One of the greatest things about programming in the internet age, is how you can learn to code from anywhere in the world for free, thanks to the myriad of resources available online. Blogs, forums, Stack Overflow, FreeCodeCamp, YouTube tutorials, open source projects. The list goes on. You just need a computer and an internet connection and you can become a programmer.
Yes, developers with better socioeconomic status have always had advantages: formal education, access to expensive courses or certifications, more powerful computers. The playing field was never completely level. But the differences were manageable. A motivated developer could overcome them with time, effort, and free resources.
In the age of AI, this dynamic is changing.
A junior developer in India spends hours debugging a complex React state management issue. Their colleague in San Francisco asks Claude to analyze the code and receives a detailed explanation with three solutions in seconds. Both developers have the same skills, same dedication, same work ethic. The only difference? One can afford $20/month for AI tools. The other cannot.
The Cost Barrier
Currently, AI tools are expensive, and the costs can add up quickly. A basic tier for Claude or ChatGPT will cost at least $20/month. And if you are serious about development, you might need to go to their upper tiers, or else you will get out of tokens quickly. Claude Code Max plan costs at minimum $100/month.
This can get even more expensive in a multi-model world, where different models are better suited for specific tasks. Now you might need a Claude Subscription for coding and a ChatGPT Subscription for research or planning, for example.
Self Hosting is possible, but difficult. While there are some good Open Source Models, they are not yet at the level of commercial models, like Claude. And you would need a big investment in terms of hardware and infrastructure to host them. And with the rising prices of GPUs, RAM, and storage, self hosting is getting prohibitively expensive.
While these costs may seem insignificant for a developer in the US or Western countries, they can be a significant burden for developers in developing countries.
The Productivity Divide
You might argue that AI is just another tool, and developers can still code the “old way” without AI assistance. This is technically true, just as it’s technically true that you could still write code in Notepad instead of VS Code.
But here’s the critical difference: when VS Code became the standard, free alternatives existed. When Git became essential, it was open source. When modern frameworks emerged, they were free to use.
AI tools creating a productivity gap of 2-5x have no free equivalents. The free tiers of Claude and ChatGPT have rate limits so strict they’re unusable for professional development.
The impact of AI on free access to knowledge
There is a nefarious side effect of the AI era that should not be underestimated, that can also contribute to the increase of inequality.
With aggressive scraping of online content and complete lack of attribution, many creators have starting to gate their knowledge to protect it from being exploited without credit. This leads to more content behind paywalls, private Discord/Slack communities instead of public forums, fewer high-quality blog posts and tutorials, and declining participation in open knowledge platforms.
Stack Overflow being almost dead is a great example of this. While this decline was already happening for some time, even before the AI boom, it’s undeniable that AI have greatly accelerated the trend. As people become used to getting answers from AI chatbots, there is a reduction of incentive to contribute to public knowledge platforms.
This creates a vicious cycle: less free knowledge available → more dependence on AI tools → greater inequality → even less knowledge sharing.
The next generation of developers may face a world where both the knowledge and the AI tools to access it are behind paywalls.
Conclusion
Everything is evolving so fast, that it´s hard to predict how big the impact of AI on the future of software development will be.
Independent of what your opinion about the usage of AI tools and their real impact on software development might be, it’s undeniable that it´s here to stay.
Personally I don’t believe in the more extreme vision that AI will completely replace human programmers. I believe in using AI as an assistant to automate common tasks and as a productivity booster, so that humans can focus on more complex challenges. While I personally love to code and to understand how things work at a deeper level, the end goal of using technology should always be to solve people´s problems and to make the world a better place. If the usage of AI tools can make humans more efficient and productive on that regard, so be it.
But we must keep the ideals of the open web alive! The internet democratized software development, by allowing millions of people worldwide to learn to code using free resources, launching careers and building products that changed the world. We risk losing this progress if we allow AI tools to create a two-tier system that widens the developer inequality gap.
The cost of AI tools may decrease over time as technology evolves. better chips, more efficient algorithms. We’ve seen this happening throughout computing history. But we should remain alert.
Building software should not be a privilege reserved for those who can afford premium AI subscriptions. It must remain accessible to anyone.