balances what consumers are willing to spend with what is reflective of our costs, work, and premium brand positioning.

Outcome: A greater proportion of our sales came from garments rather than accessories compared to last season

Last time, nearly 60% of our sales came from accessories and lower-entry products like t-shirts. This time, the majority of our sales came from garments. Serena always says what she loves most about doing this is seeing people wear her designs. This doesn’t mean we’re going Costco mode going forward, but this experiment helped us get closer to a sweet spot within the premium range that our brand positions itself in.

Design

Hypothesis: Our pieces are too “out there” and unwearable to most consumers

There’s a difference between FASHION and fashion. Collections are usually structured as pyramids, with a broad base of more commercial and casually wearable products, and a narrow top of highly artistic pieces. Our last collection was not shaped like this as only 3 of 8 garments fell into the commercial category.

Experiment: Offer simplified garments that have elements of higher-level design, but are tailored for casual wear

This collection was rooted in functionality. The materiality of each piece is a cotton-nylon blend, which is commonly used in lifestyle garments, unlike the PVC garments in our last collection. Our design process was rooted in engineering as we deconstructed and tested functional-wear, which led us to include large pockets, use water-resistant fabrics, and include features like packability and a convertible jacket hood.

Outcome: We sold more pieces to a target audience who weren’t just fashion enthusiasts, but also wanted the garments for a specific purpose

Since the collection was inspired by cycling, we targeted our event towards the biking community and encouraged people to bring their bicycles. Upon speaking to a few people, we received a lot of questions about the functionality of the pieces. People liked the pocket size and placements as it made it convenient to not need to bring a bag on rides. They spoke about biking in the rain and how water-resistance would be helpful for that. A model from our campaign even mentioned that the pocket in the hat made it convenient for her to store her cochlear implant while she wore a helmet on rides.

Display

Hypothesis: People didn’t buy the pieces because they were afraid of touching the display and trying things on

During our last pop up, we displayed garments on crowded racks which included every design and size we had available. We thought this would mimic the experience of shopping in a store and encourage people to try things on. However, we realized this made it difficult to see the products as people had to sift through the racks themselves. In a crowded space of over 300 people, this was extremely difficult.

Experiment: Create a more interactive display that let people see the garments and get up close to touch and try things on

From our high school summer jobs at Aritzia, we both knew that getting a garment on a person’s body had the highest conversion rate, so being able to interact with the products is important. We created an exhibit-style display that had the garments individually hanging eye-level so that people could walk through and interact with every piece. We also printed information cards which explained the pieces and their functionality. Having size options on a rack off to the side with a mirror made it clear what was for try-on.

Outcome: People played dress up and we made multi-item sales

Seeing people take mirror selfies in the pieces was my favorite part. Circling back to the first point, we don’t expect everyone to buy the garments. Sales help keep us going but seeing people engage in their love of fashion is what brings us all together, and that means styling a garment to our individual tastes.

Giving people the chance to wear the pieces helps them confirm fit and imagine how they might style something, which is the practical side of garment sales. We also believe style is deeply personal, and the more you get to interact with an item, the harder you fall in love with it. When people got to style the pieces with other garments on the racks or even with the clothes they had on, it bridged the gap between imagining how this piece might fit into their wardrobe to actually being able to test it.

Conclusion

Before anyone says anything about making art for art’s sake, I’ll briefly explain why we care about sales. We believe that designers create things to be shared with the world. That means making clothes that are worn and loved rather than sealed away on a studio display. Creating something that both aligns with our vision and is something that people actually want to wear is important to us. On a practical note, fashion is an expensive business. Besides the cost of actually producing each garment, brands must shoot campaigns and e-commerce images, as well as cover overhead like labor and studio space. For our pop-up, we wanted to create an event that gave various communities the opportunity to meet and connect — unfortunately, our expenses rack up to more than just a monthly Cursor bill.

We think it’s important for creatives to be able to sustain their work. We’re far from having it figured out, but each experiment gets us slightly closer. As we systematize our process and design technology to gain greater leverage, we hope to share this and help other product-focused entrepreneurs as well. That’s why Serena and I have been working on a new project called Inseam alongside a larger technical team (stay posted)!

