Parsnip’s vision is to create AI-accelerated learning for anyone, supercharging how fast you can absorb knowledge into your brain. This opens a path toward creating superintelligent AI assistants.
Our core insight is to adapt the idea of “skill trees” for character development in video games to real life, giving you a personal map of what you know and don’t know, and showing you what you need to learn to accomplish any goal. Skill trees complement GenAI / LLMs by creating “physics” for guardrails against hallucination, and allow LLMs to deliver personalized content.
This approach provides effective education outside of the classroom — and in fact anywhere, turning learning from boring lectures and tests into interactive, engaging, personalized consumer products that are integrated into your life: a continual, repeated process of learning and doing.
We see food & cooking as an ideal first vertical for this technology. It’s a massive, underserved market, a straightforward application of skill trees, and a (so far unclaimed) platform business with a huge moat if we succeed.
<aside> 💡 Our thesis memo, How to personalize learning with AI 👩🎓, explains the theory behind the tech and why food is a fantastic vertical from both a technical and business perspective.
</aside>
Andrew Mao (CEO): PhD in computer science @ Harvard; computer science + Wharton @ Penn; Postdoc @ Microsoft Research; ML scientist + PM at CTRL-labs (deeptech startup sold to Meta for ~$1B). In his research career, Andrew conducted a prisoner’s dilemma experiment on the Internet that was >10 times longer than any extant experimental economics study.
Andrew ate very poorly growing up in an immigrant family and didn’t learn to cook until well into his PhD. After finding academia quite unfulfilling, he realized that the questions (1) “how do I cook?” and (2) “what should I make for dinner?” were both solvable at scale with AI, and here we are.
Dan Sosa: PhD in AI/bioinformatics @ Stanford, computer science + molecular biology + Sloan @ MIT. Dan spent his PhD fine-tuning LLMs to build knowledge graphs from millions of biomedical research articles for drug discovery — the exact skillset for building Parsnip’s AI system. Dan also found time to conduct and compose an original score for the Stanford Symphony Orchestra during his PhD, as well as founding a smart AI cookbook startup, which led him to Parsnip.
Coming into COVID-19 lockdown, Dan was frustrated by his cooking and realized that applying his research would make leveling up so much easier. A son of immigrants, cooking has re-connected Dan with his Guatemalan roots. Building Parsnip to empower others is the most fulfilling thing he could do.
We’re backed by Luis von Ahn and Severin Hacker (the co-founders of Duolingo), Rana el Kaliouby (CEO of Affectiva), and Mark Miller (founder of 7 restaurants & James Beard “Best Chef” laureate).
Cooking is a core everyday skill that is no longer taught. As a result, demand for learning to cook increases every year [1]. We can’t solve this problem all at once, but it turns out that beginner cooks want a solution very badly. See examples of why Learning to cook is a painkiller, not a vitamin for them.
The Parsnip app is the painkiller folks were looking for, but didn’t know they needed. It has 4.9⭐ on both iOS and Android with $0 CAC for the >40k downloads we've had, and was featured several times in the App Store. It has also reached #1 on HN and went viral on Reddit several times. See what Users who love Parsnip (Testimonials) say, a **slew of data indicating product-market fit,** and a more recent Parsnip product data review (2024).