Building a Gen Z Art Platform without the Algorithm Trap: Devika Sarin of Soal
with Devika Sarin, Soal
Building a Gen Z Art Platform without the Algorithm Trap: Devika Sarin of Soal
Show Notes
Devika Sarin spent years watching the art world fail the exact people it should have been welcoming. Millennials and Gen Z who were genuinely curious about art - but had no context, no confidence, and no on-ramp. The discovery tools that existed were either too academic, too expensive, or too captured by the same algorithmic logic that had turned social media into an endless scroll of gym selfies and outrage.
Devika built Soal to be something different: a daily ritual for art discovery that uses AI recommendation built on scientific and behavioral principles to help people figure out what they actually like - and then connects them directly to living, working artists they can follow, visit, and buy from.
This episode is about what it means to build a platform with genuine aesthetic intention in an era when every algorithm is optimizing for the wrong signal - and why the founders who refuse the engagement trap might be the ones who build something that actually lasts.
What Soal Is Building
Soal is not a marketplace and it is not a social network. Devika describes it as a third space - not a physical place, but somewhere you go to learn, explore, and develop your taste over time. The daily ritual format gives users a curated set of artworks to move through, with the AI learning from their responses to surface more of what resonates.
The artist focus is intentional: primarily contemporary working artists doing 2D work - painting, photography - who can be discovered alongside the canonical names they connect to. The goal is to walk a user from recognizing Monet to realizing there is a painter in Los Angeles doing something adjacent who has prints available for $500.
Nearly 90% of the artists who joined the platform organically during the pilot were women - a product of the community Devika cultivated and the culture she built before launch. The platform facilitated print transactions during the pilot. The monetization model going forward is community-first, with print sales, brand partnerships, and artist-facing tools all in development - but no ads.
Frameworks from This Episode
These frameworks have been added to the AI for Founders Frameworks Library. Filter by Devika Sarin to find them.
Curate First, Democratize Second
Every platform that opens cold fails. Set the standard through curation, then invite the community in. The initial editorial layer defines the culture that self-publishing later inherits.
- •A platform's culture is set in its first hundred pieces of content and its first hundred community members. What you put in determines what kind of people show up.
- •Curation is not exclusion - it is intention. Once the standard is established, the community can self-select and self-publish because they already understand what the space is for.
- •Devika did not plan for 90% of Soal's pilot artists to be women. It happened because the culture she cultivated attracted like-minded artists, who then referred other like-minded artists. Culture compounds.
- •The alternative - open publishing from day one - produces a generic marketplace that looks like every other marketplace. The moat is the taste, not the technology.
The Algorithm Trap
Engagement signals and preference signals are different things. Every platform that optimizes for engagement ends up showing users what they react to, not what they actually want - and users eventually leave because the product no longer reflects who they are.
- •The trap: a user pauses on a video for three seconds. The algorithm reads that as interest. It floods the feed with similar content. The user never asked for this - they just had a human reaction.
- •The result is a feed that reflects your worst involuntary impulses, not your considered preferences. Users feel embarrassed by what their algorithm reveals about them.
- •The alternative is to optimize for depth of engagement over speed of engagement - time spent with an artwork, return visits, comments, purchases - signals that reflect genuine preference rather than reflexive reaction.
- •Soal is explicitly not building rage bait or click bait mechanics. The goal is a space that makes you feel good - which requires resisting the optimization pressure that every ad-supported platform eventually surrenders to.
- •Banksy is the rare artist who does both: provokes a strong reaction and makes you feel something real. That is the standard, not the exception to be emulated.
The Third Space for Culture
The most valuable thing a platform can build is a place people want to go when they are not working and not consuming. A third space is defined by the feeling of being there, not by the transactions that happen in it.
- •First space: home. Second space: work. Third space: everywhere else - coffee shops, parks, community centers. Digital third spaces are almost all gone, replaced by engagement-optimized platforms.
- •The value of a genuine third space is not measured in sessions or clicks. It is measured in whether people feel better after spending time there.
- •Art has historically lived in third spaces - galleries, museums, public murals. Soal is building the digital equivalent: somewhere to go to explore, not to be monetized.
- •Substack, in Devika's observation, is becoming a third space for long-form thought. The demand is real and growing as people consciously step back from algorithmic feeds.
- •A product that makes people feel good has naturally higher retention than one that addicts them through anxiety and outrage. The business case for genuine third spaces is better than it looks.
Community Before Monetization
The founders who try to monetize before they have community end up with neither. The founders who build community first often discover that monetization follows naturally - and is more defensible when it does.
- •Devika bootstrapped Soal through years of prototyping, testing, and interviewing before taking outside money. The result: deep product clarity and a pilot community with genuine attachment to the platform.