That being said, we are incredibly blessed to get to do this. People are often surprised when we reveal that the team is just two people (my co-founder/designer Serena and I), but that doesn’t speak to all the help and support we’re so lucky to have. We want to extend the biggest thank you to the production team behind our visuals and campaign, our seamstress for making the clothes real, the New Systems team for making the event happen, Harry from Giraffe who decked out our lights, and Sammy and Kristyn for dealing with our crash outs and playing bartender, sales associate, clean-up crew, and the many other hats that we can’t fill ourselves. As a final note, we want to extend the biggest thank you to everyone who came by to support our event and have been following along our experimentation. We hope you loved it and we can’t wait to see the pieces in the wild.

Lastly, big credit to Serena — she’s truly brilliant and none of this happens without her.

I officially left my job two months ago to pursue something that very much does not look like a traditional job. When people asked what my plans were, I played naive and danced around the question with “I don’t know, I’m still figuring it out,” even though I knew full and well that I wanted to pursue a startup. I’m trying to get better at this and break my habit of diluting my pursuits to make them more palatable to others.

In this process, I’ve realized that adopting the identity of someone who sets down an unconventional path is just as much of a learning experience as it is actually going down the path itself. With every customer conversation, I’m not only learning about them, I’m also learning about myself as I never thought a statement along the lines of “I have this idea and I’m dedicating myself to seeing it through,” would feel so weird to say out loud.

I won’t pretend like I’m not afraid, but I’m also not going to hide the fact that I have never felt more conviction about anything. So while I haul myself up every morning to face the scary unknown abyss that lies ahead, here are the things I tell myself to keep me going:

1. Have an opinion

You don’t need to be a contrarian, and it’s okay if you need more information. There’s a huge advantage to having an opinion at all when a lot of people don’t, and often the only barrier between having one and not is doing a bit of thinking. That doesn’t mean being rash or being stubborn, it just means to think. Pick a starting point and ask yourself “What might this mean for X”. If you don’t have thoughts then ask questions.

2. People value others who have something interesting to say

Bring something to the table. If having an opinion gets you into the conversation, having perspective, ideas, and outputs gets people to listen.

3. Get yourself into uncomfortable situations

If I had a dollar for every time I was scared I would never need to do anything scary again. Ironically, I’ve never left any of these scary conversations or presentations feeling regretful, in fact, I usually leave feeling great. As much as I curse my past self for getting us into these situations, I’m always thankful that one dimension of my consciousness knows that fear is only momentary.

4. Exist in the world

It’s easy to live in private mode, but I think staying lowkey is overrated. Your ideas will never materialize if no one knows you exist or of the thoughts that you have. If anyone thinks that’s cringe, I invite you to put that idea into the world and say it out loud too.

5. Don’t just use the internet, contribute to it

The internet gods bless you when you give it an offering. The people you want to connect with are in much closer reach than you think, but you need to give them a compelling reason to respond. This all circles to the above: have something to say and actually say it. The internet is only a platform if you use it.

6. Commit

The bar is surprisingly low. Most people talk about wanting to do things more than actually doing them. There’s a huge advantage to just trying. It doesn’t even need to be good per se — everything is relative.

7. But try so hard

The only person that gets hurt from a half-assed attempt at anything is yourself. Your brain can’t differentiate between a 50% effort vs. a 100% effort, it just thinks you’re trying. There’s no shame in wanting to win.

8. Everything is interesting

The way you can always have something to say is by always listening. Gaslight yourself into caring even if you don’t because the payoff of unexpected knowledge is immeasurable. The best studying advice I ever got was to imagine that the topic you’re learning about is the most interesting thing ever, even if it’s not. You do so much better when you care.

9. Learn from anyone

Everyone has lived a different life, so everyone has a different perspective to share. Learning from someone doesn’t mean you have to be like them. You can learn a lot about what kind of person you don’t want to be too.

10. You are not a piece of shit <3

Or maybe I am, idk. I always thought I needed to justify myself, but we are all trying our best:)

If you’ve ever come across anything I’ve written or said, you might know that fashion and tech take up 90% of my mental load. I’ll be sharing in greater detail going forward and excited for the road ahead!

I recently came across a post on X suggesting the creative class is being turned entirely into therapists and entrepreneurs. This is funny timing considering the conversations around how people, notably Gen Z graduating into a diabolical job market, are being pushed to pursue their dreams whether they like it or not. If traditional jobs aren’t available, chasing your passions might carry just as much risk and being unemployed.

