letterpress blocks saying AI Anxiety

Beyond the Hype: Navigating AI Anxiety and the Future of Development

There is no doubt that AI is changing the world, particularly within the web development community. From rapid app prototyping to the more peripheral world of image generation, the initial launch was incredibly exciting. However, lately, I’ve noticed a significant psychological shift. We seem to have split into two camps: those who are embracing the change, and those suffering from AI anxiety.

I recently saw a prominent open-source PHP framework announce they were “done” – not that they were quitting, but that they had reached a point of total acceptance. AI is on the scene, and it is here to stay.

Conversely, there is a growing school of thought, one I encounter more and more lately, suggesting that AI might inadvertently create a digital echo chamber of “dumb users.” The worry isn’t just limited to our industry; it’s the fear that by relying on these models, we’re breeding a generation of people, developers included, who can produce a result without actually understanding the process or the logic behind it.

Personally? I think that’s a bit of a reach.


The Builder’s Analogy: From Mud Huts to Steel

I’ve been coding for over 30 years. In that time, I have navigated countless syntax changes, frameworks, and languages. To me, AI is simply another paradigm shift. We can look at this through a historical lens:

  • The Mud Hut Era: Once, we built houses out of mud and straw with primitive tools.
  • The Modern Era: Then came bricks, steel, and heavy machinery.

The introduction of a crane didn’t make the architect obsolete, nor did it make the builder “dumb.” It allowed them to build higher and faster. You still need to understand structural integrity; if you give a master builder power tools, they’ll build a cathedral. If you give those same tools to someone who doesn’t know a load-bearing wall from a window pane, the house is probably going to fall down.

AI is no different. It won’t replace 30 years of experience; it will enhance it. You can see more of my thoughts on the evolving landscape of technology over at elliottrichmond.co.uk.


A Tool for Every Skill Set (Not Just Code)

This brings me to a more personal example of why I’m optimistic. Historically, when I wrote blog posts, I would always pass them to my wife and business partner. I’ll be honest: I’m dyslexic, and writing copy has never been my strongest skill. She would help me restructure my thoughts and polish the grammar so they were ready for the world.

Now, I have a different tool in my kit.

Full disclosure: this post was polished by AI. However, I didn’t ask the AI to generate the ideas or the concept. I recorded my thoughts as audio, transcribed them to text, and gave the AI very clear instructions to take my “brain dump” and structure it grammatically.

This is the perfect example of using AI to enhance a skill set. It allowed me to get an idea out of my head and onto my blog without the usual friction. It isn’t just about code; it’s about using the right machinery to do a better job, regardless of the task at hand. At the end of the day, it’s just a tool – and I’m quite excited to see what we can build with it.


  1. Mark Wilkinson Avatar

    I think the different between AI and your crane example is that AI does “understand structural integrity” and it does have 30 years experience.

    And, if it doesn’t quite understand them yet – well it will, very quickly.

    Yes, at the moment it is still not as good as the architect for example, but the rate of change is so fast that it is inevitable that it will be that good soon.

    When that happens, this is the real worry.

    1. Elliott Richmond Avatar
      Elliott Richmond

      Obviously you are right. I think in my analogy I’m trying to get across the fact that these are just ‘new’ tools; in the right hands they can do a really good job, but in the wrong hands they will likely do a bad job. Currently, we have models that simply make stuff up and convince you they’re correct, whereas a trained eye would clearly spot an error or where a hallucination might have occurred, that was my point. Will AI continue to make stuff up without context in the future? Probably, I’m sure, but that still isn’t going to be obvious to the untrained eye. The tool simply does the job quicker; ‘the prompt’ (or the way you use the tool) is where the expertise comes into play, as it is the expert who in turn checks how well the tool has performed.

  2. Grégoire Noyelle Avatar

    Thank you for this article.

    I use Claude Code a lot — it’s an amazing tool. And like any tool, it’s how you use it that makes the difference.

    My dev experience definitely gives me more control and allows me to take a more architectural approach.
    But it doesn’t stop there: I learn much faster with Claude than before, and it’s even Claude that writes my notes in Markdown and keeps them updated.

    One last point: the real skill lies in structuring Claude’s reasoning in a certain way. It takes time, but the long-term time savings are huge.

    Ps: which tool do you use for comments notifications ?

    1. Elliott Richmond Avatar
      Elliott Richmond

      Thanks for the feedback, much appreciated. For comments and notifications, it’s just WordPress core.

Leave a Reply

Your email address will not be published. Required fields are marked *