- •The pressure in San Francisco is to build for investment, not for users. Those two objectives are not the same and frequently conflict.
- •Product-market fit for a community product looks different than for a SaaS product: it is when people tell other people unprompted, when artists refer other artists, when users feel the platform reflects who they are.
- •Once community is established, monetization options multiply: print sales, brand partnerships, artist tools, events, subscriptions. None of these work without community first.
- •The Depop comparison is instructive: $115M ARR from a vintage fashion marketplace. Art discovery with the same community mechanics has a comparable addressable market that is largely uncontested.
Founder Experiment: Map Your Own Taste Profile
Devika's insight is that most people already have taste - they just cannot articulate it yet. This experiment is about making your implicit aesthetic preferences explicit, which is the same process Soal automates at scale.
- 1Open a blank document. Write down five things you own that you genuinely love the look of - not because they were expensive or because someone told you they were good, but because you find yourself looking at them. Objects, images, spaces, anything.
- 2For each item, write one sentence about why. Do not use the word "nice." Force specificity: color, texture, feeling, association, memory.
- 3Search for the artist or designer behind each item. Find three other works by the same person. Note what stays consistent and what surprises you.
- 4Now find one artist whose work shares qualities with at least two of the five items you listed. You have just done the first iteration of what Soal's recommendation engine does algorithmically.
- 5Sign up for Soal's waitlist at thisissoal.com and spend one week with the daily discovery ritual. At the end of the week, compare what the platform surfaced to the taste profile you mapped manually. The delta between those two results is where your taste is still opaque - and where the most interesting discovery happens.
For founders: The same exercise works for any personalization product you are building. Manual taste mapping before algorithmic taste mapping reveals the edge cases your model will miss - and those edge cases are where your most loyal users live.
Key Terms
These terms have been added to the AI for Founders Glossary. Search by Devika Sarin to filter them.
Tools from This Episode
Soal
Art discovery platform for millennials and Gen Z. Daily ritual format surfaces curated artworks, builds your taste profile over time using AI recommendation grounded in scientific and behavioral principles, and connects you directly to living contemporary artists you can follow, visit, and buy from. Pilot facilitated direct print sales. No ads. Community-first. Waitlist open.
Q&A
What is Soal?
Soal is an art discovery platform designed for millennials and Gen Z who are interested in art but lack the context or confidence to get started. The app creates a daily ritual for art discovery, uses AI recommendation built on scientific and behavioral principles to learn your taste over time, and connects you to living contemporary artists you can follow, visit, and buy from.
Who is Devika Sarin?
Devika Sarin is the founder of Soal, based in San Francisco. She has been building in the art and technology space across two startups - her first involved blockchain and art. Her personal aesthetic runs toward abstract art, Mark Rothko, nature-inspired dreamlike imagery, and strong portraiture. She bootstrapped Soal before bringing on early investors.
What types of artists are on Soal?
Soal focuses primarily on contemporary working artists doing 2D work - painting and photography across multiple genres. The platform also includes recognized historical artists as context, helping users understand how living artists connect to masters they might already know. Approximately 90% of artists who joined the pilot organically were women, a natural result of the community culture Devika cultivated.
How does Soal's AI recommendation work?
Soal's AI recommendation engine is built on scientific and behavioral principles rather than pure engagement signals. Users move through a daily set of artworks; the system learns from their responses over time to surface more of what resonates. The approach is designed to distinguish between what users react to involuntarily and what they genuinely prefer - specifically to avoid the engagement trap that plagues social media feeds.
What is the business model?
Soal facilitated print transactions during its pilot and print sales are a component of the ongoing monetization model. The platform is also developing brand partnership and collaboration revenue, and is exploring ways to help artists monetize their participation beyond just posting and promoting. There are no ads. The stated philosophy is community-first, then monetization.
Is Soal doing AI-generated art?
No. Soal is explicitly focused on human artists and is not building a generative AI art product. The AI is in the recommendation and personalization layer - helping users discover artists and develop their taste - not in the artwork itself.
How is Soal different from social media art discovery?
Social media surfaces art based on engagement signals - pauses, clicks, follower counts - rather than genuine aesthetic preference. It also rewards artists who behave like influencers (consistent posting, personal brand building) rather than for the quality of their work. Soal is building an algorithm that can surface artworks based on the content of the artwork itself, and a platform that gives artists a community without requiring them to perform for an algorithm.
How was Soal funded?
Soal was bootstrapped by Devika for a long time before recently bringing on early investors. The extended bootstrapped period allowed for extensive prototyping, testing, and user interviewing - resulting in deeper product clarity and a pilot community with genuine attachment before any outside money was involved.