Building upon your thoughtful insights here (and acknowledging my bias as your co-founder at Airheart), early-stage startups indeed follow a trajectory remarkably similar to the algorithms you’ve outlined. Pre-seed founders, as you emphasized, must behave like a greedy algorithm—making quick, impactful decisions—not simply because it’s convenient, but because they face severe constraints: limited resources, limited runway, and incomplete knowledge of the real-world problem space. Your experience at Airheart illustrates this clearly. Rather than exhaustively mapping out every nuance of travel-planning complexities at the start, you chose to quickly address the most immediate and obvious pain point—COVID-related travel restrictions—allowing us to rapidly build audience trust and initial momentum despite uncertainty.

However, precisely because of these resource and knowledge constraints, it’s equally vital at this stage to develop solutions that are inherently flexible. The irony of the greedy approach is that while it prioritizes immediate returns, success depends on consciously avoiding rigid designs that would prematurely lock in assumptions about the problem space. By intentionally structuring early solutions for extensibility, startups create pathways for future growth and pivots, efficiently leveraging scarce resources while continuously deepening their understanding of customer needs. While decisions made early on won’t guarantee graceful scaling—especially as experimentation and complexity inevitably continue—thoughtful flexibility at the outset can greatly support the transition into later stages.

Hello! I'm the author. I'd love to hear what you think about this post. My long-term goal is to come up with a generalized model and/or algorithm for startups. I didn't want to make this Substack post too technical, so a broader audience could also enjoy it.

I think there are many dimensions for startups (funding, team size, micro and macroeconomics, etc) and I want to see how these fit into a a high-dimensionality model. My background is mostly in computer science, so I will need to do a bit more research in a few other domains (economics, business management) to understand what makes startups work and fail and draw on existing research.

Eventually I am hoping this model can explain what makes some startups work, and others fail in a very broad sense.

The evolution described is a necessary progression. I have worked for a company that resisted their progression and it showed. Very sloppy when a large org has all the resources to build quality software but still opts for speed

applying algorithmic strategy concepts to growth strategy is an interesting idea. I wonder what other domains this could be useful for