The Overhang Manifesto

Two wards, two responses

In 1772, Joseph Priestley discovered nitrous oxide. For the next seventy-two years, it was a party trick. Scientists would inhale it at gatherings, giggling as they stumbled around ballrooms. Meanwhile, in hospitals across the world, patients screamed through amputations, held down by assistants, biting leather straps.

No one connected the dots.

It wasn't until 1844 that a dentist named Horace Wells watched a friend inhale nitrous at a stage show, slam his leg into a bench hard enough to draw blood, and feel nothing. Wells' response was: why aren't we using this for surgery?

The answer is: no reason. It just hadn't occurred to anyone to ask.

Now contrast this with insulin.

In 1921, Frederick Banting and Charles Best isolated insulin from dog pancreases. In January 1922—less than a year later—doctors walked into a ward of dying diabetic children at Toronto General Hospital and began injecting them.

One by one, comatose children woke up. Parents who had come to say goodbye watched their doomed children stand and walk out of the ward. A disease that was 100% fatal became manageable. Lab to bedside in months.

Same type of discovery. Radically different response time.

The difference wasn't technology

The gap between nitrous and insulin wasn't about capability. Both were available the moment they were discovered. The difference was recognition—seeing the overhang and acting on it.

This is technological overhang: the gap between what's technically possible and what we're actually doing with it. It's the distance between discovery and application. And it's almost always longer than it needs to be.

We're in an overhang right now

Since late 2022, AI capabilities have advanced faster than any technology in history. Models that can reason, code, analyze, and create are available to anyone with an internet connection.

Yet most organizations are still treating AI like nitrous oxide at a party—fun demos, impressive tricks, but not fundamentally changing how work gets done.

The question isn't whether this will change things. It's who will be the ones connecting the dots, and when.

What I'm doing about it

This site is my attempt to map the current overhang. Not theoretically, but practically:

What can we build today that was impossible two years ago?
What problems are newly tractable for people with moderate resources?
Where are the opportunities that others haven't recognized yet?

I'm particularly interested in what I call Avenues of Investigation—domains where the technological overhang is significant and where different types of contributors could make genuine impact:

Desalination — 2 billion people lack clean water, and small-scale solar-powered systems are newly viable
Landfill robotics — Precious metals buried in waste, newly extractable with vision + manipulation advances
Solar-enabled applications — 90% cost reduction since 2010 unlocks combinations nobody's tried

For each avenue, I try to identify:

What interested amateurs could do with limited time and AI support
What small teams might tackle with modest funding
What well-resourced entities could pursue at scale

The ward is filling up

Every day that passes with AI capabilities sitting unused is another day of unnecessary difficulty—maybe not screaming patients, but certainly problems that don't need to be problems anymore.

I'd rather be the 1922 doctors than the 1800s party guests.

Interested in exploring similar ideas? Get in touch.