Comment by echelon
2 days ago
Does anyone need Drupal anymore? You can just Claude Code all kinds of shapes of software.
I don't see how Drupal or anything like it survive. These types of softwares seem fitted to 2006, not 2026.
2 days ago
Does anyone need Drupal anymore? You can just Claude Code all kinds of shapes of software.
I don't see how Drupal or anything like it survive. These types of softwares seem fitted to 2006, not 2026.
Using Claude to build things quickly is great - but once you hit a certain level of scale, or have a need for governed workflows & oversight, you're more or less going to want some sort of CMS. When that happens, you can either go it alone, or build on top of a lot of well discussed well thought through solutions to these sorts of problems that the Drupal community has dealt with for a couple of decades.
> you can either go it alone, or build on top of a lot of well discussed well thought through solutions to these sorts of problems that the Drupal community has dealt with for a couple of decades.
That isn't needed anymore. One person can maintain an entirely in-house solution and easily make changes and adaptations.
Not nearly enough time has passed to make any conclusions about long-term maintainability of LLM-written code.
Have you tried letting AI agents make design decisions on their own? (technical or otherwise), in my experience you still need to properly understand what you want to get good results - you would need to have a deep understanding of all the potential warts, security holes, usability issues etc - developing a CMS via AI is still a substantial project.
>You can just Claude Code all kinds of shapes of software.
Drupal is used at large organizations with hundreds of sites, thousands of users, and multiple more-or-less coordinated development teams. It's actually a great target for agentic coding, because it has such a well-established internal API. It can definitely feel like an anachronism, but I think there are few projects that could replace some of it's use cases. It's the B-52 of CMSs.
not a whole lot of documentation, I feel like a lot of knowledge is held within orgs. At my first job I often had to venture into the source code to figure out how something is supposed to work. Copilot has just came out then and it was terrible at helping with a Drupal 10 migration for example because the training data only had 8 and below and a lot of stuff obviously changed in-between.
I reckon AI must be much better now at this since it's doing all sorts of things, but I wonder if you still would need to have the repo cloned so that it can crosscheck with the source code. Since the knowledge cutoff should be around 2025, most of the stuff it knows should be plenty
This is how I view it. If you're going to be hosting dozens or hundreds of sites having a consistent platform like this is important even if Claude Code is doing all of the implementation work. It has understood and workable flows for things like content approval, timed releases, etc. All the things people think about when they talk about "managing content" and for the most part organizations still want human in the loop, especially for gated approvals and auditability.
If I'm thinking of building a single one off site, then yeah I'd probably opt to just let CC handle it without worrying about a framework which is overengineered for my current need. That's what I do with most of the personal projects I work on these days. But I do have a skill to build consistency around the parts that should be consistent for everything I host. Like package shape and types of hooks available for deployment. How sites are publicly exposed and DNS management.
> Drupal is used at large organizations with hundreds of sites, thousands of users, and multiple more-or-less coordinated development teams
Sadly none of this applies to me, so I wonder whether it makes a good argument when my question is "should I use Drupal for my personal hobby website".
I'd say no, at least to start - there's an inherent cost in using a CMS a framework etc. If what you're building is simple enough that a flat file hosted on cloudflare etc, works, then start there.
Once you find you do have a need for something more complex (i.e. you have a team, want to give agents governed access etc), then it's pretty easy to go from a flat file to something more complex.
Claude still has to write software in a target language, often using an established framework, especially one with a lot of examples.
PHP is, in fact, exceptionally good as a language for creating websites.
A web accessible PHP runtime is, in fact, a giant security vulnerability and always had been.
lol my defense is still timely I swear!
also the best code is usually less code, and frameworks can be really helpful there.
One of the appeals of LLM-generated code, to me, is that you can often write things without a framework and reduce the total amount of code in the system, which tends to make it much more understandable overall.
2 replies →
Frankly everything is now. Anyone can use Rust or Go for this now.
One of the best things about Drupal (years ago, when I used to write plugins and themes), was that you could completely customize the backend, as well as the frontend, and that was a good thing, if you have low-tech users that author content. I despise having to walk people through WordPress' backend.
But that rewriting the base system, every couple of years, was why I dropped it.
Nowadays, with LLMs, it may be time to look at it again...
>But that rewriting the base system, every couple of years
To be fair, they did it once, from v7 to v8. But yeah, a major release does break APIs. I'm concerned with how Drupal Canvas is bolted on, but it looks like it's basically a content editor UI that is separate from the site admin UI, which is a good idea.
I think from v5 to v6, as well. v6 was the last one I worked with (been awhile).
Also, there were a whole plethora of extensions (not just "plugins"). Some of those broke, while others, did not.
As someone who has recently started Pair Programming with LLMs—something I never managed to bring about with a human, sadly—I'd like to say: Sure you can do it. Sure, team up with a bunch of friends, a forest next door and build your dream cabin. Then compare that to a prefab from the Sears Modern Homes line. It's still 30'000 parts, 25 tons, est. 90 days build time, but it will be vastly easier, especially if one of your pals happens to be great at digging up and bringing together assembly instructions, general technical infos and past experiences from other builders for all the many snags you'll still hit when erecting that thing.
Can't answer the specifically "Drupal" part (as for me I once had a certain pressure behind me to use it, found it atrocious at the time and won't look to repeat that; if their software is much better / much more to my taste now, good on them). But in principle, at least from a certain app size upwards, I'd rather start with something working in principle, then improve on it.
That said, it might still be more difficult to get things right with Drupal than with some other softwares where you have better continuity between versions. At least traditional search engine results tend to handle version changes rather poorly meaning you'll get tons of advice that don't really apply anymore because they're years old. That happened a lot to me when trying to help out a friend with a WordPress installation.