Software has been a compounding game. For about 30 years every new app is built on libraries, sits on frameworks, which sit on runtimes, which sit on operating systems. Most of it being open source, created by people who never met.
That stack is the reason solo devs exist and ship in a weekend that would've taken a year with a whole team in 1995.
There are three forces going full press against this stack:
Maintainer Burnout
Single individuals hold up infrastructure as unpaid maintainers. Used by trillion dollar companies. This used to be the classic problem. What's new? The emotional cost of running critical infrastructure for free, without support (and potentially even haters) is so large that these unpaid maintainers are leaving faster than they are being replaced.
2. Code is not the bottleneck anymore
Coding used to be the expensive part of creating software. Now PR numbers are exploding making it hard to find ones, that are NOT generated by AI, looking plausible, being well formatted and consisting of garbage. The bottleneck now shifted from writing the code to trusting the code.
3. Vulnerabilities that are being exposed faster than humans can patch them
Period. Okay, almost. The same speed which is encountered in the increase of PR numbers is found in your inbox as a maintainer listing a new bug, undiscovered for 27 years (1OpenBSD's TCP stack).
The conclusion is wrong
2Cal.com's response: going closed source. The logic: if AI can scan your blueprints, hide it.
Does this fix any of the three pillars above? Maintainers still burn out, plus there is no community to help them? AI still floods the codebase, closed or not. AI can still scan the binary (yes compiled binary analysis exist, OpenAI even shipped a model for this).
For every project that goes private, there is one less brick others can build on. The 30 year old stack works because each layer is auditable, forkable, reusable. If you remove the bricks in a wall... guess what.
What that actually indicates
"AI makes developer useless". The opposite is the case. AI makes the humans the scarcest ressource. Reviewers, maintainers, judgement. The economic model of open source (trust the commons, donate the labour) was already fragile. AI is the stress test thats revealing it was broken.
The fix is not closing the source, it's funding the humans who hold the stack together.
Sources
„Vibe Coding Kills Open Source“: https://arxiv.org/abs/2601.15494