Web Directions Year round learning for product, design and engineering professionals

This week, there’s just a limited number of articles for your reading. Partly—and no small part indeed—that’s because we’re currently in the final stages of producing three conferences in about ten days starting next Wednesday. In these stages, you tend to get very busy with all kinds of things that just need to be done. We’re also organizing something new, big and really exciting for 2026 that will be announced on November 24th. So my attention has been somewhat elsewhere, meaning I tend to read fewer of the blogs I might typically follow and spend a little less time on social media, so I come across fewer articles than I might in a typical week.

But there’s something else you might notice from the articles we’ve gathered this week: none of them are about web design or development or product design. They’re all about AI, particularly with a focus on software engineering.

A Shift in Focus

Thinking back over the last three or more years, that’s increasingly been the case more generally. Yes, we continue to post a lot of things around front-end design and development, product design—the things that have taken up the majority of our attention over the last 20 years or so. But the impact of AI on software engineering in particular, and also on design, has become an increasing focus for us.

I first started focusing on the web in the early 90s, well over 30 years ago now. I’d already been writing software for the better part of 10 years by then—while still at high school, then studying computer science at university, and then as a fledgling developer with my own product.

While it took me a year or so to recognize the impact of the web—on both the field of software development and more broadly on the world—from about 1995, I came to believe its impact would be very significant, and that it was where I felt I should focus a lot of my attention and energy above all else. So I kind of walked away from the product I’d been building for two or three years, that we’d been selling online, and started really focusing on the field of web development.

Not so much as a web developer, although I did quite a lot of that, but as someone building software tools for web developers, developing training and resources for the fledgling field of web design and development, which at the time were very much intertwined. If you were the web designer, you were likely the web developer, you were probably what we then called the webmaster. You probably wrote a lot of the content.

This evolved over time into online workshops, into in-person workshops, and ultimately into our conferences. All of which—the through-line, the arc of the last 30-plus years since, let’s say, 1995—has been helping the field of web design and development grow, and helping professionals within these fields develop their capabilities.

The End of an Era

As I talked about in an essay I wrote last week, I feel that’s something that’s coming to an end. Or at least to a hiatus. I feel the field of web development in particular has atrophied. We have come to a local maximum, of Next.js, and React, and a handful of other frameworks and libraries. Frontend development is now effectively React development. What we develop, indeed what we can imagine, is constrained by the capabilities of the assocated abstractions, despite the ever-growing sophistication of the underlying web platform.

Perhaps things will change there. I don’t know. I’m not optimistic, as I wrote last week. But what ironically I’ve also realized is that I have been optimistic, if that’s the right word, about something else that has been happening in the field of software development. It’s not entirely unrelated to the web. It will have a huge impact on the web, how we develop for it, what we develop, but there is no getting away, however we might feel about it, from the impact that AI and large language models are having. The practice of software engineering has already transformed. In particular, its economics has transformed, as Software Development in the Time of Strange New Angels explores. I really highly recommend you take a few minutes to read that.

But more broadly, it’s the nature of computing itself that is transforming. Were moving from deterministic to probabilistic systems, from CPUs to GPUs. The nature of human-computer interaction, which includes all the abstractions and metaphors we use about computers, is changing.

One funeral at a time

What I’ve noticed is that in both design and development it’s often very experienced professionals—those who’ve been working in the field for decades—who are some of the most thoughtful voices about the adoption of these technologies. I think of Simon Willison in particular when it comes to software development. I think of Swyx (Shawn Wang), who was heavily involved in the JavaScript and TypeScript world before coining the term “AI Engineer” and creating an extraordinary community and event series around that emerging field.

In design, I think of Scott Jenson, who has worked for decades on everything from the original Apple human interface guidelines through to distributed web ideas at Google. I think of Dan Saffer, Luke Wroblewski, and Josh Clark—all incredibly experienced professionals in the field of design who are engaging with the emerging challenges and opportunities of large language models.

