Internal Tools Are Back, and LLMs Are Why
A thread titled 'The Year of Internal Tools' had founders and engineers describing a clear pattern: LLMs have made building custom internal tooling fast enough that the calculus on buy-versus-build has shifted. Comments described canceling SaaS subscriptions and replacing them with purpose-built tools, using AI coding assistants like Claude Code to grind through backlogs of internal tooling ideas that were never worth the time before.
This connects to a separate thread where a 60-year-old engineer described Claude Code re-igniting their passion for building, and another thread where engineers described replacing Claude and GPT with local models for daily coding tasks. The throughline is that the cost and time required to build custom software has dropped enough to change what is worth building yourself.
The counterpoint in the internal tools thread is that SaaS companies selling to enterprises have more defensibility than SaaS companies selling to developers. Developers who can now build their own tools are also the most likely SaaS churn risk.
So what?
If you are selling developer tools or internal productivity SaaS to technical teams, your churn risk is higher than it was 18 months ago. The customers most capable of replacing you with a custom build are now more likely to do it. The response is to go deeper on the integration and workflow complexity that is still genuinely hard to replicate, not to compete on features that LLMs can now stub out in an afternoon.