I built my best website ever. AI did 95% of the work.

I've been designing and building websites for 25 years and I've just built the most ambitious, the most crazy, the coolest website I've ever built in my entire career. But, and this is a big butt, AI helped me design and build about 95% of it. And now, I don't know how I feel about that. And even a couple of months ago, I wouldn't have believed AI would have been able to make me build something so creative and so unique. Because only a couple of weeks ago, I thought if I would have asked AI to make me a website, I would have got some generic slop [ __ ] And that's not what I got.

This is the homepage. Everything moves. It's tactile. It's physical. You can throw things around.

Things react and move. There's hidden bits everywhere. Easter eggs. There's an achievements page for God's sake. There's music.

There's sounds. It's all been custom made for this project. And even on the about page, there's an audio guide for the museum that I wrote and recorded myself. I wanted the website to feel more like something that you're playing with rather than a boring website that you're browsing. It basically takes the spirit of an old flash website if you was around at the time and puts it on modern technology.

But I'm deliberately not showing you the whole website here because that's not the point of this video. And if you want to experience it properly, there's a link in the description where you can go play with it. Let's rewind for a second. I've been doing this a long time, as I said, and the early internet was a completely different beast to what it is today. When I first started designing and building websites, things like flash websites were around.

And flash websites were crazy experiential things. We were just making cool [ __ ] because it was an excuse to make cool [ __ ] And the whole industry was thinking like that. We had things like Stumble Upon that was out there where you went on the website, you hit Stumble Upon and it took you to a new and very different website. And pretty much every web agency that ever existed, they had a unique weird website and it was always something to experience. In that day, my delicious bookmarks were full of crazy interesting websites that I couldn't wait to go back to and experience again.

Look at the Space Jam website from 1997 or whatever it was. Websites could be a game. They could be an experience, a toy, something that made a point rather than something that just had a contact form on it, an about page and a blog. The web was inherently more curious and more interesting and more experimental. And I feel like now we've lost a lot of that because everything got better, didn't it?

The iPhone came around, mobile web, the real mobile web came around. After that, things like responsive design came around and then accessibility and then standards and then websites all just started looking the same. We were making websites that were useful that had to work on the front of a fridge as well as work on a mobile phone as well as work on a laptop. But I think somewhere along the way with that, we made the web work for everybody. And that accidentally made the web really [ __ ] boring.

Genius Division didn't need a new website. We had a website already. It was perfectly serviceable. It wasn't very interesting, but we had a website. I didn't want another website.

I wanted the opposite of whatever that was going to be. I wanted a strange, a tactile, an unexpected, and more importantly, a fun experience. That was what I wanted from a website. In fact, I started with this conversation with Claude. I said, "I don't believe you can make genuinely creative work.

I think we're going to end up making another generic corporate website. Prove me wrong. I'm going to work as just as hard as you to get to a website that nobody has ever seen before, that something's truly different, that it isn't like other things out there. And we're not going to stop when we create something competent. We're going to continue pushing.

And when we get near where I think we're going to get to, I still want something more surprising than that. And that's where we're going to keep going until I think we've got something truly creative. The whole idea from this came from a Felix his website that I saw recently. He works at Claude and he used Claude to make this crazy 3D desk, literally a desk as a website. And Felix said all of this was made with Claude.

And I was blown away by it cuz this is a crazy insane 3D monstrosity. Basically, he rendered it all in 3D in AI. And I thought I want to do something like that. I wanted to test can AI make me something that is truly creative. And at first Claude proved me completely right.

It can't make anything creative. This was attempt one. This was Claude thinking that it was being creative. It made me a website that looked like a desktop. And I I guess it's it's fine.

It's okay. It's maybe a little bit twe maybe a little bit obvious. We've seen the idea before. A website looking like somebody's computer desktop. And this version of it here that Claude spat out didn't particularly inspire me.

It wasn't exactly like the Windows 93 website. Hat tip to that cuz that is very very good. That concept pushed it further. And then it came up with this idea, a website sliding across. So we have the normal corporate website on the left and then we have uh a boring old website on the right from 10 20 years ago and when you slid it across it changed websites.

I thought the concept was cool actually kind of really cool and but again it wasn't enough. This was me giving it basic prompts and it returning basic results and I kept saying no this is too generic. You need to try harder. I've seen this before. you need to push further, be more creative.