Max Planck, the originator of quantum physics, once rather cruelly observed that “science advances one funeral at a time.” The essence of his observation is that typically once we get locked into a paradigm, say classical atomic physics or object-oriented programming, it is almost impossible to shift our thinking. Thomas Kuhn‘s famous work on the philosophy of science, The Structure of Scientific Revolutions–which is where the term “paradigm” in this sense comes from–observes that it’s typically new entrants to a field, younger people who bring novel ideas and new ways of thinking, while existing practitioners are wedded to their existing perspective, to the thing they’ve invested so much of their life in.

But with AI and large language models from the perspective of software engineering as well as design this doesn’t seem to be the case. Not for a moment do I think younger, newer entrants into our fields are resisting AI and large language models. But while there is definitely resistance—indeed, outright hostility at times—from some experienced designers and developers to these technologies, there’s also an enthusiastic adoption and engagement with them by folks like those I mentioned, closer to the end of their careers than the beginning. I certainly feel I fall into that category.

Echoes of the Early Web

I think too that there are aspects of the earlier web that find echoes in what’s happening today. Which might cause some old web hands to be aghast, but I think it’s incontrovertible. The web as a field pulled itself up by its own bootstraps. Even today, very little of the field is covered at universities and in formal education. The mid-to-late 90s and then into the following decade saw the emergence of communities around designing and developing for the web. Practices that ended up becoming standardized or even part of the browsers (like doctype switching) emerged in these communities. This came from people blogging and later posting on social media. And before all that, posting on newsgroups and mailing lists. More traditional channels of education never really caught up. Even today, almost none of this is really addressed in universities or other kinds of formal education.

I feel something very similar is happening around software engineering and large language models. There is a sense of community, despite it being a hyper-capitalistic field in some ways, with huge amounts of venture capital money pouring in—in a way that was never true for the field of web development. There are podcasts. There are blogs. There’s endless posting on social media. Papers published on arxiv. Models and experiments on Hugging Face. Code on Github.

This is what I’ve found myself drawn to more and more over the last three years or so. The code I write continues to include a lot of HTML and CSS and JavaScript. I continue to embrace the newer features of the platform like cascade layers and view transitions. But I find that to be an increasingly diminishing part of the code that I write. Just as I find the field of web development a diminishing part of my focus. The code I write is increasingly developed with large language models.

The Economic Reality

Whether you like it or you don’t, if you write software, you will be writing it with large language models and code generation tools. Or slowly then suddenly, you will be completely marginalized economically.

If your hobby is to continue to write bespoke hand-tooled HTML, CSS, and JavaScript, no judgment, but it will be just that—a hobby. Economically you will not be able to compete. It will be like woodworking or baking: something you enjoy doing, but certainly not in and of itself that will enable you economically to keep up.

And if you think large language models just generate the average of what they have been trained on, and the quality will be bad, that it will be inaccessible, insecure, poorly performant—well, that’s already not the case. Yes, these tools need to be steered, to be guided toward the right outcomes, and that’s where your knowledge and capability comes in. But the code they produce is already very often acceptable at the very least.

What matters is not that you type an angle bracket or curly brace. It’s that you understand how to write secure code. You understand how to write performant code, how to identify performance bottlenecks, how to fix those. You understand how to write accessible code, how to identify when something isn’t accessible, how to address and fix that.

What This Means for You

So, what are the conclusions from all this? Well, firstly, that your knowledge still matters, and continuing to improve it matters more than ever.

But also, armed with the capability of these technologies, good developers won’t just specialize in narrow fields. They will specialize across fields. Small teams and individuals will produce what large teams and entire organizations were necessary for not long ago.

There’s a need to continually develop your capabilities and understanding and knowledge. Not just of the particular area of practice that you focus on, but also importantly what is happening with these code generation tools. How are they evolving? What practices and patterns are emerging? What’s still nascent but transforming at a pace that we simply haven’t seen before, not just in software engineering, but probably in any field ever (perhaps Los Alamos in the 1940s matches it)?

