Tailoring taste.
Some reflections on building Inter Alia, a visual text-image fashion recommendation search.
It’s been a year since I built and launched Inter Alia. This project is also the one that tries to change my life the most — from giving me a deeper grasp of the problem to some exciting co-founding opportunities.
I often look at fashion as a tool for self-expression and think of it as a medium to express one’s taste. I made Inter Alia with the mindset of building broader toolkit for others to enjoy the process of discovering and buying items that inspire. If style is a creative outlet then we need to build better infrastructure enabling others to view shopping as an act for self-expression.
We are all on the spectrum on how much we want to use this tool in our lives. Some people are inherently more interested in using it, others only occasionally (e.g. during holidays). If you are working in fashion you are definitely living it. Nevertheless, even if you don’t think of clothes this way — you can inevitably find yourself, clicking from one brand’s site to another to find what to wear for your’s friends wedding, spending hours narrowing down options to find a couple of dresses, and then another hour choosing which one you are going to buy. I’m also omitting a huge chunk of details in the process — evaluating items based on how much you can afford in a given moment, imagining in your head how you’d look like in each of the options, waiting for your friend to text back to get their opinions, and more.
I remember one day in the coffeeshop in NY I accidentally found myself observing how a person is searching a winter scarf on their Chrome Book. They were searching for a winter scarf on their Chromebook. After some scrolling, they copied a product description - "fringe wool scarf" - and pasted it into Google to find more options. They proceeded to browse and click through various websites for about 40 minutes. But ultimately they gave up their search empty-handed and started reading a book instead.
In thinking about the nuances of fashion search, I realized we must discern precisely when the transition occurs in the shopping journey—from a stage of genuine enjoyment to one of sheer frustration and solve for it. Pinpointing the moment when a seemingly enjoyable experience turns painful is deeply rooted in the psychology of choice and the nuances of user experience.
I built Inter Alia not because I am a frequent shopper, but because I often look at various fashion items as a source for creativity in exploring styles and trends. Even if I don’t end up purchasing the item I want to leave this rabbit hole process with more inspiration than pain. However, current shopping sites lack key human elements that enrich in-person shopping. In my view, the new kind of search experience can facilitate:
Expressive inputs - allowing people to describe desired styles vs just keywords
Meticulous refinement - an iterative process of tailoring options to taste
Contextualized decision-making - seeing items in a variety of lifestyle context
A tighter critique loop - quickly evaluating and adjusting results
Social collaboration - getting input from friends
A creative sandbox - freely combining styles and elements
This makes me think back to the Victorian era, when people would purchase fabrics, linings, trimmings and threads to make their own clothes or work with a tailor. This was cheaper than buying ready-made garments. I feel we should think about search like that too: mirroring the one-of-a-kind tailoring experiences of the past. I don’t know where this project is going to take me, but I definitely think rather than carelessly throwing people into the sea of all options, I envision search to be a creative medium - a rich toolkit where users can expressively blend styles and elements to convey their unique tastes.
Curiousities
I love Patrick Collison’s quote on why making things beautiful matters
Define CTO by Greg Brockman — very genuine reflections on and returning to the thing that he loves the most, writing code.
How to Get Startup Ideas by Paul Graham — “Live in the future, then build what’s missing.”
Remarkable progress of CLIP models since 2021, however, we need better benchmarks on real-world use cases
Some of my most favorite fashion shows of all time:
Wow, what really impresses me about Inter Alia is how snappy it is. What're you using to decide which img urls to load?