Learning is painful, and there is no going around it. Yes, it can be really gratifying, both immediatelly, like when it gives you a small reward, or maybe years down the line, when suddenly things click, and you can understand a new concept. Learning requires you being at the edge of your comfort zone, constantly improving upon your last version, correcting errors and trying new things. Therefore, learning is usually done with a goal in mind, which gives this pain a meaning. I love when I manage to find the beauty in learning…

Expiring Skills

Yet it you talked to me lately you know I am freaking out cause I do not understand what is worth learning anymore. Cause? Frontier Large Language Models. Since February, when I started working at Tufa Labs, and I started properly taking advantage of the AI tools, I realized how powerful they are and how much they are changing the way we work and think. These models are directly produced to target programming, and the ugly truth is, the programmers managed to be the first people to automate themselves away. As of now, with tools like Claude Code or Codex, you can literally describe in plain English what you would like to be built in the virtual world of code, and it will build itself. You can just sit, provide API keys, sign up for accounts online, and the projects just grows. Everyone now controls around 2-5 agents at the same time, as they can execute their tasks autonomously. Every programmer has upgraded into a small team manager. More code, even though of lower quality – some might argue – is getting shipped every day.

You might be thinking: isn’t that great, why freak out? The coding tools are not just good. They are amazing. If you use the ones that are state of the art, you get exactly what you want, provided you have given good guidance. Many programmers have not written single line of code since January 2026. This is slightly terryfying. This feels like skills, that I have been carefully building up over the last 4 years of my studies are suddenly worthless. I was always anxious about my career, but seeing something fly out of the window at such speed is stunning me. It feels like I could have just as much been studying biology, or history, and the value of my knowledge would have been the same. Let me give an analogy.

Imagine you have just finished education as a baker. You have learned how to mix dough out of ingredients, how to let it sit, knead up bread, bake it and how to maybe even package it and sell it. Suddenly, there is a new machine: you just put in ingredients, and the recipe in. And voila! You get the baked goods out, exactly as per description.

What is the baker thinking? Yes, their job has just gotten so much easier. Yes, they understand the machine better than people than never baked a thing. Yes, they can bake whatever they want. Yes, they are baking faster. Yet anyone who wants to bake anything now actually can, and they will get it done at ~75% quality of the baker. This means that their skills themselves decreased in value. Everybody wants bread, but nobody needs the baker.

And this is how I feel about being able to code. It is amazing, but nobody needs it. It breaks one of the pillars of Ikigai: Do what the world needs. The worlds needs code, not coders.

My Future

It feels like AI has given me a future, and now it has taken it away. I have joined bachelor’s of AI in the Netherlands, and I always imagined a career around that. I have always identified with the skills and understanding that I was gaoning, and I was really proud of moving forward along this career trajectory. I always loved automation, I like to describe it with the following story.

Three years ago, a friend who studies Economy has described his job to me. He had to download invoices and put them into folders, and read some information from them, and put that into a spreadsheet. Simple work that would take him a couple hours each week. I coded up a small script that has automated this, so it was done in a couple of minutes, just needed to be verified that everything was parsed correctly.

I had high affinity towards coding, so it was fun. It blew my friend’s mind. I really liked this cause it felt like the skills that I have accumulated have brought value to the world. The skils that I have learned were so natural to me, but others could not replicate them so easily, I thought I found my thing. Now the engineers of Anthropic have played the UNO Reverse card on me and the have automated me. More so, I have identified a lot with jobs in AI/ML engineering. It was something that I wanted to do, and I tried to learn skills to be able to do it. It was the reason I went to the Netherlands and Switzerland. Suddenly, this sense of path, and identification is gone. I am back at square one.

Counterarguments

I usually yap to other people about my ideas before I write my blog. When I share my worries, most people think there is no need to worry. Here is what people tell me, and my responses to it.

“People still need to take responsibility.”

Yes, that is true, but they are taking increasingly less and less responsibility. Recently a friend writing a paper told me when we were talking about how he padded his data: “I do not know, I used Claude Code so much I do not know anymore what is in the final version.” That is the world we are moving towards. If we do not need to know, we will not know. Until fire breaks out, nobody will be taking responsibility.

“You are doing Master’s, you can learn anything.”

Yes true, but that does not mean I will like it. Coding was something that I liked. As said before, this puts me on the same line with somebody who has barely ever coded. And that is okay, but suddenly the easy path is not there anymore.

“You have learned so much more than just coding, you learned to cooperate and organize yourself.”

I would disagree. The pace of work with coding assistants is quite something different to pace of work with human written code. Yes, from personal point of view I might have the same perseverance, ability to motivate myself. Yet the ecosystem is changing so fast, that you might just as well know nothing about it and do great in it.

I like to compare it to the internet descreasing the value of memorizing and copying information. This is a new revolution, where knowing how to use digital services is a solved thing. What should one then focus on? I think the ability to learn is still important, but not in the sense of learning to perform a skill efficiently, but rather understanding a domain quickly and learning to automate it. Setting up agentic workflows is the new coding. In hindsight, it actually makes a lot of sense that coding is the first job to be automated. Programming is literally about writing re-usable code that works for you. It is natural that these people are the first to automate their own work. However, it now seems that initiative now matters more than long-term learning. Being first, and being fast is more than accumulating 40 years of expertise in the world where anything tha can be learned is consumed by machine.s Being reliable, being able to communicate and lead retains its value. These are things that should be practiced more.

End of Coding

So if coding is dead, where shall we go? Coding always attracted people who liked solving problems. Same as math, it is perceived as something challenging, that not everybody can do. This has made it attractive venue for ambitious people. Coding was, however, also doing something different, being kinda futuristic. Automating everything. I wonder where other people with this hobby now go, if all these problems are automated? When I say my cousing back home last week, He told me about a JS meetup he is organizing. The event talks about that it is dead, and this is a final farewell. I am curious what new hobbies these people will find now. Here I would like to pay respect to all the people who contributed to creating current AI. People who loved problems and got us where we are. The early adopters, the ones that are now “needlessly clever”. Their mission seems to be finding its goal.

From people that I talk to it seems that the more senior they are the happier they are about AI. It boost them, it reads the emails they cannot keep up with. Generally, I think if you are in a good position, AI helps you to get even further. If you have not gotten there yet, you are worried even more. A lot of people around me: well educated people in STEM, are losing their hope for making a meaningful contribution to the field. I think this is a real problem, and giving young people a perspective is an important thing.

One perspective that I got from my older coworker is that in this era, one should be falling with the outcome, not the way of getting there. Good research is good research, regardless of who wrote it, AI or human. This is quite a good remedy. Let’s see how long we can keep up with the machines in that regards. Generally, I think it would be sad if coding went extinct. I think coding is remarkeable activity, and it the same like when I am writing this blog. I write a bit of text, come back to it, edit it. This blog is the manifestation of the flow of my thoughts. The Code has been the embodiment of practical thoughts. Now those are dead. I started looking into registering coding as UNESCO Intangible Cultural heritage.

Silver Lining

I think as with any revolution, the benefits outweight the risks, I think now is the best time to…:

• …to understand new things. Need tailored classes in quantum physics? AI can give you those, and read any textbook with you.

• …to do reaserch. Pick up a paper, let AI explain it to you, design an experiment and execute it. Especially in computer science. If you have something interesting, publish it using AI.

• … build a business. For 20$/month you have access to PhD level intelligence in all fields, including practical tips on accounting. Once again, access to knowledge has never been more democratic.

Curious about your thoughts, and what you think where this new revolution will lead us…