2025 has been a year full of AI hype BUT also of real improvements of AI models and tools. And the job market, especially for junior devs, has been ... rough, to put it mildly. How will things change in 2026?
Audio and video play straight from the show's own hosting. PodSubs stores and serves no media.
Transcript
From the show. Tap a time to play from there.
2026 will for sure be another interesting year and I think in some regards, like the job market, it will be better than 2025 for developers. So I'll share my perspective on AI, obviously, the job market and why I think it will get better, and web dev and web dev trends and technologies in general. So let's start with AI because that obviously will still be a very important thing next year, and here
I think we have kind of reached a plateau when it comes to the intelligence, if you want to call it like this, does not mean that we won't see more development or better models or tools around these models next year, and let me explain. So if we had a chart and we would start in 2022, obviously where ChatGPT was released, and then we would end at the end of 2025, I would say the "intelligence," and I put that in quotes because they're not really intelligent, they're token generators, but of course they do that pretty
well and they can use tools and I'll get back to that. But the intelligence, the capabilities of these tools, I would say developed something like this maybe, so a pretty good development. Obviously we all saw that, but I think that now we're kind of at this point, so I don't think we'll see an exponential improvement or anything like that. Instead, my feeling for the last year whenever new models were released was that there was a lot of hype, there were benchmark numbers where the model was better than most other
models of course, but at the same time not that much better. And especially once people started using than once that it was kind of meh in the end. I mean, take for example the LLaMA 4 last year. When LLaMA 4 was announced, the model family, we had impressive benchmark numbers. We saw that the top LLaMA 4 model, maverick, basically was better than all the other top models
or other frontier models we had at that time in benchmarks. But shortly after the release, this where people were kind of, "It doesn't feel that great," and we had that more than once last year. Now don't get me wrong, models did get better for sure, especially of course compared to 2022 or 2023 or 2024 and so on, but we're at this point where they won't get drastically better every year, and I'm talking about
large language models here. If we take a look at video generation Veo or Sora 2, obviously we saw a drastic improvement there compared to the years before, and the same is true for image generation models. We had the release of Nano Banana, not the pro version but the normal version, in August of this year I think, and that was really groundbreaking. And then of course we also had Nano Banana Pro which raised the bar even more, and we had other models like the
Flux models by Black Forest Labs, Flux 2 recently released, and all these models definitely got much year before. But if we take a look at large language think we're reaching that plateau. Now I will say there's a lot of development on those models to make them better at specific tasks. I think that's the clear trend we can already see that will continue in 2026 because of course we now have those thinking models. That is pretty much the standard now for
models. We also have models that are better at following instructions and at tool calling, and that is a big one. This here really is a big one because that is important. It's that improved tool calling capability of new models that makes them way more useful in the applications that use these models, and that is important. With this chart here, I'm referring to how models or their raw
intelligence kind of improves. Now if I were to copy this chart and judge the capabilities of applications that are built on top of these models, it would probably look more something like this, more linear, not exponential still but linear. Because when we look at applications that use AI like ChatGPT or Gemini themselves obviously, then they got better. Oh, also of course because all these models have great
multimodal capabilities. They're good at, uh, taking a look at images, understanding images, generating images by integrating other models, and that's basically what these applications are doing then which takes me back to this chart here. It's important to understand that of course we are exclusively almost interacting with applications that use AI, unless you use the raw models via the API to build your own applications of course. But as end users, we interact with applications like
ChatGPT or Gemini of course, so that would be example applications we have, but also something like Cursor or Claude Code. These are applications built on top of AI models and those of course got much better over the year and I think they will continue to get better. Why? Because the developers of these applications get better at integrating those AI models, there's a lot of wiring of different
models behind the scenes, there is of course prompt and context engineering, and there are tools that are exposed to those models. If we take a look at ChatGPT here, the ChatGPT of today is of course nothing related to the ChatGPT we had in 2022. It has, for example, web search. It can, uh...... write and execute code. It can generate images and that, of course, is technically done through a different model, not through the GPT large language
model, but through an image generation model, ChatGPT application that wires these different models together. And other applications like Claude Code are also potentially wiring different models together, doing prompt engineering, and they expose tools to these models like web search, which is exposed by ChatGPT and also by Cursor. It can also search the web. And that is one part where those models, of course, did get better because they are optimized and
fine-tuned for being good at tool calling. Now, just to make sure we're all on the same a tool really just means that the developers of an application describe the available tools to the model and then the model emits tokens that describe an intended tool use if it feels like calling that tool. Now, of course, whether it feels like calling a tool simply on the prompt engineering, the system prompt and how the model
has been trained and fine-tuned. So for example, if a tool has a get weather tool available and the user asks it what will be the weather tomorrow, there is a very high likelihood for the model to suggest that it wants to call the get weather tool. Now as a next step, it will emit those tokens that describe that tool call and then it's again the application like ChatGPT that will make the actual tool call. So that it will execute the actual code
behind the tool and that will then pass the results of back into the chat history, so to say, to have the large language model then generate the final output that should be shown to the user. So all that tool calling and these intermediate results are not necessarily shown to the user, but the final result, which call, is. And that, of course, is where those models did get much better, or the applications I should say, they will continue to get better because we, as developers, can leverage
those large language models that have been optimized for being good at tool calling and so on to build more capable applications on top of them. So that is the trend I see continuing I see Cursor or Claude Code and all those other applications and all those applications that are not related to coding at all become even better in the next year, not exponentially, but linearly I would say. Now what does that mean for us developers? Well, it means two main things I would say.
