_           _     _
  ___ ___ ______ _ _ __  | |__   __ _| |__ (_)_ __
 / __/ _ \_  / _` | '__| | '_ \ / _` | '_ \| | '_ \
| (_|  __// / (_| | |    | |_) | (_| | |_) | | | | |
 \___\___/___\__,_|_|    |_.__/ \__,_|_.__/|_|_| |_|

cezar babin

Hi, I'm Cezar.

This is my little corner of the internet, where I share my work. Most recently I led software development at Installation Made Easy, a PE-owned enterprise software company that has grown 10x over the last couple of years. Before that I spent time building consumer and infrastructure products at Amazon, and have been around the AI-application space long enough to be just a little bit dangerous.

Most of what I publish on this site is an attempt to clarify my own thinking on questions I keep returning to.

• How does AI-powered software abundance change the business world, especially for non-tech companies? Studying business is all about studying competitive dynamics grounded in the laws of reality. Despite software eating the world, much of the literature on competition and moats has stayed relevant, yet AI is flipping a lot of those rules on their head. Do moats still exist, and what do they look like in the age of AI?
• How do the software rails change with AI? LLMs are now deterministic enough that they no longer require much surrounding Turing-complete code. Software can be instantiated on demand and is becoming increasingly ephemeral. All an AI model/agent needs is a set of primitives to run the computation as efficiently as possible.
• How does the abundance of production change distribution costs? We increasingly live in an economy where the only scarce factor is human attention. Distribution architectures will need serious rethinking to allow emerging actors to participate in the discourse.
• What does a world of frictionless capital → capability look like? AI promises to turn any idea into an implementation at a fixed cost. This is both incredibly inspiring (we will be able to cure any disease by simply allocating resources to it) and extremely unsettling (how do we decide what to allocate resources to?).
• How does social media change with AI? A small minority of social media users produce content while the vast majority consume (the 1% rule). The first-order effect of LLMs is an overflow of AI slop, which social networks are trying to rein in as best they can. But AI also presents an opportunity to elicit content from the silent majority. After all, TikTok is TikTok because it removed barriers to creating music videos, and Instagram made low-res photos look professionally edited. How will X/Meta lean into this?

I've been writing more here since May 2026. If something here sparks a thought, the fastest way to reach me is @sygmo1d on X.