On a Paradigm Shift

"Code is cheap now..." - I've been hearing this phrase a lot lately. Because of which I think more and more about the role of developers and the experiences that I have gone through over these 13 years.

At the beginning of my career, I loved typesetting. I liked making interfaces, although at that time they weren't interfaces, they were more like some kind of websites, endless landing pages, promotional projects, and so on. Not everything worked out right away; we had to struggle a lot with the support of different browsers, and with the widespread development of smartphones, also with adaptive layout.

The workflow was like this: first you make up the layout looking at it in Adobe Photoshop, then you upload the PNG, open the PixelPerfect extension for Chrome, compare sizes, try to adjust the fonts, worry about the problems that some Internet Explorer users have or something else.

Now, of course, this is not the case. I remember even looking for layouts on Dribbble and laying them out for practice, and then adding this layout as an example of projects to my resume. Then it seemed important and served as a kind of indicator of the developer's skill.

And I thought about the fact that the development for which I came into the industry is changing, and rapidly.

The other day, in a chat, a novice layout designer shared a landing page that he had designed. Most likely it took him 1-2 days, maybe a little less. Out of curiosity, I decided to replicate this landing page using Cursor and the almost ready-to-release page was ready in about 2 minutes. Fill it with content and you can upload it to your hosting.

Price? $20 per month, but I probably spent less than $1 in tokens on this session.

You connect Figma MCP, choose Codex or Opus, a couple of minutes and your design is ready in code. There were days, now minutes.

And it seems to me that now is a very interesting time to improve yourself, to find something new in your own field.

I am very glad that I became a manager, literally and figuratively, because telling agents what to do is still much more pleasant than doing something yourself.

Fundamental knowledge is important, but even more important now is to be flexible to everything new that appears in our industry.

To be in denial now is to admit that your peak as a developer is already behind you.

Yes, you get tired of the noise around AI.

But when there are models who write code in much the same way as I did a couple of years ago, do it many times faster and are almost not distracted, it is simply irrational to ignore their use.

The full version of the phrase sounds like this:

"Code is cheap now. Software isn't."

We are still needed - perhaps for now. But our role is changing. And the sooner we accept this, the sooner we can benefit from new models and tools.