For one, it gives us opportunities. There are opportunities for building interesting stuff on top of AI, on top of large language models, models, and potentially combined capabilities. Like the Build My Graphic app I built, I built it for me because I need infographics for videos like this and it's not just about infographics. You can generate all kinds of graphics there. And I built this on top of Nano Banana Pro, but also by using many other models to process the user input,
do research, whatever, and then generate graphics based on style templates provided by me. So I built an application on top of think there will be many opportunities for small and bigger useful applications that you can build on top of AI because the AI models have all these capabilities now which they didn't have two or three years ago. So I think the first important implication takeaway is that this gives us opportunities. The second important implication for us developers, I
think, is that we must be able to work with AI assistants for coding. And I know that's not everybody's cup of tea, coffee, whatever. You, of course, don't have to if you have the luxury of not needing to, but if you are in a job and want to keep that job or want to find a job, you really need to build the skills to effectively work with AI assistants for coding.
And you might wonder, what does that mean? I mean, I can write a prompt. It's about writing the right prompts, asking for the right things, providing the right context, not overwhelming the model and asking for too much, small, incremental demands or requests. And I will create separate resources on that to share more about how I work with AI and how I think you can effectively work with AI because I definitely want to help there. But I think you need to build that
skill unless you're really in a position where you're your own boss and you can do whatever you want, and then of course it's totally fine AI at all. Let me be very clear here. I'm also not saying, and I want to be clear about that too, that thing that will generate all perfect code and that we should go into live coding. Quite the opposite. I'm talking about AI assistants here, not live coding. You as a developer can and should steer and control the model. You should tell it what
to do and you should get good at being precise there about providing the right context, about being able to evaluate the results you get back, about building a plan before you even prompting. And again, I'll share more about that. But that is the, the second important imple- implication, I think. That is my take on AI in this, uh, new year, uh, from a developer's perspective. Now of course, AI is one thing, arguably the most important one, but let's talk about the
job market because that clearly is also important to many people. And here, we probably all know this chart here, this horrible chart where we had this peak here during the pandemic and, and now we're down here. Now the good thing, I guess, about this chart is that it looks like we, bottom, right? Maybe there even is a slight improvement coming for us or already happening, and I definitely think there will be an improvement n- next year. Now we also have to differentiate of course. We have to differentiate between jobs for
senior developers, and I'm only talking about the developer here, and junior developers. I think for senior developers it was already okay or good and it definitely will be good in the next year because all those companies are scrambling to get senior developers, especially the ones that are capable of working with AI, of steering the AI, and of judging the output by AI. So senior developers
h- have a relatively good position I'd say, get much better in the coming year. The problem, of course, is that not everybody's a senior become a senior by being a junior first and here, we definitely had a- a bad (laughs) state of that job market in 2025. But the good thing is, I expect that to get better in-- not badder, (laughs) but better in the next year. In 2026, I expect that to
improve. Now, why do I expect that to improve? Of course, we have all these reports about layoffs. We have all these reports about layoffs because of AI that happened in, uh, 2025 but we have to be careful here. It's pretty easy to mention AI as a reason for a layoff. That's a nice reason to have. It's especially a reason that might please investors because, they wanna hear that you're an AI first company and that you're
embracing AI. And I'm not saying that AI is not having any impact on the job market, it clearly has, and even if it's just that many executives expect AI to be able to do your job in the future, so they rather don't hire anybody new for now. I'm pretty sure though, that more and more companies will realize that going all in on AI or relying heavily on AI instead of humans that made and work with AI, is not a good strategy. I'm sure we'll see in 2026 that more companies
want to have a workforce that's capable of using AI as a tool because it is a great tool, at least in certain industries. Encoding, I would say, it can be a very useful tool but it's that, a tool, not a replacement, and I think more companies will realize that in 2026 and they will also realize that yes, of course, it would be nice to have more seniors that are capable with AI, that's why this will be such a good job
think, but that there aren't enough seniors squeeze and that's why I think more companies will be open to hiring juniors again. But the role of developers will change, not just next year, but in general. As I mentioned, it will be more about working with AI, about controlling AI, and junior developers that feel confident working with AI will be more interested next year than junior developers or developers
