On Deck Fellowship
In February, I joined On Deck Fellowship and flew to San Francisco for a one-week in-person onboarding experience as part of its 27th cohort (ODF27).
I started the onboarding week feeling depleted and burned out. I left feeling much more energized and excited about what comes next.
It has been one of the most valuable communities I’ve been part of. I made genuine friends, including Nilu Kulasingham, Joy Cai, and Chinmaya Joshi, and met many others I’ve stayed in touch with since. A few of us have even continued collaborating on small projects after onboarding ended.
Here are some of the lessons I took away from onboarding week.
Project → Product → Company.
A guest speaker who had built three successful companies shared this framework for how he starts companies. All of them started as projects. Most projects do not become products, and fewer still become companies. This really resonated with me because I often get frustrated by my seeming lack of progress toward starting a company, but starting (and killing) projects is an essential part of the process for building a successful business.
The lesson: start more projects. Motion creates clarity.
Three types of companies: Rapid Fire, Total Immersion, Long Aim.
Another framework that stuck with me was the idea that there are different kinds of companies, and each one needs to be built differently.
The one that resonated most with me was Total Immersion: a company where the founder lives and breathes the problem, deeply understands the customer, and often is the customer. The next company I build will almost certainly be of the Total Immersion type.
Solo founding is underrated.
It is conventional wisdom in Silicon Valley, and among VCs, that every founder needs a co-founder. But I’ve become more skeptical of that advice. Some of the best companies were built by solo founders, and even in companies with multiple founders, the reality is often more asymmetric than the mythology suggests. Solo founders also have a structural advantage: they retain more equity, which can make it easier to recruit exceptional early talent.
AI had its second “iPhone moment” in November 2025.
When I first started using Claude Code in spring 2025, it honestly did not feel like much of an improvement over using ChatGPT. Wow, has that changed. Claude Code became way more powerful in late 2025. Everyone was obsessed with their newfound powers from using Claude Code, and ChatGPT Codex to a lesser extent. After I left On Deck, I really started tinkering, experimenting, and building artifacts with these tools in earnest.
Fractional work is an underrated entrepreneurial path.
Before ODF, I had never seriously considered working part-time. One of the people I met in my cohort, Michael Fajardo, worked fractionally for multiple clients.
He used that not just to fund himself, but to immerse himself in specific roles, including chief of staff, bookkeeper, and others, in order to uncover opportunities for products or productized services. Later, I also learned about Manual Override, an accelerator focused on helping fractional workers become independent entrepreneurs, through Taylor Crane, an ODFer building FractionalJobs.io.
This is a path I am exploring during my sabbatical to extend my runway, perhaps indefinitely, learn new skills, and build a bridge toward a more ambitious business.
