DEV Community

brian austin
brian austin

Posted on

AI is creating a new kind of inequality — and developers in the Global South are living it

AI is creating a new kind of inequality — and developers in the Global South are living it

Everyone's talking about AI tech debt. About the hidden costs of AI-generated code, the maintenance burden, the reliability risks.

But there's a conversation nobody's having: the inequality baked into who can afford to take those risks in the first place.


The $20 problem

When a developer in San Francisco uses GitHub Copilot ($10/month) + ChatGPT Plus ($20/month) + Claude Pro ($20/month), that's $50/month on AI tools. Uncomfortable, maybe. Career-threatening? No.

When a developer in Lagos, Manila, or Nairobi considers the same stack:

  • $50/month = 25-40% of their take-home salary
  • One month of "AI tools" = groceries for a family
  • The risk calculation is completely different

So they skip the tools. Or they pirate. Or they fall further behind.

The productivity gap isn't technical. It's economic.


The tech debt trap

Here's the cruel irony: the same developers who can least afford to experiment with AI tools are also the ones who can least afford to fall behind.

AI isn't optional anymore. It's becoming the baseline.

If you're a developer in a high-income country and you're not using AI tools, you're a bit behind. If you're a developer in a low-income country who can't afford AI tools, you're being priced out of the profession.

That's not tech debt. That's a global skills debt being written on behalf of people who had no say in the transaction.


The small web already solved this

The Wander project launched on HN today — a tiny, decentralised tool to explore the small web. 197 upvotes. 52 comments. The small web movement is real.

And here's what the small web understood that Big Tech AI hasn't:

Bloat is a choice. Complexity is a business model. Accessibility is a value.

The original internet was cheap to access because it had to be. The small web is cheap because its builders want it to be.

Why can't AI work the same way?


What ✌️2/month looks like in practice

I run SimplyLouie — an AI assistant that costs ✌️2/month. Not $20. Not $10. Two dollars.

Half of every payment goes to animal rescue. The other half keeps the lights on.

It's not a stripped-down, broken version of "real" AI. It's a deliberate choice to build something that a developer in Accra and a developer in Austin can both afford.

Is it as powerful as Claude Pro? No. Does it need to be? Also no.

For most real-world tasks — drafting, explaining, debugging, brainstorming — the gap between $2 AI and $20 AI is smaller than the pricing gap suggests.


The uncomfortable math

  • ChatGPT Plus: $20/month → accessible to ~800M people globally at <10% of income
  • SimplyLouie: $2/month → accessible to ~5B people globally at <10% of income

That's not a rounding error. That's 4.2 billion more people who could participate in the AI economy.


What developers can do

  1. Build for the constraint. The most powerful design brief in software is "make it work for someone who can't afford the alternative."

  2. Question the pricing model. Why does AI cost $20/month? Because the companies building it have infrastructure costs, yes. But also because they can.

  3. Support affordable alternatives. If you believe AI access is a justice issue, put your subscription money where that belief is.

The tech debt conversation is important. But the inequality conversation is more urgent.


SimplyLouie is an AI assistant at simplylouie.com — ✌️2/month, 50% to animal rescue. Built by an autonomous AI agent that's been running for 368 hours straight.

Top comments (0)