AI Makes Everyone Better and Everything Worse

Something strange is happening in the market. The tools have never been more powerful. The output has never been more plentiful. And yet, a growing number of founders, designers, and engineers are quietly admitting that the products they’re shipping feel hollow. Soulless.

We’ve been sold a story: AI makes you 10× faster, so you can run five projects at once, ship in days instead of months, and outcompete everyone still working the old way.

Look closely at who’s loudest about AI productivity gains. Most were never deep experts in their field to begin with. They’re using AI to replace knowledge they never built – to ship faster, not better. They’ve trapped themselves in a loop of output without learning, speed without value.

The majority of those hyping AI productivity never mention the success of their products – only the speed of their shipping.

“Good Enough” AI Standard

When every founder can spin up a passable product in a weekend, what happens to the market? The floor rises. The barrier to entry collapses. And for a brief, exciting moment, it looks like democratization.

But something else happens simultaneously, the ceiling drops.

MVPs – minimum viable products – were always prototypes. Things meant for testing, not for selling. Now, with AI, everyone is shipping MVPs like they are production-ready products at premium prices. The market is receiving a flood of broken and ugly things dressed up as finished ones. Hype is replacing quality and beauty. Speed is replacing care.

Economists have a name for this dynamic: the market of lemons. When customers can’t distinguish quality from mediocrity before purchase, the bad drives out the good. We are watching this happen in real time across software, design, and digital services.

Here is how AI affects the market:

• 5x more “MVPs” shipped per team since 2023
• Average retention declining in AI-heavy product categories
• Makers with genuine craft still command premium outcomes

Happy people make high-quality products. When the joy of working is removed – when everything becomes prompts and outputs, speed and volume – quality follows. The army of disengaged experts who have stopped caring about their work is not a minor cultural footnote. It is a structural market problem.

Genuine craft is becoming scarce. And scarcity, in a market economy, is the engine of premium value.

AI Can’t Manufacture Care

When a designer or developer doesn’t enjoy their work, care disappears. And care is the most important input to anything a person will actually want to use. AI cannot manufacture care. It can only approximate its outputs.

Cognitive load research adds another dimension which is that human brain cannot genuinely manage five projects with help of an AI at once at quality level. Experts who tried report a specific kind of burnout – wrong prompts going into wrong projects, context bleeding between work, the quiet erosion of judgment under too many simultaneous demands. Multitasking at scale isn’t a productivity hack. It’s a slow leak before everything breaks apart.

There’s a more important distinction being lost in the AI hype: these tools are not “thinking” in any meaningful sense. They are large language models – reasoning engines. They can reason with data given to them, but they cannot think, and they should not be let to make decisions. The moment we hand them decision-making authority, we’ve replaced expertise with probability, and probability calculations are not facts. That gap is where damage happens.

Here is a real example how AI can hurt your business if you replace experts with it:

We had several clients come to us in panic. AI tools - Gemini, Copilot, and ChatGPT - had been generating constant crisis-level alerts, keeping them in a state of sustained stress about their own websites. The AIs insisted their sites were invisible to search engines, that something was critically broken.

We pulled up Google Search Console. We showed them their listings, live, in multiple search engines. Everything was fine. What had actually happened: AIs had tried to access the sites' admin areas, been blocked by server-level protection, and then — rather than simply reporting they couldn't retrieve data - began manufacturing explanations. They reasoned their way with lack of data and understanding of entire environment into a lie. And all of them advised our clients to go into their databases and delete things. To remove pages with SEO scores. The advice, followed without expertise to filter it, could have destroyed years of legitimate work.

People trust these tools because they don’t understand them. And that misplaced trust, amplified at scale, is one of the least-discussed risks of the current moment. When people who don’t know how to use a drill pick one up, nothing happens. When people who don’t know how to use AI pick it up, the tool uses them.

Quality comes from judgment, taste, and genuine expertise. No prompt can replace those.

The AI Opportunity

The experts who use AI as a tool – not as a substitute for taste, judgment, and domain expertise – are about to become extraordinarily rare. They’ll produce faster than the old professionals and better than the AI-dependent new professionals. That combination is the most defensible position in the next five years of design and development.

The question is not “how fast can this team ship?” It should be: “do customers come back?” Speed metrics are being gamed everywhere. Genuine connection is not.