America's Biggest Itches Nobody Is Scratching

A few days ago I read that Microsoft quietly cancelled thousands of AI coding licences, the same tool their engineers loved using. Not because the tools did not work. Because they worked too well.

Engineers used them constantly. The bill got out of hand. Uber, around the same time, burned through their entire 2026 AI budget in four months. Their whole year's budget. Gone before summer.

That stopped me. We have been told AI is the great equaliser, the thing that lets a 3-person startup move like a 300-person company. But if Microsoft, with a $3 trillion market cap, cannot sustain the cost of AI at scale, what does that mean for everyone else?

I sat with that for a while. And the more I thought about it, the more it felt less like an AI story and more like a symptom of something bigger.

Then I Remembered the J-Curve

I have been going through some startup masterclasses recently, and one concept kept coming back to me: the J-Curve. The idea that most startups do not die because their idea was bad. They die in the dip, the valley between early momentum and sustainable scale, where costs spike, complexity multiplies, and "this is working" quietly turns into "we cannot keep this up."

The Microsoft AI story is a perfect J-Curve moment. The tool worked. Adoption shot up. Costs followed. And right when things should have stabilised, the math fell apart.

And I got curious. Is this just an AI thing? Or is the same valley showing up everywhere (healthcare, immigration, SMB operations), where the gap between "this exists" and "this actually works at scale" is where companies quietly go to die?

So I went looking.

How I Looked, and What I Borrowed

I did not want to just collect complaints. Anyone with a Reddit account can do that. I wanted a way to actually rank these problems. Which ones were urgent, big, and genuinely unsolved versus which ones just felt loud?.

That was when I came across a framework called Fix My Itch, something Razorpay built brilliantly for India. They scraped Reddit, X, and forums for real problems, then scored each one on severity, frequency, and market size. The "Itch Score." It is such a clean idea, and as far as I can tell, nobody has applied it to the US market yet.

So I borrowed it. Three inputs, simple math:

5 Gaps That Stood Out

I went through dozens. These five had the highest Gap Scores. Weirdly, they all started looking like variations of the same underlying problem. Click any card to see the breakdown.

What I Noticed Across All Five

Here is the thing that surprised me most: none of these gaps are about missing technology. The AI tools exist. The healthcare mandate exists. The immigration system exists. The enterprise platforms exist.

What is missing in every case is the layer that makes the existing thing usable, at the right price, at the right scale, for the right kind of company. Microsoft has AI but no way to control what it costs. Doctors have software but no way to escape the paperwork. Startups have tools but the ones built for them assume they are either enormous or tiny, never the in-between.

Every gap I found sits in the same valley, the one between "this exists" and "this actually works for me."

That is the J-Curve at industry scale. And it is where I think the next interesting wave of companies will get built.

Where I Am Going With This

This is just one afternoon of curiosity turned into something. I went down a rabbit hole, found a framework I loved, applied it to a different market, and noticed a pattern I genuinely did not expect.

I would love to keep doing this. Go deeper on each industry, refine the scoring, talk to people actually living inside these gaps. If anything here resonated, or if I got something wrong, or if there is a gap I missed entirely. I would really love to hear it.