This isn't creative enough. These were all the kind of things that I was trying to say and this went on for days and I think actually weeks at this initial stage. It can generate endless possibilities of design, but the results it produces are largely dull. So, after about a week or so of going backwards and forwards with these okay, but fairly generic ideas, I went back to Claude and said, "Build me a taste profile. Let's try again.

Let's go from the beginning and do something a bit different. And it was finally starting to get somewhere with this. It was colorful. It was strange. There was a bit of a weird obsession with road signs in it, but it was getting somewhere.

This was the first time I'd seen with Claude where it had gone away and made something unexpected. This time it was giving me something that was a little bit different. And the process now between me and Claude felt genuinely collaborative. It felt like we were actually working towards something different, i.e. a creative process.

But when I got to this version, I think this is where most people would have stopped. I mean, it looked good. It was different. I mean, you can see it here now, but I still didn't think this was enough. And actually, here was the real point where I thought I was going to give up.

I thought this was good enough and it wasn't going to get any better than this. And then something happened that completely changed this project. And yes, I am doing that stupid YouTube thing where we're trying to farm retention. But what happened next was genuinely really surprising. And this was it.

The thing you saw at the beginning, this was the first version that it showed me. And suddenly the thing that appeared wasn't a website anymore. It was tactile. It was different. Objects could be moved around, thrown away, played with.

It had even written some music for me. At this point, this felt like the true breakaway. It felt like all of that effort that I'd spent for days and days and days of battling with AI to try and figure out my taste profile, what I actually wanted. Here's some examples of things that I've done in the past. Here's this sketch that I've done that I want this to look like.

These are road signs that I like. This is the thing that I like. I like these stickers. I like this thing that Designers Republic did over there. I like all these things.

And they all went into a melting pot and they came out and this was here. And this version one, I knew I was on to something. But here, I said, "Make it crazier. Make it more extreme. Show me the turbo extreme version." And when we got there, I knew that was it.

There's three games in the website. There's a hammer at the bottom of the site where you can bash and things fall down on the website. And I've watched the whole team in the office and my other half and lots of other people play with this website over the last week and they've been mesmerized by it because it isn't a website. It feels like Flash but modern. And that's what I love about it.

I got a comment from someone said, "I can't stop playing with this website." And I knew that was it. We've cracked it. We've got where we need to get. But the really crazy part of this, this was just version 0.1. When I got to this, I then went away for another five or six days and tweaked and changed everything about this to show you the final thing that we got today.

I think this website is the best work I've done for 20 years. But the uncomfortable part of it is that I barely designed any of it. And I certainly didn't build any of it. AI built it all. I prompted my way to this solution.

I was truly a creative director in this project, controlling the look and the feel and the taste and the things that weren't in there and things that were going to be in there. I didn't design specifically with my hands any part of it. So that leaves the question, did I actually design this? Can I call this my website? Did I do this or was it just AI that does this?

And that's the bit that makes me uncomfortable. It's the most creative thing I've ever made, but I also didn't make it. AI wrote the code and AI generated a lot of the visual execution and it suggested ideas, but AI made that possible. I wouldn't have been able to do this by myself. I couldn't write the music.

I couldn't do the physics or the animation. I couldn't do the majority of the things on here. I knew when something wasn't good enough. I rejected hundreds, if not thousands of decisions. I set the specific creative direction and brought references and I knew what I wanted in my head before we got there and I wouldn't settle for any less.

I did keep pushing it and I brought those 25 years of taste and experience to that entire process. One of the most valuable skills over the last 25 years that I've learned is that simply knowing when something is not [ __ ] is actually the skill that I'm going to be taking forward. I don't know. And I guess AI doesn't remove any of that. So, it left me wondering, is AI creative?

I certainly went into this thinking that AI wasn't creative. I didn't think it could build something like this, but I've been proven wrong. AI is obviously creative because it's done this along with me. But can a human and AI working together build something truly creative, truly different? Well, the proof is in the pudding, and the pudding is on the screen right now.

not literally a pudding. It's a website, but it's possible. And now that leaves me with more optimism than I've had for a really long time about the web industry and also our job as designers. We have this tool that we can now turn to anything and any idea we have and it can be created. The limit now is our own imagination.

Genuinely, for the first time in our careers, the limit is our imagination. And I think that's pretty damn [ __ ] amazing. And I can't wait to see what comes next. Make sure you go check out the trophy cabinet and the Flappy Bird clone that I made on the