One problem with a lot of the skepticism around AI produced software is that it focuses on existing ways of packaging and delivering software. PyPi packages are one example, shipping “apps” another.

While it’s interesting to see that in open source software the increase is not dramatic, this ignores however many people are now gen-coding software they will never publish just for them, or which winds up on hosting platforms like Replit.

Theres tons of ai apps. They're all general use chatbots or coding agents. Manus, Cursor, ChatGPT. Almost every app that has a robust search uses a reranker llm. AI is everywhere.

As far as totally new products - I built one (Habit.am - wordless journaling for mental health) and new products require new habits, people trying new things, its not that easy to change people's behavior. It would be much easier for me to sell my little app if it was a literal plain old journal.

well, many apps i made are really good but i would never bother to share it, takes unnecessary effort and i don't really know what works best for me will work like that for others

There are actually a lot of new startups coming out with agentic workflows, and they're probably moving fast. But to your point, there's probably still a lot of friction that keeps the average person/dev from launching new companies.

Even taking the “we’re all 100x more efficient at writing code” argument at face value… there’s still all of the product/market fit, marketing, sales, etc “schlep” which is very much non-trivial.

Are there any agentic sales and marketing offerings?

Because being able to reliably hand off that part of the value chain to an agent would close a real gap. (Not sure this can be done in reality)

This is just counting pypi packages. Why would I go to the effort of publishing a library or cli tool that took me ten minutes to create? Especially in an environment where open source contributions from strangers are useless. If anything I'd expect useful AI to reduce the number of new pypi packages.

- this would be much more insightful if the author takes the number of submissions to producthunt and the top 10 saas directories as the measure to see how many new apps were created pre AI and post AI era

- product hunt or app sumo is something i believe everyone tries to get a submission to which would truly measure how many new apps are we having per month these days

By "apps" this author apparently means "PyPi packages". This is a bafflingly myopic perspective in a world of myopic perspectives. Do we really expect people vibecoding "apps" to put anything on PyPi as a result? They're consumers of packagers, not creators.

I don't blame people for responding to the title instead of the article, because the article itself doesn't bother to answer its own question.

I don't think people are using AI to create new dependencies that they're then submitting to open source package managers (which is what this shows)

This is more useful for discussing what kind of projects AI is being used for than whether it's being used.