How a startup loses its spark#



  • Article touches on the topic of scaling engineering organization from 10 to 100.

  • Not trying to introduce any clever trick to fight the inherent properties of scale:

    • More specialization.

    • Less impact.

    • Less awareness.

    • Higher communication overhead.

    • Lower risk tolerance.

  • The advice is not accelerate the effects of scale:

    • Don’t introduce heavy-weight project management tools (hello, Jira). Approach scaling issues from the first principles instead of copying old method popular among big enterprises.

    • Keep each team small and independent. Citing Rippling as an example. Also noting, that this approach doesn’t work in practice due to high coupling between components of a product.

    • The best incentive is to give engineers fat chunk of equity and the power to influence company’s value.

  • But the actual solution is to stay small. Maybe with assist from AI.


  • When inspecting the issues of scale from the perspective of “fun”, author runs to about the same conclusions as Brooks in Mythical Man-Month. But Brooks didn’t have ChatGPT.