The term “Creative Class” wasn’t on my radar until that post, and digging into it, I learned that it came from Richard Florida’s 2002 book, The Rise of the Creative Class. He raises the idea that jobs centered around creative skills are a significant driver of economic prosperity. This includes the Super Creative Core, people like artists, designers, writers, and scientists, and the Creative Professional, mostly knowledge workers whose jobs involve some kind of problem solving and ideation. In my opinion, this is a generous definition of the word ‘creative’.

I’m the last person to gatekeep what is and isn’t creative — call my excel formula of 12 nested IFS a “creative solution” and I’ll take it. However, as ambiguous as creativity is, there is something instinctively off about categorizing engineers and fashion designers in the same bucket. Apart from the differences in their work, the chasm between the two is evident in the way people respond to “I’m an artist” and “I’m an engineer”, which may be a cultural reflection of the economic situations of the respective professions.

When I see books with titles such as “The Art of ___”, it urges me to consider if the word “art” is being used in a definitional sense, as a means of expression, or if it’s (for a lack of a better word) just marketing.

Back in February, I wrote about the tech industry’s AI coding panic and sudden interest in having taste. “If AI can write our code, the only thing we have left is taste!” Suddenly, there was a shift in the sentiment towards creative pursuits. Fashion became less frivolous, UI/UX took center stage, and just look at the surge in Zora’s (a Web 3 platform helping creators monetize) popularity as more emphasis was placed on creativity as a differentiator. Even Y Combinator issued a call for startups requesting founders with design backgrounds. It seems like everyone and their dog started "building for designers, by designers”. Creative Professionals — the founders and engineers who dominated the economic gains of the creative class, were trying to converge with the Super Creative Core.

This would be a great thing for the Super Creative Core if it truly meant greater economic resources by virtue of technological or monetary empowerment, but at large this isn’t the case. Firstly, industries are relatively siloed, with many artistic creatives (designers, writers, musicians) expressing limited exposure and education of emerging technologies. Over the course of the past month, I spoke with fashion designers ranging from independents to large enterprises and have noticed a hesitation to adopt AI tools. Part of this comes from fear of replacement and threat to artistic integrity, and part of it is out of unfamiliarity and intimidation. Secondly, there is still an observable difference in the monetary benefits artistic creatives receive relative to professional creatives — for example, UI/UX designers comparatively earn less than software engineers do, despite now being told their “taste is the final moat”.

By no means am I saying the surge of emphasis on creativity is a bad thing, however, I would criticize that much of it feels…disingenuous. It makes sense for AI to reconfigure the way creatives operate like what Adobe and Figma had done, and it makes sense for the people leading these initiatives to have a genuine interest in designing good products. However, I don’t think it’s absurd to question whether the newest ~AI-native canvases and dashboards~ are built by people who actually have experience, engaged with, and surveyed who they’re building for. It’s worth questioning how many of the startups who put design at the forefront of their product experience actually have a founding UI/UX designer and whether their compensation reflects the company’s supposed values. An example of a company doing this well is Flora (an AI design canvas), whose founder actually has a background in creative fields.

We can all say we’re anti-slop, high-taste, design-valuing creatives, but without the supporting evidence, we’re doing nothing more than performatively reading The Creative Act.

Katalin Karikó and the Vaccine that Almost Wasn’t

In 2023, Katalin Karikó received the Nobel Prize in Physiology or Medicine for her research that led to the development of mRNA-based vaccines, the basis of both Pfizer and Moderna’s vaccines which helped contain the spread of Covid-19. While the research itself was complex and groundbreaking, much of the uphill battle came from Karikó’s difficulties in securing funding. Karikó began studying mRNA at UPenn in the 1980s, however much of the scientific community viewed mRNA as unstable and did not think it could be used in any treatments, making the research risky due to low potential returns. In 1995, she was demoted at UPenn, denied tenure, and rejected from numerous funding programs such as the NIH (National Institutes of Health). Grant organizations wanted proof that her hypothesis could be proven, however, this was not possible without conducting the research that she needed funding to do.

