Introduction
Over the past few years, tools like ChatGPT have sparked bold claims about developer productivity - 2×, 10×, even 100× gains. Stories of individuals building full applications or even browsers in days have fueled the narrative that we are entering a golden age of software creation. Naturally, this raises a simple but important question: if building software has become so much easier, where is all the new software?
To explore this, researchers including Alexis Gallagher and Rens Dimmendaal turned to PyPI, the largest public repository of Python packages. If an AI-driven productivity boom is real, we should expect to see clear evidence here - more packages, more updates, more activity. What they found, however, tells a far more nuanced story.
Key Insights
1. No explosion in software creation
Despite the rise of AI tools, the total number of new packages on PyPI has not surged dramatically. Growth continues steadily, but there is no visible “AI shockwave” in software production.
2. Spikes are misleading
Some increases in package uploads exist, but many are linked to spam or low-quality submissions - not meaningful software development.
3. Quantity isn’t the right metric
Creating a package is easy. What matters more is whether software is used, maintained, and improved over time.
4. Updates show only modest change
When looking at actively used packages, update frequency has increased slightly—but not dramatically enough to signal a major productivity leap.
5. The trend started before AI
The rise in update frequency began around 2019, before modern AI coding tools became mainstream. This suggests other factors, such as better development workflows, are at play.
6. Software still ages the same way
Packages are updated frequently early in their lifecycle and less often as they mature. AI has not changed this fundamental pattern.
7. AI packages are different
Software related to AI stands out. These packages are updated much more frequently than others—often more than twice as often.
8. Popular AI tools lead the surge
The increase is especially strong among the most popular AI packages, which show significantly higher update activity compared to both non-AI and less popular AI projects.
9. Not a universal productivity boom
If AI were making all developers dramatically more productive, we would see broad increases across all software categories. That simply isn’t happening.
10. Follow the money and hype
The most plausible explanation is not just productivity gains, but intense investment, attention, and competition in AI. More funding leads to more developers, faster iteration, and higher update frequency - especially in high-visibility projects.
Conclusion
The idea of an AI-driven software explosion is compelling - but the data suggests a more grounded reality. AI has not yet transformed software development across the board. There is no widespread surge in new applications, nor a dramatic increase in overall development activity. Instead, the impact is focused and uneven.
What we are witnessing is not a Cambrian explosion of all software, but a targeted acceleration within the AI ecosystem itself. AI tools may indeed be making some developers faster, but the larger effect appears to come from where the world is placing its bets - capital, talent, and attention are flowing into AI, and that is where the visible change is happening.
In other words, AI is not yet rewriting the rules of software creation everywhere - but it is rapidly reshaping the space where it matters most: software about AI itself.
[The Billion Hopes Research Team shares the latest AI updates for learning and awareness. Various sources are used. All copyrights acknowledged. This is not a professional, financial, personal or medical advice. Please consult domain experts before making decisions. Feedback welcome!]
