The AI Middle Ground ⚖️

There is no light without dark. Right now, though, it feels like everything is black and white; black or white. The way that many express themselves feels very polarised.

One thing is evil: bringing certain doom to humanity, destruction and the end of the world. Another thing is good: a promised utopia, human progress—the one thing we have been waiting for throughout the history of mankind. Today, we don't have simple answers to the challenges facing the world we are creating for ourselves. It isn't as easy as "do this", "do that" and we're fine. Equally it isn't as easy as "don't do this", "and god forbid, don't do that". No. It's more complicated than that.

But one thing is certain: this is a battle. It's left against right, yes against no, truth against lie. Or is it? Instead of looking at what divides us, can't we—for once—look at what unites us? But I'm getting ahead of myself. All I really wanted is to talk about AI.

I've been probably thinking about it every single day since the launch of ChatGPT. If you asked me what is the hot shit in generative AI right now, I wouldn't be able to tell you. Supposedly a completely different thing it was two weeks ago. And that's fine. In a way, I think that's exciting. I haven't joined the hype though, and I won't, but I like the energy. I like that there's something in the air; something everyone's talking about, something everyone has their own unique opinion about. And while there are days where I definitely have AI fatigue and I just can't stand spending another thought on it, there are other days—exciting days.

On one of said exciting days, I realised something. I realised how my own stance on AI had been drifting toward one of the extremes over the last few years. When AI entered our lives 4 years ago, suddenly everything was changing. I didn't want to accept this change. Instead, I got really annoyed how everything was going to shit. Because AI—I was convinced—is shit. This is destroying the designed world around us, I was convinced; it's killing our creativity; it's spreading lies. It's being used in weapons; it's being misused in attempts to scam people online. And isn't it destroying the web, too? It's shit, and if I woke up the next day, in a world without AI, I said, I'd be happy.

Again, the day I'm referring to was actually an exciting day, so here comes the exciting part: It was when I remembered that there is still good out there. And it has been right in front of my nose throughout all these years that AI became part of our lives. Let me start with the AI tool I have been using most in the last 4 years. Any guesses? No, it's not ChatGPT and it isn't Claude or Gemini either—but then again, it is all of these three (and Mistral too). It's Duck AI, DuckDuckGo's AI wrapper. It's exactly what you'd expect: an AI tool that lets you choose any of the frontier models and switch between them according to your liking.

While that may not be anything special to many, because I'm sure that there's dozens of tools like that out there, it is special to me. And it isn't just the LLM switching part. What sets this tool apart from many tools I have used is the fact that I gain access to AI without the need of an account and the option to delete all my chat history. You might have expected something more exciting here—but as I've said previously, I'm not one for the hot shit in AI.

But I'm a man of principle—someone who deeply cares about privacy—in my mind—one of the most important aspects of using software in 2026. And the exciting thing about Duck AI is just this: it's an AI tool like any other but with privacy embedded by design. The reason why this realisation is so important is because it reminded me that whenever humans build something—really no matter what, but especially on the internet—there will come someone else around and improve it. DuckDuckGo added a privacy layer onto a—in my mind as of lately—purely evil technology: it complimented it with a human value—a value many people care about.

What happened next is that I took a closer look on my AI use with this new perspective. And: I found more exciting things. One application I have for AI is using it to build on the web. I'm not a good programmer—in fact, I wouldn't call myself a programmer at all. When I built my personal website some years ago, I was still relying on StackOverflow and the generous help of my developer friends when I faced issues. Today, whenever I build a new feature for any of my websites, I default to not using AI, but when I hit a dead end, it's incredibly convenient to have AI as a backup.

Also when it comes to playing around with CSS, AI is there to assist me. I love building on the web and the mission to master CSS has been with me for well over two decades. I write CSS frequently—again, I default to writing it own my own at first, because it's fun and I want to learn it—but every time I can't for the life of me, think of the proper CSS selector, it's super nice to have AI. And then of course, AI's there too whenever I have IT trouble. Usually, AI is really good when it comes to fixing your own, very individual IT problems. I've recently returned to Linux after a decade apart and AI has been generously helping me fix many of my issues as of late. Of course—on the other hand—very often, it can't. Whenever it fails and that tends to happen a lot, it always goes along the lines of:

• You are correct, that solution won't work.
• You are right to correct me. Try this instead...
• Of course this solution doesn't work!

It is frustrating to say the least. Why haven't you told me in the first place, oh mighty AI?

But overall, I'm happy to have it assist me.

Of course, there are further useful applications for AI: practicing new language skills (as I did when I wrote a blog in Spanish), looking up meaning of words or my new favourite use of AI: simulating job interviews. Altogether AI is most usefully applied, when we use it to challenge us, instead of offloading all our tasks and questions and ideas onto it, which basically means shutting our brains off.

There will always be good and evil people and they will make use of the technology at their disposal. As such, scammers find in AI tools a haven for their sketchy business. Phishing emails today look deceptively real. The other day I received a phishing email—supposedly from my email provider—claiming that my password was about to expire. The design looked exactly like my provider's, there were no typos, it looked exactly like the real thing. Only the email address looked sketchy.

Also recently, I've stumbled on a website scam that promised handcrafted wood work by supposed Austrian senior citizens. There was a heart-warming story of how the products were made complemented by AI-generated images that were supposed to establish trust. Luckily by now, it's been shut down. Watchlist Internet, an Austrian platform for consumer protection wrote an article about it (only in German). To me the images looked obviously like AI, but what about senior citizens, who in this case would be potential buyers? I don't think so. When you looked at the imprint of the website, you'd see an address in Shenzhen—which isn't Austria at all—and if you placed your order there you got minor quality products mass products in China. Altogether, I believe that the internet has become a much more dangerous place for the older generation, but that's not to say that younger people shouldn't watch out, too.

On the other hand, there's good people building awesome things with AI. Jan Kus, a German technologist, vibe-coded the original Twitter from 2006 (video in German) and called it Chrrp. My friend Manu built a web app that lets you draw your travel itinerary on a map He built it, of course, with no coding experience at all. And then there's my own vibe-coding projects (or ideas)—more on that in a future post (maybe). AI enables people who were never able to build software to do just that.

This post is neither about praising AI to the skies and nor about saying how shit it is. It's about finding my own individual middle ground and I believe it's important that everyone else finds their own middle ground, too. The discussions we are having are polarised as it is. Oftentimes it's totally legit to complain about the fact that supermarkets are using AI-generated images of frozen chicken in their online store to promote their products. But equally, there are times, when we should all just take a step back and think about the positive sides too. AI shouldn't be hyped the way that it is and it shouldn't be dismissed altogether either.

If we all could just find a less emotional way of dealing with it and take a more objective look at the way to go forward, we can all get excited again. Because there is no light without dark and one thing is crystal clear: AI is here to stay, no matter if you like it or not.