Startup update 14: In search of a nail

Originally posted 2025-07-28

Obligatory disclaimer: all opinions are mine and not of my employer

Progress update

This week, I published what I have of the AP chemistry course, and then started looking around at competing education products. I wanted to understand the very complex Venn diagram of existing products, target markets, competitors’ features, my features, and figure out what my differentiating features were, so I could start tuning a marketing message.

I realized then that I’d been going at this just completely backwards the whole time. Instead of building an AI course and then trying to find an audience for it, I really should have just started with the marketing message and then built the product around that… Really stupidly obvious in hindsight, and now I’m back to the drawing board.

This time, I’m focusing directly at the chemistry olympiad prep market, which I estimate to be pretty tiny (probably a 1M ARR market), but is a tractable enough space that I can actually just analyze all of the market offerings and likely make something better. I’ve done the research, figured out all of the existing products/services/people/offerings in this sector, as well as used a bunch of similar products from adjacent, better resourced sectors (math olympiad prep, software interview/coding competition prep), and have found the relevant communities/discord channels where chemistry olympiad kids are actually hanging out. Based on this research, I have a list of features for an MVP. The marketing message is simple: “the best way to prepare for the chemistry olympiad”. Given how tiny the market is, the bar is pretty low here.

This week, I’ll try to get out this MVP of a problem solving interface that is superior to anything currently existing for chemistry olympiad students. I might dump the backend/frontend I developed over the past month, just to maximize dev speed. I have all of the learnings on how to use LLMs to teach classes, as well as the raw content (course curriculum/lesson sketches + problems + solution generating routines), so it’s not a big loss.

Thoughts on the fragmented state of education

I am starting to realize “education” is an umbrella market in the same sense that “cancer” is an umbrella disease.

In any organism, every system needs to stay healthy and functioning, which typically means some mix of cell repair, replication, and refresh. Different cells have different natural lifecycles, environmental exposures, and propensity to damage. “Cancer” is a failure of this process in some subset of cells. The treatment method for any cancer is a very specific function of the type of cell, type of mutation, proximity/accessibility of various treatment routes, and so on.

In modern civilization, every person needs to become and remain gainfully employed, even as the market for jobs shifts over time. Different professions have different Schelling points defining the qualifications for a job, whether it is a certification, a license, a degree, or an exam score. The market is often free to decide that a specific qualification is no longer serving its purpose and start requiring a different qualification. Every time somebody needs to realign their resume with the necessary qualifications, we call this “education”. Education products are a very specific function of the desired job, type of qualification needed, student ability, existing skillset, and availability of time.

The fragmented state of education is a reflection of a healthy and diverse job market. Chemistry is relevant for pre-meds hoping to do well in the MCAT, for chemistry undergrads hoping to do well on the GRE, for high schoolers hoping to get good scores on the AP/IB chemistry exams, or even for software engineers trying to pivot into climate work, but all of these are really different markets, and hard to target with a single tool.

I would even argue that education is not fragmented enough. In the U.S., K-12 and then four years of college is the baseline expectation for a variety of professions, when it shouldn’t be. Outside of the U.S., medical degrees typically do not require an undergraduate degree, and I think given the stupidly long timelines of medical school + residency + fellowship that most U.S. doctors go through, four additional years in undergrad is really overkill. Vocational/trade schools probably shouldn’t even require the completion of high school. Ultimately, educational products that reduce the amount of time required to satisfy unnecessary requirements are probably a net gain to society.