commons.systems

This blog is built to my own specification - it has few dependencies on third-party platforms, and it uses the technologies I like and am comfortable maintaining. The dependencies it does have are deliberate. This blog relies on standards that are open source and freely available (HTML, JavaScript, golang, git), or measured choices to adopt third-party infrastructure with deliberate lock-in costs (GitHub, Firebase, Claude).

Coding agents are reaching a level of efficacy where they are changing how technology like this blog is deployed. Common value propositions for agentic coding overlook an important strategic advantage - recovered autonomy in technology deployment.

There will be a new breed of platforms seeking to monetize this trend by "enabling organizations to leverage" the new technology, and in doing so, position their platform as the new gatekeepers. Platforms will "accelerate adoption" with a new generation of low-code, automation, and governance solutions. This legacy approach of platform capture will fail to deliver on the most disruptive promise of the new technology.

Autonomy is not recovered with naive delegation to coding agents. That approach neglects the professional experience needed to build software that is deeply aligned with business objectives. Likewise, if the agents are only augmenting business as usual, it won't result in the velocities needed to decouple solutions from legacy dependencies. I will not be able to build a blog site without my business becoming a blog site business.

Recovering autonomy means applying expertise with enough velocity that the logistics don't become the core business.

When delegating implementation work to coding agents the recovered time must be used for augmented planning and design workflows. Whether coding by hand or using agents, the time to develop a v1 may be similar. By using agents, the developer can dedicate themselves to a well-aligned architecture, while maintaining a total cost of ownership on par with third-party platforms.