Simon Willison's personal observation, while intriguing, underscores how thin the evidence remains for AI's transformative impact on software creation. The spike could be coincidental or driven by other factors entirely. The real test will be whether this activity sustains and translates into meaningful feature development or simply reflects more frequent, trivial updates.
In our view, the industry needs more rigorous, longitudinal studies before declaring a new paradigm. For now, it's a data point suggesting that powerful new tools might be getting used, not proof they are making better software.