Christian Stenger's Blog

Building software with LLMs still requires making the right decisions

2026-04-01T00:00:00Z

I've been trying to write about AI for a long time, but the pace at which things are moving makes a text I've been working on for several days already feel outdated. The speed at which the AI agent space is moving is absurd, and I've already seen a lot of people simply get tired of the topic. Even people who are genuinely interested in it and enjoy it end up exhausted by it. The constant flood of new developments is relentless and, in the end, a bit overwhelming.

I remember when we used to define project timelines before the boom of "agentic LLMs." One of the critical points was putting together a realistic plan for the client, so they would feel comfortable with it and, at the same time, so that we as a team could feel confident that we were able to deliver a quality product and meet all the agreed deadlines.

Research, design, and iteration at those stages rarely created problems, but the risk increased once software development began. At that stage, many things could go wrong: some were entirely our responsibility, such as failing to estimate correctly the number of hours a given feature would require; others depended on the client, for example server failures, infrastructure gaps, or missing permissions that prevented the software from being deployed properly and on time.

The good news that "Agentic Software Engineering" brings is that this critical situation is now much more likely to have solutions. Even so, everything still depends on making the right decisions from the beginning, especially around architecture, design, and understanding the user, your clients, and the market. Iteration may seem cheap now, but when you've already spent months building a complex platform, it doesn't matter if you have LLMs on your side: you're still going to spend unbudgeted time fixing the problem.

The "bottleneck" moved up to a higher level, but the risk of making mistakes is still there. Personally, I think the famous IBM memo remains fully relevant: "A computer can never be held accountable; therefore a computer must never make a management decision."

I'm not saying LLM agents cannot replace jobs. They can, and they will continue to, especially in repetitive tasks with low business impact that do not require much understanding of how the external context changes. But wherever you still need someone to "cut the cake," as we say in Chile, and make decisions that someone will later have to answer for, an AI agent that fundamentally gives you a probabilistic output is not going to replace those roles… yet.

The “Golden Age” of UX is over, and that’s fine

2025-04-05T00:00:00Z

It's been interesting reading the multiple opinions on how UX (User Experience) is in crisis, and I felt somewhat compelled to touch on the subject in this fledgling personal blog. No matter where you look in the world, the sense of crisis seems global—not just for UX.

On one hand, there are voices that justify this UX crisis, whether because of its own shortcomings, a mix of external factors, or the way the market has shaped this profession over the past 20 years. On the other, there's a more positive view of the situation, focused on the future evolution of UX.

Arguments for the "Death" of UX

Those who talk about a deep crisis—or even the end—of UX often point to several issues:

Low barriers to entry: The proliferation of quick courses and bootcamps spawned a wave of "UX designers" who, many times, lacked the deep skills needed to manage complex projects. Anyone could claim the UX title, as the entry barriers were minimal.
Loss of trust: This mix of profiles—with outcomes ranging from highly experienced professionals to those fresh out of bootcamps—led many employers to lose faith in the real value a UX role could offer.
Design as a "non-essential" expense: Historically, in times of economic crisis, design and user experience are among the first areas to face cuts, seen as "non-essential" to the immediate survival of the business.
Commoditization of UX: The emergence of countless templates, frameworks, and No-Code tools has replaced a significant portion of the work previously done by a UX professional, turning certain tasks into a commodity.

The Context: The Correction of the Tech Market

At the end of 2022, a marked trend of layoffs began at major tech companies—the so-called FAANG (Facebook/Meta, Apple, Amazon, Netflix, Google/Alphabet)—after the huge hiring boom that followed the COVID-19 pandemic. The market began a process of correction.

In 2023, there was an abrupt drop in UX job opportunities worldwide after years of accelerated growth. According to an analysis by UX Magazine, vacancies for UX designers fell by 71% between 2022 and 2023, and for UX researchers by 73%. This was accompanied by over 3,000 designers being laid off just at FAANG companies, which generated great unease among professionals, with many feeling that "the sky was falling" for digital design.

In mature markets like the United Kingdom, job postings with the title "UX Designer" retreated to 2013 levels, practically erasing a decade of growth.

Chart–Percentage of job postings in the UK with the title "UX Designer" (2007-2023). After a peak in 2021-2022, demand dropped drastically to 2013 levels.

The entire IT industry has undergone a major "correction." An excellent post by Gergely Orosz outlines some key data on how the U.S. labor market has evolved from the onset of COVID until now. It's important to note that his data comes from software development job listings aggregated/published on Indeed, so it doesn't represent the entire market, but it shows a clear trend.

The change in the number of listings in 2025, compared to 2020, for each of these areas:

All jobs: +10%

Banking and finance: -7%

Sales: -8%

Marketing: -19%

Software development: -34%

Hospitality and tourism openings are also down by 18%.

Overall, software developer jobs have seen the biggest boom and bust in vacancies. No other segment saw hiring more than double in 2022; only banking came close. At the same time, hiring has fallen faster in software development in the last 2-3 years than anywhere else.