in general that don't. And therefore, that is one important takeaway for next year, the first important takeaway. As a developer that wants to get a job, you must be able to use AI. And again, sounds like a simple thing but you must feel confident using it. You must be able to show potential employers that you can get a real value out of AI assistance. That's one thing, but I also think there is a second thing, unfortunately, that will be important
as a junior developer, especially if you wanna get a job next year, and that second thing is that you will need connections and/or a strong social media presence or anything like that in addition to having a good portfolio, a good resume, and so on because there is this huge crowd of junior developers. I mean, look at this chart, right? We're coming from here. Lots of people worked in the development industry, lots of people started getting into it because there was all this demand here and therefore, all these people are
still here today but there are less jobs. So in order to stand out from that crowd here, you will need something that makes you stand out and for sure, that is proving that you can work with AI, employers will want that, but it will also be about, well, having connections, having a strong social media presence, anything that makes you stand out. And I know that these are not great news because you wanna work not as an influencer. I totally get that.
I just think that will be something that will be needed to vastly improve your chances of getting that job. And again, you still need that portfolio and all that fun stuff, but that will be important too. So that is the job market next year. The last thing I wanna take a look at dev in general, so the fun part now. Which trends do I see there? Which technologies, uh, should you learn and use? And here, I would say we have one
clear default stack, the standard stack you will use in most projects next year and that is TypeScript, React, Next.js, Tailwind. That is AI's favorite right here. That is what you get in most, uh, applications if you just ask AI to generate something for you, some web app. Uh, that will be important next year too. Now, by the way, I of course have the JavaScript angle
PHP world. Let me be very clear, PHP, Laravel, all amazing technologies which can land you jobs. I just have that JavaScript lens, which is why I'm talking about that. But if we take a look at the ReactJS download numbers, for example, it's pretty clear to see that AI has a favorite here in web development in general and especially the JavaScript web development world, of course. So that is that default stack we'll have next year.
That is the stack which you should learn are that most companies or many companies will be using that. Now-If you have the luxury of deviating from that because you're not looking for a job your own, or because you wanna kind of expand your wanna have more than just one stack that you bring to the table, I think for next year, one interesting thing to dive into would an alternative to Next.js like React Router or TanStack
Start, which is amazing and TanStack Start, especially something I wanna cover more next year. I think it also might be interesting to look at modern browser APIs like the, uh, Popover API, for example, which I recently covered on my other, uh, channel and I'll do more about and about CSS because a lot has improved here. Uh, raw CSS and also what the browser can do for you has gotten
so much better in the last two years. It's just that nobody noticed because of, well, AI. It's the big thing which takes all the attention but there has been such amazing development here that you can get a lot done by just using what's built into the browser which is available across all major browsers. So it's production ready and/or by using modern CSS features like native, uh, CSS nesting where you can write CSS code
like this and it just works, no compilation step needed, by using CSS layers to control precedence and much, much more. There's a lot of amazing stuff available in the browser right now. It's also something I wanna teach more in the next year to, well, bring some more awareness to that because you might not need all those extra libraries you're pulling into your project or all that code that's being written to be wr- you don't need to manage your own modals
code and your own divs which are then maybe not accessible or not always on top. You can let the browser do a lot of that work and the result will be better. And then, of course, I mentioned it before, you can build stuff with and on top of AI, on top of large language models, image generation models and so on, There are all these opportunities, uh, which you as a developer can try to take advantage of because you know how to write code.
You can use AI to accelerate the development process and, therefore, you have these opportunities stuff on top of AI by using AI, that can really also be an interesting area next year, and I got exciting stuff to share next exciting content I'm working on and, um, yeah, I'm, therefore, really looking forward to the next year. I think it will be a better year there for developers than 2025 was to some extent. Definitely still challenging, definitely also exciting, I think.
GPT 6 Astra was released. It's a good model, I have to say. Especially for 3D work it seems to be a real game changer. But due to all the hype, you may be disappointed...