“Only a very small amount of protein could be produced because RNA is so quickly degraded, and everybody thought it would not be enough to treat any disease. In addition, we learned from work we did with Drew Weissman [a professor at the University of Pennsylvania], RNA was extremely inflammatory in human immune cells. So, in the ’90s, many left the field because they couldn’t get funding, or they had run into other problems.” (Karikó, 2022)

This circular problem appears in numerous power-asymmetric industries, ranging from fashion to venture capital. Resource-rich decision makers act as gatekeepers and aim to de-risk themselves by selecting projects with foreseeable success—something that is hard to achieve without having the resources to try.

When Department Stores Were Cool

Parallel Karikó’s story to the experience of numerous designers in the fashion industry. In my previous article ‘Why Now is Time for Systemic Change in the Fashion Industry’, I describe the dynamic during the early days of fashion’s commercialization. Department stores and media publications acted as incubators for designers by purchasing products from emerging designers and placing them in stores, thus providing financing for production runs and creating a distribution channel to consumers. This is explained in detail in the book ‘The Beautiful Fall’ by Alicia Drake, as she delves into case studies of Anne Klein, Stephen Burrows, Halston and Bill Blass—all of whom benefitted from large institutions’ willingness to take a risk on them.

We Live in a Catch-22

Between the US federal government’s proposed cut towards research funding and retail buyers showing increasing favor for brands with existing customers, it is evident that major institutions are solving for risk-adjusted returns rather than pure innovation-driven returns. There is a fine balance to be achieved since any organization needs to consider risk when planning for longevity, however, organizations that exist with the purpose of supporting innovation must consider the cost of de-risking, particularly in industries where few gatekeepers act as the primary incubators to progress. A healthy amount of de-risking is effective in self-preservation, however, optimizing for return-focused metrics can pose just as much of an existential threat as ineffective risk management.

Take Vogue for example—the publication recently faced backlash for featuring an AI generated Guess ad. Known for its role as an industry titan and tastemaker, readers felt an aversion to the AI generated image for a myriad of reasons such as “devaluing hard work and art of real people”. It is unclear why Vogue approved the ad—or if it was even within their control. However, a likely reason is due to its advertising relationships as advertising comprises a significant portion of revenue. While decisions like these reflect in financial gains, over the long run they have led to questions about the defensibility of Vogue’s position as fashion’s authority.

Beyond causing harm to the institutions themselves, applying return-focused optimization to innovation-driven organizations will inevitably lead to stifled discovery and slower advancements in the field. Exemplified in the book ‘Why Greatness Cannot Be Planned’, many of humanity's most profound breakthroughs were made by chance rather than creating a targeted goal; this includes the laser, the internet, computers, and evolution itself. The concept of novelty search details the importance of casting a wide net around discovery and prioritizing exploration over achieving defined metrics as it is often not possible to know the steps needed to achieve ambitious targets.

Currently, research grants—particularly from major institutions such as the NIH and NSF—are evaluated with a focus on likelihood of success, having preliminary data, defined objectives, and impact (or ROI). This creates barriers to scientists like Karikó, whose work was dismissed as fringe and nearly abandoned before it could prove its true return. In discovery-driven industries like science and art, optimizing for return metrics can be not only misdirected, but even arbitrary as attempting to define an inherently ambiguous endeavor applies unnecessary limitations that stifles true potential. Though it would be naive to ignore resource scarcity, it is important for institutions to acknowledge the power law distribution of the risks they take and how pursuing novelty can generate significant unquantifiable returns such as preserving Vogue’s reputation as fashion’s head tastemaker, or making scientific discoveries that can be applied to unexpected uses.

We’re De-risking Ourselves of Progress

I’m not a scientist or a cultural tastemaker—I just think a lot. I remember doing a case study back in school on a financially struggling airline. In an effort to reduce costs, it decided to retract the number of routes it offered—this ultimately weakened its positioning as it offered customers fewer options, a less robust network, and closed revenue streams. While each industry is nuanced in how it balances stakeholder interests, the underlying pattern I want to highlight is how easily innovation and its non-immediate benefits are neglected in de-risking processes.