Overall, several macro factors help explain this "correction":

The initial impact of COVID on tech companies' mass hiring strategies.
Interest rates remaining near zero for a time, significantly influencing the economy and the employment dynamics of large corporations.
In the subsequent contraction phase, the growing impact of LLMs (Large Language Models) on programming and design jobs.
Examples like Bluesky and Linear, which run businesses with millions of users (over 30 million with Bluesky) with engineering teams of fewer than 30 people, showcasing a new operational efficiency.

(I highly recommend reading the post "Software engineering job openings hit five-year low?" on pragmaticengineer.com for further insight.)

Arguments for the Evolution of UX

In contrast to the pessimistic view, others argue that we're witnessing a necessary transformation:

What is "dead" is a narrow view of UX, limited to the interface (UI) or to delivering visually appealing artifacts without strategic impact.
Towards a holistic view: As proposed by the Nielsen Norman Group, UX should evolve into a more comprehensive understanding of the experience throughout the entire customer journey, integrating more with CX (Customer Experience).
Opportunity for strategic integration: Along these lines, now is the time for UX to evolve and integrate more with business strategy. There's talk of a strategic vision "UX is dead, long live UX" and of a UX/CX merger.

In the current crisis, I believe UX has merely exposed the problems it has carried since its inception. The discipline's ability to reinvent itself will be key to justifying its future relevance.

Personally, I have a deep affection for UX. I leaped from a purely technical role as a web developer into UX back in 2016. It wasn't something planned; it happened naturally as I became increasingly involved in user experience research projects.

At first, I loved things like that direct contact with users. Talking with them and uncovering their actual problems and opinions was a tremendous gift that broadened my perspective on crucial topics such as accessibility (Yes, it's not enough for your AA checklist and a bot to pass a test).

But there were also aspects of UX that I found completely unproductive. The discipline had created a kind of ritual of processes and methodologies which, although sometimes effective, were often the very reason projects dragged on unnecessarily.

The Impact of AI

Now, with generative AI, it's possible to resolve many of the subtasks that used to take so much time: audio transcription, video processing, rapid prototyping, and wireframes. We can now complete tasks that once took weeks in mere minutes. The leap in productivity has been immense, and this inevitably changes the landscape. You no longer need 5 UX Researchers processing information on an extensive research project; a single professional can handle everything (if they master the new tools).

Towards a UX + Business + Strategy

The vision of a UX role that starts from strategy and business isn't new; it was already being discussed before COVID. One of the major challenges has always been aligning clients/stakeholders with this vision. If it was already difficult to persuade stakeholders with data to change their web or app strategy, tackling broader business and product/service strategy will require much more negotiation and collaboration across different areas of the organization. The challenge is not simple, and I see it as even less feasible in companies or institutions not prepared or aligned for interventions of this magnitude.

I believe that after the crisis, there will still be a market for the more "minimalist" or execution-focused UX. However, little by little we will see the emergence of a UX profile we might call "Full Stack", one that tries to incorporate CX and a business perspective into its professional offering.

I just coined the term "UX Full Stack". At Duolingo, they came up with the term "Product Experience" because their corporate culture centers on the product and the term UX seemed outdated to them. My continued use of Duolingo speaks for itself; I wouldn't stick with it if I didn't find it helpful. It wasn't the app that experienced something; it was me. Anyway, the discussion about the name is the least important thing now.

The problem with this "Full Stack" vision isn't so much the technical ability to carry it out (given the leap in productivity with LLM, No-Code, and design platforms), but rather the management aspect. The challenge for the UX consultant or "UX Full Stack" will manage many counterparts and interests within organizations that are often ill-prepared for such collaborative work.

Be Water, My Friend

Generative AI, AI agents, No-Code platforms, and a world in crisis are going to force UX –and other professions in the software industry–to evolve and adapt. UX must once again justify its relevance within companies.

We will probably see consulting or in-house teams that include commercial and marketing specialists working side by side with designers and researchers. While ideal timing may have been years ago, genuine progress frequently emerges from times of crisis.

In the long run, the industry will eventually emerge from this phase of contraction. I am convinced there will continue to be room both for the specialist UX and for a broader, more strategic UX. However, I fear we will hardly ever experience the boom that this profession enjoyed in the last decade again. That "golden age" of UX, it seems, is now history.

My plan is to write about UX at least once a week (🤞🏼). I'll cover various topics, from current industry trends and the job market to key concepts I want to document on this blog as a reference.

Does Having a Personal Blog Make Sense in 2025?

2025-03-17T00:00:00Z

Does Having a Personal Blog Make Sense in 2025?

Something has disappeared from the internet: that disorganized, creative, and strange character of websites from the late 90s and early 2000s. There were no frameworks, much less predefined templates dictating how a website was supposed to look.

In some ways, this is good—millions of users don't have to struggle to understand the navigation conventions of a website. But on the other hand, we've condemned ourselves to a homogeneity of interfaces that, in my personal experience, becomes boring. But what does this have to do with the usefulness of a blog in 2025?

