# In the AI era, stack debates are dead

Published 18 May 2026

[Originally posted on LinkedIn](https://www.linkedin.com/feed/update/urn:li:activity:7462112759053471744/)

**In the AI era, stack debates are already dead**.

Engineers love arguing: microservices or monolith, TypeScript or Python, Rust or Go. For years I just picked whatever fits the business better. 🤷 Now, working on my own CLI tool that moves the source of truth from code to specs, I’m fully convinced: **in the AI era, these holy wars are irrelevant**.

If code stops being the source of truth, the importance of technologies falls away. You essentially don't care on what this gets implemented. What you actually need is to learn how to explain to AI what's required of it. On what technology AI does it — that's a tenth-rate question, because it can do it on basically any of them anyway.

My guess is we'll end up with one or two — maybe three — AI-native technologies, where a specialist picks the one that fits what the product is supposed to do (Nanolang is the early example). Languages where AI makes fewer mistakes when generating code.**The choice gets driven by where AI makes fewer mistakes. Not the human, specifically AI**.

This is a mindset shift, and from what I see, 98% of people still think differently — it's a relatively new idea in the mainstream. I don't quite know how to say this gently, so I'll say it plainly: what you know in any current language today will become unimportant a year from now. You'll need to learn one of these AI-native technologies and how to explain things to AI in it. The language part it'll figure out on its own. 🐍

**Honest caveat**: AI can solve any task you give it — that part I'm confident about. On which stack it'll solve it cleanly — there I'm still not sure. Maybe on PHP it won't quite manage. 🤔