3 Comments

how important is revenue growth if that revenue channel isn’t protected by a ‘moat’ - if important, how do you assess moat/defensibility?

Expand full comment

great point and moat is definitely worth considering when assessing the long-term viability of a company. some thoughts:

- the big AI labs + some infrastructure companies (Together/Fireworks) have a natural moat with the capital intensive nature of their products that serves as a barrier to entry

- AI product companies technically have less of a moat. Still, they're making a lot of money, and growing really fast, making them amazing place to work and grow IMO.

For example, it's not clear to me *yet what Cursor/Codium's long term moat is, and VSCode is quickly catching up. So it's hard to say whether those companies will be around in 10 years (I hope they are). But the metrics show that they're in real hypergrowth, and they're pushing the envelope forward by consistently being first to market (ex: Windsurf) - and thus they're worth working at for the sheer fun and career growth that comes with being at a place growing that fast

Expand full comment

There's a typo for "SaaS" in the paragraph under the revenue cycle diagram. Great points, but I too wonder whether hypergrowth is reliable. Much too often, VC-backed companies grow too fast, enshittify and then fall off just as quick. This happens less so once ARR is upwards of say 100M, but it still happens. Relevant clip: https://youtu.be/BzAdXyPYKQo?si=k5Y_bT2QLra29QUj

They have amazing revenue but unit economics is very poor, which eventually bites them back. Another issue is when they grow too fast, they may hire exponentially which causes immense slowdowns, culture wars, poor cohesion and eventually attrition. Idk really know what a good signal for growing startups is, but I think a good guiding principle is to follow really, really odd (in a good way), passionate and skilled people. Nearly always a fun place with non-linear experiences. Being in great company of cool oddballs is highly underrated imo

Expand full comment