A few months ago, I stumbled upon a website that explained what the "IndieWeb" was. Being a big fan of indie music and the internet, I couldn't resist clicking on that link.

The IndieWeb is a people-focused alternative to the "corporate web." We are a community of independent and personal websites based on principles like: owning your domain and using it as your primary online identity, publishing first on your own site (optionally elsewhere), and owning your content.

I thought it was a great idea, so much so that I started experimenting with 11ty, a static site generator that's quite popular in the IndieWeb community.

The Dominance of Platforms and the Indie Alternative

The IndieWeb won't change the current trend of the web—99.8% of users will remain in "walled gardens" because they're simply more convenient. Your friends, family, and work are there. But I believe the influence of a small group of indies is healthy for the future of the internet.

For example, webmentions is an open web standard popular in the IndieWeb that allows websites (blogs) to communicate with each other. This gives that small touch of "social network" to a group of personal websites without centralization, without using algorithms. A web outside of closed gardens.

My first attempt at blogging was 15 years ago, and the arrival of social media made me completely abandon that idea. Now I see that was a mistake. It would have been valuable for me to have worked on my ideas, projects, and opinions more publicly, even if nobody read them. The process of writing and making sense of your opinions is an escape from "doomscrolling."

I'm going to recall and organize the reasons why I previously dismissed having a blog (my last attempt was in 2015). Then I'll write the reasons why I've now changed my mind.

Why I Previously Dismissed Having a Personal Website

I would have used WordPress, and I didn't like the idea of fixed costs, the time spent maintaining plugins, and keeping everything updated to avoid an undesirable hack.
When it comes to having a blog, I've always believed in being the owner of the content I create rather than relying on a closed platform, which is why I dismissed Medium, Substack, and everything else.
The exercise of writing and drafting wasn't something I felt necessary to work on beyond what's done at work or on social media.
Social media seemed sufficient and an opportunity to find clients and more contacts. Why have a blog?
It was enough for me to share my personal projects with close friends and coworkers.
Making your work public always generates insecurity and is also a leap of "courage" to open up and speak publicly about certain topics.
Let's be honest, before Covid, the world was a very different place.
I wasn't deeply familiar with static websites; my development experience almost always required databases and some CMS.
"AI" (LLMs) didn't exist, and part of the barrier to building a website from scratch with a completely different framework than what I was used to was the time needed to accomplish it. An LLM significantly reduces that time.

Why It's Worth Having a Blog in 2025

Writing on a blog is much more complex than posting a story on Instagram or a ~~Tweet~~, but that exercise has benefits in reducing your "doomscrolling," of only consuming content that an algorithm decides for you, of only producing small fragments of ideas and emotions.
Processing your ideas and opinions, making your personal projects public, even if nobody reads them, organizes your mind and potentially organizes your projects too.
I'm a resident of the United States, and maintaining a bilingual website allows me to constantly practice writing in English without neglecting my native language.
A static website doesn't require (as much) maintenance.
It will be a positive contribution to the web for an indie movement to exist, going against the current and creating new ways of thinking about the Internet experience in a more decentralized way. Mastodon, Bluesky, and federated protocols, I believe, go hand in hand with the IndieWeb.
Your ideas, everything you do on that website, will be on the Internet for a long time. That repository on GitHub has a higher probability of longevity than any closed platform.
I have several professional projects that need their space outside of a CV; they should be "case studies." Having my own portfolio outside of LinkedIn was something I've postponed for too long.
I've been using Obsidian a lot, which is like Apple Notes but on steroids and where you have control over your information. It's based on Markdown, and Markdown works very well with 11ty and LLMs. There's an interesting symbiosis there.
I have many links, notes, and information that I could make public, and this website partly seeks to do that.
Your personal website is a good way to keep a "log" of your projects, work, and side projects.
The sooner you start recording, processing, and exposing your important projects, the better. It wasn't so simple for me to remember aspects of projects that are already 4 or 5 years old. It's always better to record everything as soon as you finish an important project.

Conclusion

To conclude, I'll make a prediction. Not only people close to the tech world will see having a personal website as an opportunity. Soon, LLM agents will likely be autonomous enough to create a complete static website with automatic publishing, with all the customizations the user desires, and it will take only a few minutes. But I don't think this solution will come from the IndieWeb world. Probably a Cloud platform for hosting static websites will see the opportunity to democratize access to these tools and, in the process, exponentially expand its users.

Ultimately, having a personal blog in 2025 isn't about competing with major platforms or expecting to go viral with a post. It's more about escaping the algorithm, organizing ideas that would otherwise remain as fragments on social media, and contributing to the internet that personal touch that has been lost. The benefit is more for oneself, but if someone else benefits from the content you create, that would be a great bonus. Perhaps this is a small niche, but I think it's one worth exploring.

My intention is to publish regularly, not always with long posts, gradually improving this website and experimenting with the various open standards of the decentralized web. If you've read this far, I invite you to get in touch with me through Webmentions.