Despite my being much closer to the end of my career than the beginning, I’m genuinely excited about this. It’s what I’ve been increasingly focusing on. And what I really encourage you to do, regardless of your area of expertise—whether it’s front-end or back-end, whether it’s operations or mobile. Focus on working with these tools, learning them, understanding them, and then sharing those ideas. Just as we built the practice of web development over the last 3 decades.

This week’s reading

Software Development in the Time of Strange New Angels

AI Native Dev economics LLMs software

Five months ago, my lifelong profession of software development changed completely. My profession was born in the 1940s, created to help fight demons. Our first encounter with the strange new angels of agentic AI is changing every aspect of it. Hardly anyone has noticed yet.

Source: Software Development in the Time of Strange New Angels

Every software developer, everybody involved in the delivery, development and design of software-based products should read this. It’s deeply insightful.

How I use AI (Oct 2025)

AI AI Native Dev LLMs

I believe I use AI fairly typically for a software engineer, if slightly more than average. I’ve been working in the AI space since slightly before the announcement of GPT-3 in 2020.

I use AI in a few particular ways:

• Coding
• Research & search
• Summarization & transcription
• Writing
• Art & music

This is, of course, purely my opinion & speculation. If you have different ways of using AI, I’d love to know!

Source: How I use AI (Oct 2025) – Ben Stolovitz

Another “how I use AI” type post which we have been collecting because it’s early days, and seeing how different software engineers use the technology is a very interesting way to expand our own possible uses and use cases.

You Should Write An Agent

AI Engineering AI Native Dev autonomous agents LLMs MCP

Some concepts are easy to grasp in the abstract. Boiling water: apply heat and wait. Others you really need to try. You only think you understand how a bicycle works, until you learn to ride one.

There are big ideas in computing that are easy to get your head around. The AWS S3 API. It’s the most important storage technology of the last 20 years, and it’s like boiling water. Other technologies, you need to get your feet on the pedals first.

LLM agents are like that.

People have wildly varying opinions about LLMs and agents. But whether or not they’re snake oil, they’re a big idea. You don’t have to like them, but you should want to be right about them. To be the best hater (or stan) you can be.

So that’s one reason you should write an agent. But there’s another reason that’s even more persuasive, and that’s

It’s Incredibly Easy

Source: You Should Write An Agent · The Fly Blog

This, in just a few minutes, won’t just give you an understanding of what agents are but will enable you to build your first one and understand what some of the fuss is about. You might also ask why you need MCP after seeing this.

When Everyone’s a Developer, How Do We Promote the Web Platform Over React?

AI front end LLMs react

2025 is a strange time to start a newsletter about web technology. The past four editions of WTN have focused on the intersection of the web and AI, because frankly that’s where most of the excitement is on today’s internet. But as a few people I link to this week point out, the web platform has improved to a point that it now does much of what frontend frameworks do. So why isn’t there as much exciting activity to report on regarding the web platform? The problem, I think, is that web platform improvements are being undermined by AI development trends. I see two main issues:

I’ve heard both Vercel and Netlify (two of the leading web developer platforms) say in recent weeks that their user bases are massively increasing. Why? Because of vibe coders. The definition of a “developer” has expanded to include people who rely on prompting rather than programming. But the problem is that vibe coders get React solutions instead of web-native ones. What’s happening is that vibe coders ask their magic lamps to build an app or agent, and the AI genie gives them a React app.

The leading large language models, like GPT-5, are defaulting to React and Next.js when asked to create web apps or sites. That entrenches the power that React has on the web development ecosystem, which means web platform improvements aren’t being utilized by AI. Which leads to point #2…

Source: When Everyone’s a Developer, How Do We Promote the Web Platform Over React?

This is a really good round-up of some recent articles, many of which, if not all of which, we referenced here in the last few weeks. It also includes a reference to something I wrote recently in a similar vein. This is a real time of transformation across all of computing, and front-end isn’t alone in that. Where exactly it ends up? It’s impossible to say.