A few thoughts after the Warsaw IT Days conference
On March 19–20, the Warsaw IT Days conference (warszawskiedniinformatyki.pl) took place. The first day was online, the second on-site. I attended in person on the second day and, first and foremost, it was worth the trip. By car. The parking lot was empty.
The conference was held at the National Stadium, a pleasant venue in itself. However, at this time of year it was cold, and I quickly appreciated my decision to keep my jacket on. Judging by other attendees, most of them were freezing.
It was no secret, but worth noting: the conference had a strong focus on job searching and connecting employers with candidates. There were quite a few initiatives aimed at people entering the market. That aspect didn't concern me, but I'm glad such things happen. Looking at the booths, the interest was significant.
Talks
The first track I attended was Jakub Dzikowski's presentation on using vibe coding for content humanization, cannibalization analysis, and site audits in the context of large-scale SEO. Initially I had mixed feelings. Jakub demonstrated tools he'd built for his own specific needs, and they were very targeted to the specifics of his work. I was hoping for more concrete examples, perhaps a list of specific products or implementations. However, after reflecting on it calmly, I need to soften my stance. It was simply a highly specialized presentation aimed at people who truly work in SEO and also have a knack for building their own tools. I lacked the SEO expertise to fully evaluate it, so I shouldn't be overly critical. I think it was a good talk.
Then I attended Aleksander Kaim's presentation on visual product recommendations and their practical implementation at Answear.com. I really enjoyed this one because Olek showed, with concrete examples, an implementation of image recognition technology. A production deployment that actually works is always great to hear about. What's more, he shared that their team built a solution that slightly but definitively outperformed an external vendor's offering. That vendor offered a paid subscription service. Specifics weren't disclosed, but it was clear it wasn't cheap. And here, suddenly, a new AI unit at Answear, a team that probably wasn't particularly experienced since it was just forming, built something better in about three months than a company that offers it as one of their flagship products. Not only did they do it better, they also drastically reduced costs. The development and maintenance costs of the software turned out to be negligible. Congratulations, great case.
Next, I listened to Radek Wojtysiak and his presentation "End of the Hand-crafted UI era? Over 10 years of front-end experience vs. AI Agent." I have to say I was very disappointed. I imagined a review of tools and methods, iterating through different solutions. Replit, Lovable, Cursor, Claude Code, etc. I pictured showing different stages of progressing from a simple solution to a really impressive implementation, perhaps something about Figma, some Figma-to-HTML tool with AI support. Instead, Radek talked about his completely basic first-contact experiences with AI. He told them well, I don't hold it against him, but it was far below my expectations. Especially since the talk was marked as difficulty 2 out of 3. What's more, the author apparently didn't explore the topic further, so he can't look at events from a few months ago with a new perspective. He simply tried some AI in December and decided to talk about it.
Michał Hakenberg's presentation on deploying Vision AI on production lines in the food processing sector was the second highlight that didn't disappoint. Vision AI interests me greatly. I'm working on the Brick Hunt project myself, which you can read a few words about on my website. Again, this is a production case, not hypothetical musings. Someone talking about how it actually worked, how it exceeded their wildest expectations, and how it runs in production. Such experiences are the most interesting. I'm convinced there will be more and more such projects, that they're becoming easier to execute, and I genuinely believe these technologies will become widespread.
Then I attended Aleksander Misztal's presentation "Agentic AI: Why most implementations don't make sense (but ours does)." The topic seemed interesting, but to be completely honest, this talk didn't captivate me. The speaker seemed quite stressed; the packed room probably overwhelmed him. I don't want to be overly critical. I generally have high standards, so I usually leave conferences dissatisfied. Several important points were made during the talk that I agree with, but the speaker's stress and very poor slide visibility with small text really hurt the reception.
Panel discussion on the future of the industry
Finally, I had the pleasure of listening to a panel discussion on the future of post-AI development. On stage: Tomasz Ducin, Sebastian Gębski, Tomek Onyszko, Przemysław Smyrdek, Marcin Czarkowski, and Artur Skowroński. I must admit, I really enjoyed listening to them. They were largely in agreement on most topics, including the fact that the future doesn't paint itself in dark colors at all.
