Claude is Becoming the Dominant AI Tool. Should That Worry Us?

At a major AI industry conference in San Francisco earlier this month, 6,500 executives, founders and investors gathered to talk about the state of AI. One thing became clear very quickly. ChatGPT no longer dominates the conversation. Claude does.

The CEO of enterprise AI company Glean described it as "Claude Mania." He said: if you asked anyone in that room which single AI tool they'd want, the answer would be Claude. He called it a religion.

I understand why. I use Claude almost exclusively. I do use Google AI Studio as my starting point for builds - which is Gemini under the hood - but in almost every other way, Claude is my default. I've tried ChatGPT. But I keep coming back to Claude because, for the kind of work I do, it's simply better. The quality of the output, the way it reasons, the way it handles complex tasks - nothing else has consistently matched it.

But lately I've been sitting with a question I can't quite shake.

Is it wise to go all in on one tool?

The WordPress parallel

Here's the thing. This isn't actually a new question.

Take WordPress, for instance. For years it has been the dominant website platform. Designers, developers and business owners built entire careers on top of it. WordPress is open source - free for anyone to use, modify and build on - and that has always felt like a guarantee of stability. Nobody owns it in the way a commercial company owns a product. The code is yours to use.

And yet even WordPress, the bedrock of 40% of the internet, had its moment of instability. In 2024, co-founder Matt Mullenweg publicly called hosting company WP Engine "a cancer to WordPress," accusing it of profiting from the platform without contributing sufficiently back to the open source community. He blocked WP Engine customers from accessing WordPress.org resources - the central repository for updates, plugins and themes - leaving over a million sites suddenly exposed. It ended up in court.

The point isn't that WordPress failed. It didn't. The platform remained stable, the community held and the legal system intervened. The point is that even a mature, open source, community-owned platform - the kind of platform that feels like infrastructure - can experience unexpected turbulence when the people and organisations around it fall out.

Claude is not open source. It is a commercial product owned by a private company with its own investors, principles and pressures. And those pressures can have real consequences. Earlier this year, after a dispute over how the military could use its technology, the US Department of Defense declared Anthropic a "supply chain risk" - a designation previously reserved for foreign adversaries. The case is still in court.

That situation is specific to government and defence. But it raises a question worth sitting with. What if you woke up tomorrow and Claude was inaccessible? What would that mean for your business?

And then there's the pace

The AI landscape is moving at a pace that has no real precedent in the history of technology. What was the best tool six months ago, or even three months ago, might not be the best tool today. The gap between the leading model and the rest can shift dramatically in a matter of weeks.

Anthropic, OpenAI, Google, xAI, Meta - and increasingly, well-funded challengers from China - are all in a race. And races have unexpected turns. A breakthrough from one competitor, a pricing change, a policy shift, a strategic pivot - any of these could change the trajectory quickly.

We've already seen hints of this stratification within Claude itself. Anthropic recently announced Claude Mythos Preview - a model so powerful they've limited access to roughly 50 companies due to its advanced capabilities. Most of us are using one version of Claude while a select group of enterprises accesses something significantly more capable. The tool isn't standing still. It's evolving in ways that aren't always visible to the people building on top of it.

The stack question

There's another dimension to this that I've thought about a lot, particularly when building solutions for clients.

There's a strong argument for picking a stack and going deep. Choose your tools, learn them properly, build your workflows around them and become genuinely excellent with them. That focus compounds over time. Trying to stay fluent across every platform as they all race to improve is exhausting and probably counterproductive.

I've operated both ways. I spent years running a tech implementation agency where we supported every platform going. And I've also been someone who picks a tool, commits to it and gets really good at it. Both approaches have merit. Both have drawbacks.

What I actually think

I'm not going to tell you to stop using Claude. I'm certainly not going to stop.

But I do think there's a question worth asking - one that most people who are deep in Claude haven't stopped to ask properly. If Claude changed its pricing significantly tomorrow, what would happen to your business? If a competitor released something demonstrably better next month, how easily could you shift?

And here's the harder question that sits underneath both of those: would you even want to?

Because that's the thing about a fast-moving space. If you shift every time something new and shiny appears, you end up in a game of ping-pong - constantly migrating, constantly rebuilding, never quite settled. There's a real cost to that. The muscle memory you've built, the workflows you've refined, the way you've learned to get the best out of a particular tool - that's not nothing. Abandoning it every six months isn't agility. It's just noise.

So maybe the question isn't "should I diversify away from Claude?" Maybe it's "am I making this choice consciously, or have I just drifted here?" Dependency on a tool isn't inherently bad. Unconscious dependency, the kind where you haven't even noticed it happening, is worth examining.

The WordPress analogy might hold. Claude might be the dominant, stable, long-term platform that founders build on for the next decade.

But WordPress wasn't moving this fast. And the companies racing to displace it weren't this well-funded or this capable.

We should be asking the question. Even if, especially if, we don't have a clean answer yet.

Building with AI and thinking through your stack? Come and join the conversation at Vibe Coding Lab - a free community for founders who are figuring this out together.