AI Changes Nothing – Why Product Success Still Requires Taste and Creativity
Crash course to product success: three critical product moments AI can't solve – creating shareable marketing, reaching aha moments fast, and retaining users forever
I feel like the AI hype is finally reaching its endpoint. At least the sentiment is starting to shift. Every housewife now knows about ChatGPT. Every LinkedIn bro has posted their hot take. Which means it’s time for us to actually think critically about where AI helps and where it doesn’t.
Recently, I stumbled across an interview with Dax Raad, the guy building OpenCode – one of the most popular open-source coding agents out there. He’s not some random influencer hyping the next big thing. He’s in the trenches, shipping actual software that real developers use every day. And he said something that many of us have been feeling but maybe haven’t fully articulated:
AI isn’t going to suddenly make you amazing!
It caught my attention because, all of a sudden, people finally started talking about what’s obvious and what nobody wants to admit: the fundamental challenges of building successful products haven’t changed at all. AI didn’t solve them. AI won’t solve them. And if you’re betting your business on AI doing the hard work for you, you’re going to lose (there’s a small chance you might get lucky, though).
So today we talk about what actually matters: the three critical moments in your product funnel that determine success or failure – and why AI is basically useless for all of them.
Marketing Is About “Wow”
Let’s start with the top of the funnel. Marketing.
True marketing isn’t blog posts announcing features. It’s not paying influencers. It’s not clever taglines or good SEO.
Real marketing is making people stop and think: “Holy shit, I need to show this to someone.”
That’s the bar. If nobody wants to share what you’re doing, you haven’t done marketing. You’ve done... corporate communications? PR theater? Whatever you want to call it, but not marketing.
Think about your own behavior. When did you last send a colleague a link to some company’s feature announcement? Never. Nobody cares. You’re scrolling through content, your attention span is destroyed (thanks, TikTok), and you’re definitely not getting excited about “announcing our Q3 product release with 14 new capabilities.”
But you know what you do share? Something weird. Something that makes you go “can you believe they did this?”
The challenge isn’t understanding this – it’s obvious. The challenge is creating that shareable moment. That requires real creativity. The kind that makes people uncomfortable because it’s risky, because it might bomb and because it’s different.
And here’s where AI completely falls apart.
I’ve tried using AI for marketing brainstorming. You know what I got? A perfectly reasonable list of perfectly boring ideas.. Cringe-worthy stuff. AI can’t do cool. It spits out what statistically sounds like marketing, which is exactly the problem – it sounds like every other piece of marketing that gets ignored.
In his small talk, Dax said: “I have not had a single good idea come out of AI.” Same. Because creativity isn’t pattern matching. It’s seeing what nobody else sees and having the taste to know it’s good. I guess you can’t outsource that to an algorithm (no matter how much you want to).
The Aha Moment
So someone clicked on your product. Great.
Now comes the actually hard part.
You need to get them to the “aha moment” before they bounce. This is where most products die. Not because the product sucks. Because users never even got to see what makes it good.
The aha moment is that instant when someone gets what your product does and why it matters. When Figma lets two people edit the same design at the exact same time. When Facebook automatically fills your feed with the people you actually know. When Spotify starts playing exactly what you wanted to hear.
Your job is brutally simple: identify that one critical moment and eliminate every piece of friction between the user and that moment.
Yep. Every, Single, Piece.
Here’s what kills me. When we try a new product and they ask for my company size before I even know if I want it. Email verification before I’ve seen anything valuable. Job title. Five-click tutorial that teaches me nothing.
Each step? You’re losing 10-20% of people.
I still remember a case a former colleague who led the development taught me: most users never reach your aha moment. They bounce somewhere in the friction maze you built between signup and value. You’re not losing people because they tried your product and didn’t like it. You’re losing them because they never tried your product at all.
And here’s the painful part – you probably think all your features matter. You want to show off A, B, and C. You want users to understand your positioning. You want them to fill out that form so you can segment them properly.
Naah.
