On (not) using Django in 2026

07 Jan, 2026

Here's a question:

"What will generative AI do to all the tools that help you set up a project quickly and bolt on features to those projects? Will generative AI offer competing, higher-quality bootstrapping and feature-bolting, or will generative AI complement those tools by making them even faster and more feature-rich?"

That's a tough question, and I don't know the answer to it. So I'll start with a more specific one:

"What will generative AI do to Django specifically?"

...and I don't even know the answer to that, so I'll narrow it further:

"Do I plan to use Django in 2026?"

Now I do know the answer: No. I've used Django a whole lot, but I have no plans to use it again in 2026 (or ever). What it's good at is less necessary, both for me (as I learn more) and in general (because of generative AI). Here are some reasons previous versions of me have used Django:

1. To keep from having to query the database directly.
2. To have a working authentication system immediately.
3. To avoid the details of hosting and serving a project (by using a "ship your Django project here and we'll serve it for you" service).
4. To avoid implementing a routing system.
5. To get a very nice admin portal, with search and CRUD operations over the database, out of the box.

Now I know a lot more than I used to, and I can use LLMs. These are all non-problems¹. Meanwhile, the downsides of Django are much less amenable to generative AI mitigation:

1. Coupling persistence and models (in the sense of Django models.Model objects) tightly causes big problems.
2. I like using non-relational databases, especially DynamoDB, in many² contexts; depending on the details of the situation, Django supports this thinly or not at all.
3. I've always eventually outgrown the built-in auth system.
4. As soon as you use React (or any other non-Django front-end system), you lose many of the simplicity benefits of Django.

Generative AI doesn't much help with the latter problems, except indirectly, by helping you build reasonable alternatives to those aspects of Django much more easily. But once you have those alternatives, there's much less reason to use Django at all.

The emergent story seems to be: accepting some long-term structural problems for many short-term conveniences is a lot less worth it than it used to be. Generative AI can replace the conveniences, and usually won't mitigate the structural problems.³

That's not to claim that nobody will, or that nobody should, use Django.⁴ But if I had to guess, I'd say that this case will generalize pretty well. So, I'd guess that generative AI, despite being good at helping people succeed with Django, will be even better at helping people succeed without it.

#potofhoney #Django #Python #generative AI