Written on 2026-04-11 by Adam Drake - 8 min read

My Medium friends can read this story over on Medium.
I’m so confused right now.
Last November, Anthropic released Opus 4.5. It seems in retrospect this was a bit of a game changer. I’ve been using it since day 1 and I have to admit, it’s a beast of a coding agent.
I am developing an internal tool right now at work and I’m mostly using Opus 4.5 to write the code. It’s in Python which I am not super comfortable with (yet! We’ll get to that). I am using it carefully to plan, build out features and write tests and documentation. I am still definitely checking things along the way. But this is the first time I’ve used AI where I don’t understand every single line of code.
Ryan Dahl just dropped this message on Twitter — “the era of humans writing code is over”. He says we still have work to do but writing syntax is not it. Ryan is one of the people I respect most in the industry and to hear the likes of him saying it really hit me.
Jeffrey Way, another developer who I have the upmost respect for dropped this video recently saying his business is basically over. AI has more or less destroyed the demand for the Laracasts service. However, he also states he’s having so much fun coding at the moment with AI. He still thinks the editor is core to getting stuff done, but ultimately the way software developers do their jobs has and still is fundamentally changing.
This begs the question — How do we build software now?
I see this opinion thrown around quite a lot online. You have to be an architect now. Everyone’s going to be architects and we’re all going to architect our way to software glory.
On the one hand I can accept this. You can see how this makes sense, but on the other hand it’s not something I can accept fully. There are too many holes in this approach.
An architect’s main skill is designing buildings, and they can be really good at that. They make decisions about what materials the building should be made out of, they decide what should go where and the similarities with that and building software are quite clear.
With this approach AI will be the builder and they will build whatever based off of our plans and specs. They’re essentially going to lay the bricks, build the roof, implement the electrics, do the plastering and then the decorating. Everything’s going to be left to AI.
For me this sounds a little naive.
Brick laying, electronics and carpentry, like coding, all require skills. We’ve all seen the meme going around at the moment with the brick layers and vibe coding.
Press enter or click to view image in full sizeIt’s funny because there is an element of truth here.
This is for me where things start getting a bit murky and grey.
Do we accept this potential issue where we’re using AI to build all the code, even though it might only code 80% to 90% of the application in the way you would if you did it by hand? Is this worth it because of pure productivity and time saved?
Most developers have seen the issues first hand using AI to code. Lack of context, memory issues, sloppy code, removing tests that fail… The industry currently is taking the approach of putting in a whole bunch of guard rails — skills.md files, small context windows, spec driven development.
There’s always a trade-off, so as an industry, are we saying we’re going to accept this?
For the foreseeable future I still think you need to know what is going on in the code. This is what my gut is telling me anyway. Sure, put the architect hat on and use Agents to produce some of your code. But when working on something important that is connected to your livelihood then ultimately you are being paid for the “Responsibility” of knowing whats going on in that codebase.
Don’t hand that all over to AI.
Then we start looking a little ahead.
What about the programming languages themselves? Are we going to continue to rely on them and continue producing code in those languages? Are we going to get AI to produce it’s own programming language? Are we going to lose the need for programming languages altogether and let AI just produce assembly code straight out? What about the future? Should we spend time anymore learning the ins and outs of syntax?
There are just SO MANY questions…
I don’t have answers for this.
There are just too many possibilities open at the moment that it’s impossible to make any kind of sensible prediction.
This is what is worrying me at the moment. The sheer amount of directions this whole thing can go in. I see some people jumping ahead first into this AI way of doing things and my gut is telling me to not jump yet.
There is this real tendency in our society at the moment to be “first”. We believe that that’s where the biggest rewards lie. With the spread of information so fast on the internet, it seems everyone’s keeping one eye open to see what the next thing is and jumping on that straight away just to be the first.
What do we do right now then? A good question I have found for that is “What can I control right now?” — this helps me guard myself against the anxiety all these changes can bring.
It’s also maybe time to take a step back and ask yourself where do you find the joy in programming? What specifically. I actually really enjoy digging. I find something out that I wanted to know and from this springs two more questions. Just take this really cool trick from “Bread on penguins” that I learnt the other day. It’s just a nice trick you can run in the terminal that has it’s used cases but I love finding out about this stuff.
I find myself asking questions to AI all the time. This is curious because even though I know AI can do it, I still want to understand how and why the thing works the way it does. This curiosity drives me. Back to my point about not knowing Python well “yet”. AI is helping me understand it more and more.
A quote I love and live by which is associated to Ryan Dahl is “You can never understand everything. But, you should push yourself to understand the system.” I think this is truer than ever right now.
I was relatively late to the game in Software. I lacked a purpose and meaning in my life before I started programming. When I finally found Software, I fell in love with it. For whatever reason it just really clicks with my brain, my personality and my way of life.
I see AI as a threat to all of that and I think that is what scares me the most. I see companies like Anthropic and OpenAI coming along with their fancy LLMs and threatening to take this meaning and purpose away from me.
This has happened in history to many people before and it will happen again. I like to think of myself as a realist. I ultimately will face the true reality and adapt as best I can but my problem at the moment is knowing what is real and what is not.
AI companies are obviously overselling the abilities of their tools but at the same time this isn’t another crypto bubble. AI is actually changing and having a real impact on almost everybody’s life.
What does the future hold for Software. Only God knows at this point.
Enjoyed This Post?
If you found this blog post helpful, why not stay updated with my latest content? Subscribe to receive email notifications every time I publish.
If you're feeling really generous you can buy me a coffee. (Btw, I really like coffee…)
I live in the vibrant city of Prague, Czech Republic, with my family. My blog is more than just articles; it's a community of like-minded developers who share a love for innovation and learning.
I'm a passionate Frontend Developer specialising in React and TypeScript. My professional journey revolves around exploring and mastering new tools and libraries within the JavaScript ecosystem.

Adam Drake is a Frontend React Developer who is very passionate about the quality of the web. He lives with his wife and three children in Prague in the Czech Republic.
Adam Drakes Site © 2026