Pick the one moment that matters most. Deprioritize everything else. And be brutal about it.
Look at the most successful products. ChatGPT gives you an empty input box where you can type literally anything and get a human-like response. That’s it. Instant value. Zero friction. The dumbest person in the world can use it and immediately get what makes it special.
Facebook had the poke. Spotify had “just start listening.” These products understood something fundamental: the aha moment isn’t about features, it’s about clarity. One clear moment where the product’s value becomes undeniable.
AI doesn’t help you find that moment. Creating it requires deep understanding of your problem space. Positioning clarity. Taste – knowing what to cut and what to obsess over. You can’t prompt engineer your way to good product sense. AI won’t do that for you.
Retention
Let’s say you did the impossible. Got their attention. Got them to the aha moment. They like your product.
Now you need to keep them (ideally, forever).
If you’re not retaining customers for life, you’re running a leaky bucket. Growing on one side, bleeding on the other, until you hit a ceiling. It’s exhausting and it caps your growth.
Retention gets interesting because now you’re dealing with power users. People who’ve been using your product for months. Who push boundaries. Who want more configurability, more advanced features, more control.
And here’s the trap most products fall into: they think there’s a trade-off between simple and powerful. They say “we’re like Apple, we focus on simplicity” while actually just building something incapable.
There is no trade-off. You can build both simple and powerful. But it requires more work up front.
Instead of building features, you build primitives.
Think of primitives as the fundamental building blocks of your product. The core components that can be combined and recombined into different functionality. You design a wide, powerful layer of these primitives first. Then you assemble a simple experience on top that 99% of users see and interact with.
When users get more advanced? They get direct access to those underlying primitives. They can customize. They can bend and flex your product to their specific needs. They never outgrow you – because the power is there, just hidden under a layer of simplicity for those who don’t need it yet.
This is how open-source projects like OpenCode serve both casual hobbyists and massive enterprises. The primitives handle enterprise complexity. The interface stays simple for everyone else.
But designing those primitives? That’s some of the hardest architectural thinking you’ll ever do.
You need to understand your problem space deeply enough to break it into fundamental components that recombine in useful ways. You need to predict what users will want to do – not just today, but as they grow into power users. You need to find the right level of abstraction where primitives are powerful but not so low-level they’re useless.
AI, again, can’t help you here. The building process itself is how you figure out what primitives you need. It’s exploratory. To even explain to AI what you want, you’d need to understand it yourself first. Which means AI isn’t solving the problem – you are.
This isn’t the type of thing where you can communicate to AI what you want to do. You have to understand it really well yourself first.
Why This All Still Matters
Closer to the end of the video, Dax shared his most honest moment:
Even though I can understand all this stuff to the degree where I can talk about it and articulate it, it’s just hard. It’s very hard to follow this advice.
I guess that’s the truth nobody wants to hear. Understanding these principles is easy. Executing them is brutally difficult.
I can sit here and tell you exactly what needs to happen: create shareable marketing, eliminate friction to your aha moment, build primitives that scale. Simple, right?
Doing it? That’s where it gets real.
AI gives us new capabilities. It lets us do things we couldn’t before. But it doesn’t save us from the day-to-day grind of making something great. It doesn’t give us creativity for marketing. It doesn’t give us taste for finding aha moments. It doesn’t give us the architectural insight for designing primitives.
My life hasn’t gotten easier with AI. Things are just as hard as they’ve always been.
And that’s actually good. Because that’s where all the fun comes from. The purpose. The meaning. The difficulty is the point.
People worry that AI will take away our purpose, make everything trivial. The problem is people expecting AI to do their hard work and being shocked when their products still fail.
The fundamentals haven’t changed. Most of the work still requires creativity that breaks through noise. The aha moment still requires brutal clarity about what matters. Retention still requires deep thinking about architecture and user growth.
AI solves none of that. The winners will still be the people who understand these fundamentals and execute them relentlessly.
Just like always.
🔎 Explore more: