Answer.AI PyPI study sees no broad surge

The real answer to "where are all the AI apps?" is that they are mostly in the AI tooling layer, not in a giant wave of new consumer software. The broader 100x-productivity story still does not show up in aggregate data, but AI-native packages are clearly churning faster.

• –New package creation on PyPI does not show a clean post-ChatGPT inflection; the obvious spikes are mostly spam and malware, not healthy ecosystem growth.
• –Overall update frequency only rises modestly, and the secular trend starts before modern AI coding tools, so continuous integration still explains part of the curve.
• –AI-related packages are the outlier: recent AI packages hit about 20 releases in their first year, versus about 10 for comparable non-AI packages.
• –The strongest effect is in popular AI packages, which jump to roughly 21-26 releases per year, suggesting money, hype, and active experimentation are concentrating effort there.
• –Because the repo and data are public, the methodology is easier to audit than a hand-wavy productivity anecdote.