I can’t wrap this up without the obligatory AI mention, so here’s what I’ll say: AI has generated unprecedented levels of productivity and capability gains across every single industry. As discussed in a16z’s podcast ‘The Future of Software Development’ where they discussed the implications of AI on infrastructure, these improvements will materialize in TAM expansions which will create greater net gains across the board. In a time like this, institutions, especially those meant to play the role of scientific and cultural leaders, cannot afford to be hyper-focused on de-risking. There is so much opportunity in serendipitous discovery and during a time of deep systemic disruption, the institutions that end up on top will be the ones who figure out how to capture idle potential in the discoveries others forgo.

Retailers have long served as gatekeepers between designers and consumers in the fashion industry, initially establishing their role as curators within the ecosystem. In 1943, Eleanor Lambert founded New York Fashion Week, a model that has spread globally and become the cornerstone of the fashion business. It began as an invite-only event, aggregating a curated selection of designers, journalists, and buyers, which generated excitement around new collections, facilitated purchases, and sparked critical relationships between stakeholders.

How It Worked

1. Curated showrooms: The event invited select designers to show pieces in one place, where they would have access to an aggregation of buyers

2. Media relationships: Select fashion editors were invited, which generated hype. Retailers depended on the media to inform trends, so the attending designers benefited from receiving exclusive coverage

3. Designer Personas: The curation and exclusivity transformed clothing from commodities that were purchased directly from manufacturers to brands with cultural value

4. Institutionalized Pipeline: Fashion Week became a commercial event that standardized the process of bringing garments from designers to consumers as it formally integrated the roles of the media and retailers, which influenced audiences, curated buying options, and delivered products to consumers

This was a time when there was a highly symbiotic relationship between designers and retailers. Fashion was beginning to shift from commodity to luxury, and retailers supported emerging designers by funding production runs with purchase orders and distributing products to aggregated audiences.

Shift in Power Dynamics

As department stores expanded and became more profit-seeking in the 1990s-2000s, buyers became more risk-averse and made decisions based on data and past sales rather than relying on editorial power through curation. Rather than acting as a launchpad for emerging designers, they became gatekeepers who required designers to have proven success before placing an order. Since retailers owned distribution, designers who were fortunate enough to pass the gate had very little leverage in negotiating terms. This forced them to bear more risk, such as delayed payment terms, have no say in markdown allowances, and pay for prime shelf space.

Relative to the broader fashion industry, the decline of retailers’ curation power and designers’ brand values was perpetuated by the rise of fast fashion, as cheaper clothing and shorter trend cycles made it impossible to compete for consumer demand. Given the power imbalance, department stores further pressured designers to produce larger collections—increasing both the pace and volume of designs—at more competitive prices, which was unsustainable for many businesses.

The redistribution of risk between designers and retailers can be attributed to the distribution leverage retailers hold. Despite the rise of direct-to-consumer distribution, audiences are still largely concentrated in the hands of retailers, with DTC e-commerce sales accounting for only 15%-19% of total sales in 2024 and 2025—the rest largely attributable to sales from retailers.

Implications on the Future of Retail

Cue The Devil Wears Prada’s Cerulean Blue monologue (here). Despite iconically capturing the role of the industry gatekeeper, fashion has evolved significantly since then. Today, curation power has shifted away from institutions towards individuals; however, product distribution still largely remains in the hands of institutions, which creates a false sense of decentralization.

First Wave Curators: Institutions (1950s - 2000s)

Prior to the social media era, curation was largely dominated by institutions such as magazines which dictated trends and retailers who selected and delivered products to consumers. These institutions had full-stack power as centralized tastemakers and acted as industry gatekeepers.

Examples include Lord & Taylor, Eleanor Lambert (NYFW, CFDA), Harper’s Bazaar

Second Wave Curators: Web 2.0 (2000s - 2010s)

The rise of the digital social sphere shifted attention to blogs and platforms which both aggregated audiences and made it easier to find niches such as Instagram, Tumblr and Reddit. Influencers became curators as selections became hyper-tailered and audiences developed parasocial trust.

Examples include Alexa Chung, Danielle Bernstein, Emma Chamberlain

Third Wave Curators: Web 4.0 (2020s+)

As the digital landscape shifted from celebrity influencers to micro-influencers and algorithms became more advanced at categorizing, consumers entered an era of hyper-individualism, fuelled by digital echo chambers. Coupled with AI-powered discovery tools, humans and computers now work hand-in-hand in selecting the perfect assortment of products for each individual.

