On Unintended Consequences
The “A.I. Flood”: An Abundance Beyond Review
A.I. is making it easier than ever to create information. But now, there’s a different bottleneck.
“Everything should be made as simple as possible, but not simpler.”
— attributed to Albert Einstein
A Reasonable Request
A friend asks me to review an open-source software project he’s been working on.
He’s brilliant, and I’m excited to see what he’s built. And what he’s built — and how fast it was turned around — is really impressive.
It’s pretty genius. I’ve never seen anything like it.
I also notice how very comprehensive the documentation is. And soon, I’m finding … more documents. And then, even more. There’s a lot to read.
I thought I’d be well into the code by now — and using it. But the process is taking a lot longer than expected.
Still, it’s nothing compared to other projects I’ve seen lately. And that’s when I start realizing the issue isn’t really about any one project at all.
It’s a much broader trend.
I start to think that things that are simple should just feel … simple. Not overwhelming.
Though too often, nowadays, I’m beginning to feel like an archaeologist.
And the more I think about it, the less this seems like a software problem.
It seems like a whole new pattern.
“A.I. Slop” vs. “A.I. Flood”
People have been using the term “AI slop” to describe low-quality AI-generated content. That’s a useful term, but it describes a quality problem.
What I’m noticing feels different. It’s more of a quantity problem.
Even when AI-generated output is useful, there can be so much of it that human understanding becomes the bottleneck.
I’m calling this the “A.I. Flood”.
- The good news is that AI can easily generate code, documentation, diagrams, and many other kinds of written work.
- The bad news is that, at some point, someone still has to read, understand, and approve all that.
The Bottleneck Has Moved
Recently, courts have reported growing numbers of AI-assisted filings from self-represented litigants. Some contain mistakes. Some contain fabricated citations.
But even legitimate filings are consuming much more time and attention than before.
The filings are now easy to produce.
The judges, unfortunately, remain on the original hardware.
It used to be that the number of filings was constrained by the cost of hiring a lawyer. But in this A.I.-powered do-it-yourself era, that’s no longer the bottleneck.
Because a court’s capacity to review cases … doesn’t automatically scale just because documents become easier to generate.
So, there’s still a bottleneck — it’s just moved.
The obvious response?
That A.I. will come to the rescue! It can review all of it too!
Maybe. Probably. Eventually.
Though it’d help if the second set of hallucinations weren’t coming from the same gizmos that generated the first set.
So let’s just say … someday.
But meantime, somebody still has to sign off on the work.
And whoever does probably ought to understand it.
And yes, the role of humans may increasingly shift from creation toward supervision. To a higher level of abstraction.
That does sound easier … until you remember that supervision also requires time, attention, judgment, and understanding.
And by the way: expectations of our output easily scale without limit.
So, the more we get done — the more there is to do.
It’s called “The Productivity Trap.”
But I digress …
The Price of Complexity
One of the oldest lessons in engineering — good engineering, that is — is that simplicity requires discipline.
A hundred-page report may contain wisdom.
It may also contain ninety pages that nobody had the discipline to delete.
Historically, complexity carried a cost because it was difficult to create.
Today, complexity is becoming cheap.
Understanding it is not.
Every artifact we create comes with an invisible obligation attached.
Someone eventually has to understand it, review it, maintain it, govern it, or decide whether to trust it.
Those costs don’t disappear. They merely move downstream.
And sometimes that someone is “Future-you.”
What Remains Scarce
Human attention remains stubbornly finite.
Which means that judgment, simplicity, and wisdom become even more valuable.
Not because information is scarce.
But because it isn’t.
The challenge is no longer producing more.
It’s deciding what deserves human attention, and how much.
Being deliberate about what we create.
And shifting the focus from generating more … to generating what is useful.
Author’s Note:
It seems appropriate to acknowledge that this essay about the “A.I. Flood” was itself developed with the assistance of A.I. 🤖
The process was not one of automated generation, but active collaboration:
the A.I. helped me explore ideas, challenge my assumptions, identify counter-arguments, and propose language — while I reviewed, revised, accepted, rejected, and/or rewrote every section.
The final essay reflects my own views, and I remain responsible for its contents.
© 2026 Ron Diamond. All rights reserved.