Personally, over the past two or three years, since AI conquered our minds and our work, I've had various fluctuations about where this is heading. From the very beginning, as soon as ChatGPT appeared, I recognized the coming revolution and tried to convince everyone to pay attention, because it was happening before our eyes. But there were moments when I thought this artificial intelligence would simply take over.
Now I try to approach it rationally. I see that this technology, however smart it may be, will have difficulty functioning without humans. Of course it's a revolution, and like every revolution it causes serious reshuffling in the market. We're only at the very beginning of this journey and we don't know where it will take us. I'd rather not comment on superintelligence or look that far into the future. The intelligence that exists now is already very useful, but it won't manage without humans.
What's very important, however, is to understand the role of that human, because it's changing drastically. The mocking label "code monkey" wasn't coined yesterday. We've long had greater expectations of developers than just writing code. There are different kinds of programmers, some more committed to high-quality code, others approaching it more from a business perspective. Both will probably find their place, but everything indicates that for most, it's rational to focus on business and solution architecture. Writing code itself is no longer any challenge. But someone still needs to see the business value in it, some utility, still needs to come up with something others need.
I'm convinced that this time we have now and the coming years will be the best time for developers. I know there's a lot of talk right now about AI taking jobs from programmers and about juniors having a tough time entering the market. But I think many of these opinions are exaggerated.
Has it ever been so easy to learn something as it is today? And this doesn't apply only to programming. I haven't been a junior for a long time and it's hard to speak on their behalf. The panelists, my peers, admitted that too. But here's how I see it: juniors can compensate for their gaps with engagement and effectiveness in using AI tools and be very valuable. They can be more valuable than people with much more experience who are resistant to this change. And there are still many such people. A junior combined with artificial intelligence can play an enormous role.
In the discussion about AI taking jobs, I always reference the WordPress example. Before WordPress, creating websites was more expensive, slower, and required more commitment. Then WordPress appeared with its famous 5-minute installation, and suddenly everyone could launch their own website without even knowing what FTP or a database is. Just go to a hosting site, click through a bill for about 50 EUR per year, use the installer, find free plugins, buy a template for 20–30 dollars, and build your site. Back then, people also said that now everyone would make their own websites and nobody would need developers. It turned out to be completely the opposite. Hundreds and thousands of agencies specializing in building WordPress sites emerged, employing thousands of developers. This ecosystem has existed for many years and made us all a lot of money.
It's not the case that when we have greater capabilities, everyone suddenly does everything themselves. We developers are in a bubble. If we suddenly started doing things 10 times faster and 10 times more, we think the whole world can do it. That's not true. Even though this technology is super useful and will become even more intelligent, I don't think people will en masse decide to build software on their own. They don't really want to, they'll never enjoy it, and as we know, we prefer to delegate responsibilities to others and pay good money for it every day.
Around WordPress, on one hand new space for developers appeared, on the other hand marketing agencies emerged, SEO grew, many businesses built themselves around this phenomenon. In the same way, we'll now go through a revolution around which many different ventures will form. Sure, programming will be cheaper, but that won't result in more programmers than work. It will automatically unlock funds and resources for people who had ideas for startups, implementations, software, but couldn't afford it. The phenomenon of building something for your company, business, or family will simply become more widespread.
The takeaway from the panel? The requirements that artificial intelligence places before us (think more in terms of product, less in terms of code, be active and engaged) are essentially traits that were always needed. Nothing has changed. Such people were always sought after and will continue to be. You should try, build, look for your opportunity, be active. You must not hold a grudge against this technology. The revolution has happened and it will only grow. Don't be afraid of it, find your place in it.
Final thoughts
All the talks I attended were obviously just a slice of the program. I'm very much hoping to watch twice or three times as many online, since recordings should be available to conference participants on the official website. There's also a VOD section with a number of talks that were prepared in advance, some of which can already be watched. I'll definitely be returning to this material in the coming days and weeks.