I've been thinking about how people talk about AI in companies, and honestly? Everyone's got it backwards.
You go to any VC firm right now, any startup, any enterprise team, and they're all asking the same thing: how do we use AI to be faster? Faster memos, faster research, faster decisions. As if speed was ever the real bottleneck.
Here's what I think is actually happening. AI is going through the same phase that every general purpose technology goes through before it disappears. Remember when companies put “we use the internet” in their pitch decks? “We use cloud.” “We use mobile.” Eventually you just stop saying it. The technology becomes infrastructure, present everywhere, noticed nowhere. AI is heading there fast. “We use AI” will sound as impressive as “we use Slack.” Which is to say: not at all.
So if everyone has the same tools, what's the actual differentiator? I've become convinced it's something people don't talk about enough, because it's hard to demo and impossible to hack together in a weekend: whether the system actually learns.
Most organizations are amnesiac
Most organizations I've seen are amnesiac. Not stupid, often quite the opposite. Full of brilliant people with rich mental models of their domain. But the organization itself? It remembers almost nothing in any useful form.
Watch what happens. A partner leaves after ten years. A decade of pattern matching, of subtle market intuition, of why we passed on this in 2019 and were right, just evaporates. Not because anyone wanted it to. Because it was trapped in a brain that walked out the door.
A founder call happens. Someone takes notes. Those notes go into a folder that no one will ever search, because search returns garbage when you don't know the exact keywords. The insight dies in a document graveyard.
Four different people, over six months, notice something about a market. Each tells someone. No one connects the dots. The information existed. The pattern didn't.
This is the thing. We have information. What we lack is structured belief formation. The ability to say: here is what we currently think, here is why, here is what would change our mind. Most companies can't do this at all. They have opinions floating around, contradictions no one resolves, “lessons” that get relearned every two years because no one wrote them down in a form that survives context collapse.
A search engine tells you what happened. A memory system tells you what happened before. But what you actually want is something more ambitious: a system that can answer what do we believe, why do we believe it, and what would prove us wrong?
The goal isn't perfect recall. The goal is continuously updating your model of reality. A company that remembers everything is useful. A company that learns from everything becomes very hard to compete with. The compounding is different.
Memory is linear. Learning is exponential.
Nobody cares about your features
Now, I've watched enough product cycles to be skeptical of the feature chasing that dominates most markets. Every founder I meet wants to tell me about their special feature. “We do X that Competitor Y doesn't.” “Their integration is worse.” I nod along, but I don't think users care nearly as much as founders need them to.
There are probably hundreds of note taking apps technically superior to Notion. Better databases, cleaner architecture, more powerful queries. People still use Notion. Same with Google Forms, objectively janky, universally used. Same with Apple. Apple didn't win because nobody else could build a phone. They won because you buy the thing and assume it will work. That's the whole proposition. Not features. Trust.
I think AI infrastructure ends up the same way. Everyone will build roughly the same thing. Memory, agents, context, search, whatever the buzzword of the month is. The surface will converge. The difference will be whether people trust it.
This is where it gets subtle and important. When the system says someone is blocked, is that actually true? When it says nobody's working on this, can you believe it? When it says we made a decision three months ago, is it right, or is it hallucinating a meeting that never happened?
Because here's the thing about trust: it's not about being right most of the time. It's about being reliably right, and transparently wrong. If you're second guessing every answer, the product is useless no matter how powerful it is. The cognitive overhead of verification kills the value.
The winners will be the companies people stop thinking about. You ask, you get an answer, you move on. Like a spoon.
This sounds like a stupid example, I know. But think about spoons. Solved technology, thousands of years old. Same job, same material, same purpose. Yet some spoons feel terrible and some feel perfect. The difference isn't some genius feature. It's a thousand tiny details nobody notices individually: the weight distribution, the curve of the bowl, how the handle sits in your hand. Each detail is invisible. The aggregate is everything.
Most products are like this. The distance between good and great is rarely a single decision you can point to. It's the accumulated texture of this just works.
The right to build
Cursor is the example I keep coming back to. They started with something new, AI native coding, and when competitors copied the obvious surface, Cursor went deeper. The whole proposition became: it just works. Reliable performance across all models, every time, everywhere. That's it. That's the product.
And here's the fascinating part: Cursor is expensive. Wildly expensive compared to subsidized alternatives. People pay anyway. Because they've established the trust. When Cursor does something, you don't spend mental energy verifying. You just use the answer and keep going.
Now they're launching Git origin, a GitHub competitor. I'm sure they'll nail it. Not because the features will be radically different, but because they've earned the right to build. They've convinced a generation of developers that if Cursor built it, it's the most reliable way to do that thing. That's an asset you can't buy or hack. It compounds.
Similarly, I think what…