Examples include Blueberry Milk Nails, Dubai Chocolate Matcha Labubu, Granola micro-influencers

Fashion’s Two-Pronged Problem: Curation vs. Distribution

As described above, the fashion industry operates on two core functions: curation and distribution. Technology has enabled curation to evolve and gradually become less centralized, with power disseminating from institutional gatekeepers to individual tastemakers. However, a sustainable ecosystem relies on thriving designers who are incentivized to continue to supply the market with innovative products, and this is where the challenge remains.

When a brand owns its audience, it owns its economics. This means having maximized control over pricing and product offerings. However, one of the biggest challenges brands face is distribution, as consumer discovery and having the right audience come across their offering is difficult to achieve. One of the most efficient ways to overcome this is by aggregating consumers and presenting them with a curated selection of products, which is why the power of distribution is still concentrated in retailers who own audiences. In order to access these audiences, brands are forced to comply with disadvantageous retailer terms. Despite losing their acclaim as curators, retailers have maintained their role as gatekeepers since they bridge the gap between designers and consumers.

A New Era for Retail

In the era of third-wave curators, information dissemination is targeted at niches and micro-communities, amplified by algorithms and AI personalization tools. The rise of the hyper-curator poses an existential threat to mega-retailers such as Saks, Nordstrom, and even SSENSE, whose business models reflect a brokenness in the status quo ecosystem. Store closures, bankruptcies, and layoffs imply a looming financial struggle as retailers depend on strong cash conversion cycles, full-price sell-through, and designer relationships to thrive—all of which have been on decline for years.

Definitions and Examples:

• Cash Conversion Cycles: The amount of time it takes to convert inventory into sales.

  Since designers and retailers must prepare products seasons in advance, lengthy lead times and high capital requirements, such as production deposits can strain businesses. This is why Sak’s has extended its vendor payment terms from 60 to 90 days as the retailer is unable to fulfil obligations with current cash flows.

• Full-Price Sell-Through: The percentage of inventory sold at original undiscounted value.

  Retailers—notably SSENSE—turn to discounting when products are nearing the end of a season or have limited consumer demand. Low full-price sell-through can indicate poor inventory management and harm brands by creating price competition.

• Designer Relationships: Retailers depend on designers to supply stores, while designers depend on retailers to distribute products.

  Strong relationships are mutually beneficial, while weak relationships result in distrust and exploitation.

Does Better Risk Distribution Lie Outside Traditional Retail?

Brands are increasingly investing in direct-to-consumer models as wholesale has always been substantially less profitable and retailers present a less promising value proposition due to poor terms and deteriorating designer relationships. However, the fundamental challenge with DTC strategies is accessing the right customers, if any, and this has been evident in the 20% plateau of e-commerce capture as outlined in the data above. Brands operating through DTC become reliant on marketing, whether through social media, SEO, or in-person. This brings the industry back to its roots, as there is no centralized place for buyers and sellers to meet.

Marketplaces may address this problem by bringing a curated selection of brands to an aggregated audience of customers. The trouble here lies in achieving balanced liquidity—while there is a critical mass of supply, there must be a compelling reason for consumers to migrate away from traditional retailers. The main driving forces on the demand side of the marketplace lie in price and quality of selection, which is crucial to the consumer experience.

This is where the increasing sophistication of curation processes will play a pivotal role in whether fashion captures the opening inflection point to create structural changes.

The Window of Opportunity is Opening Now

There is strong evidence of systemic brokenness in the fashion ecosystem. Despite monstrous power imbalances, profit takers are rare and the prevailing system is tremendously disadvantageous to designers which discourages innovation and creation—two things that underpin the success of the industry.

The rise of curation technologies such as AI tools and micro-communities enable marketplaces to thrive as social media has proven that a single aggregator can be segmented into hyper-specific niches. The industry can thrive without gatekeepers as sorting can now be achieved at scale with low human touch, evident in trends that originated and grew online, eventually seeping into offline markets.

A true open marketplace is the only way to unlock a thriving fashion ecosystem and this is possible with today’s technology. By economically empowering designers with targeted tools and an aggregated consumer base, a more sustainable equilibrium can be achieved. Designers previously gatekept from participating on the commercial side of the industry not only presents consumers with greater innovations and pricing dynamics, but allows consumers to establish true communities and take on an active participation role through direct channels to designers.