A couple of days ago, I was working on an assignment that required a slightly higher level of coding than I normally master. It was, ultimately, a rather simple Python script, but in fairness, most of my skills lie in visual programming and basic C++.

This means I can read and understand most code reasonably well, but syntax is a completely different matter. That is precisely why I turned to AI for assistance with the task.

As I was filling in my code snippet, however, I suddenly paused for a moment and envisioned the situation that would arise the moment someone pulled the plug on the service. This isn’t far-fetched at all, is happening as we speak.

I don’t expect everyone to work in IT, so a brief introduction is probably necessary to make the topic a little less obscure. If you work in software development, chances are you have used GitHub. If you are into videogames, you might recognize the name as well, as gaming and software development communities have often overlapped.

As the platform evolved, it became—much like Stack Overflow—a valuable resource for developers. The difference was that GitHub allowed users not only to discuss solutions but also to share and reuse actual source code. After all, the software industry rarely develops everything from scratch; much like writing a text, you do not invent every expression yourself. GitHub made it possible to build upon the work of others and interact with them much more directly.

Over time, as GitHub became increasingly integrated into professional development pipelines, it evolved into a critical service for software development. In many ways, even before cloud computing became commonplace, GitHub was already enabling developers across the world to share repositories, collaborate on projects, and contribute to the same codebases regardless of their geographic location.

Its value eventually became so significant that the company was acquired by Microsoft, while still retaining its brand identity and familiar appearance. That said, it did not take long for GitHub to incorporate AI through its own version of Copilot.

The move was initially met with criticism, largely because of the open secret that the model had been trained on publicly available repositories, some of which may have contained other people’s intellectual property. Despite these concerns, the tool was rapidly embraced by the developer community.

After all, it saved time and helped perform analyses and validations that would otherwise have required considerably more effort. Many developers subscribed almost immediately, as the cost was easily justified by the productivity gains. In practice, they were simply paying for a tool that made them more efficient.

Fast forward to June 2026, and GitHub announced a change to its subscription model, moving toward token-based pricing. In practice, this means users are charged based on the amount of AI-generated traffic their interactions consume.Since software projects can contain millions of lines of code, charges measured in cents can quickly accumulate into thousands of dollars when analyzing large repositories. Mind you, this is not a guarantee that the code will work; it is simply a control and assistance service.

Given that AI remains a difficult business to monetize profitably, the move is not only understandable but perhaps even reasonable from a commercial perspective. Nevertheless, it made the new rules of the game abundantly clear: you are free to use our tools, provided you can afford the admission fee.

The reaction was significant. Many began predicting the end of “vibe coding” barely two years after the trend had emerged—not necessarily a bad outcome, if you ask me. 

More importantly, however, the change shifted the discussion toward a different question: What happens when a third party controls a critical part of your value proposition? And perhaps more importantly, what happens when that third party decides that you are no longer profitable enough for them?

I suspect many companies will soon discover whether they own a capability or merely rent access to it. For years, the software industry has served as a cautionary tale about becoming dependent on proprietary systems. AI may simply be pushing the rest of us to learn the same lesson.

---

The Pocket AI Guide is out!

📙 Amazon US: https://a.co/d/gCHHDax
📗 In Europe Amazon Germany: https://amzn.eu/d/3cmlIqa
(Available in other stores Amazon stores too in Europe)

Check the free resources in this website!