Simon Willison argues that LLM coding agents are not forcing developers toward 'boring' mainstream technologies as previously feared. Recent models with long context windows can effectively learn from documentation of new or private tools, adapting to unfamiliar codebases by studying examples and testing their own output. This challenges the assumption that AI coding would reinforce established technology stacks.
Background
There has been ongoing concern that AI coding assistants would bias developers toward popular technologies well-represented in training data, potentially stifling innovation in newer tools. This phenomenon was dubbed the 'Choose Boring Technology' effect.
- Source
- Simon Willison
- Published
- Mar 9, 2026 at 09:37 PM
- Score
- 6.0 / 10