I’m not crazy. I’m just early.

Col. Billy Mitchell predicted Pearl Harbor. In 1924.

The year was 2006. The iPhone did not exist. Facebook was losing millions — it was still years away from profitability.

But I could see the writing on the wall: Newspapers were doomed.

First, I sounded an alarm. Then I offered a radical, extensive, 10-point survival plan. I called it “NewRulesForNewspapers.”

It seems prophetic, in hindsight.

I was mocked, ridiculed, and ignored. The New York Press Association stiffed me on a $1,500 speaking fee because they were offended by my message.

I wasn’t crazy. I was just early. Newspapers did crumble, just as I had predicted they would.

This is nothing new for me. Because I’ve been Stirring Shit Since ’74™

But back to the story…

After two years of frustration, I said “Fuck it. If newspapers won’t take my advice, I’ll just do it myself.”

I partnered with two newspaper-industry thought leaders, Jay Small and Janet DeGeorge, and launched a mashup of YouTube and Craigslist.

Jay used Drupal (think WordPress on steroids) and a “cranky” video processor, FFMPEG. We used AWS to reduce storage cost of the videos. He wrote a script that coordinated storage and retrieval between our site hosted by Rackspace and AWS.

We did a video site before 5G. Before 4G. We were thrilled when we found 3G.

It got named-checked by WaPo and WSJ — twice. I met with Japanese investors. I did a segment on NPR with Craig Newmark.

But I never gained traction — not until I launched another site that I eventually sold for millions.

I wasn't crazy. I was just early.

And now, it’s happening again.

This time it's AI. This time it's governance. This time it's revenue models and risk structures and the false stability of billion-dollar bets placed on unstable foundations. And again, I’m the guy with the warning. With the receipts. With the timestamped record.

They’re laughing now. They’re calling me a “crank.”

But I know what comes next. Because I’ve been here before.

I'm not crazy.

I'm just early.

Col. Billy Mitchell antagonized much of the Army brass with his arguments and criticism. In 1925, his temporary appointment as a brigadier general was not renewed, due to his insubordination, and he reverted to his permanent rank of colonel.

Later that year, he was court-martialed for insubordination after accusing Army and Navy leaders of an “almost treasonable administration of the national defense” for investing in battleships.

He argued for the ability of bombers to sink battleships and organized a series of bombing runs – less than a mile from where I live – against stationary ships as a demonstration. But still his message was not embraced. He resigned from the service shortly afterwards.

Mitchell spoke truth to power – an unforgivable sin in the military.

But wait.

There’s more.

In 1924, he produced a 324-page report that predicted future war with Japan, including the attack on Pearl Harbor.

He wasn’t crazy. He was just early.

The system is breaking, and I’ve seen it firsthand

For years I’ve said I was early, not crazy.
And now I have the receipts — not philosophical, not theoretical, but lived.

Every major piece of technology I’ve touched in the last six months has confirmed a simple, unsettling truth:

The entire software ecosystem is collapsing under its own weight.

I’m not talking about bugs or annoyances.
I’m talking about systemic rot — the architectural decay that happens when companies get too big, too siloed, and too afraid to rewrite anything.

And I know this because I just lived it.
Not ten years ago.
Not in theory.
This week.

Here are the systems I fought.
And I won.
But I shouldn’t have had to.

Google Analytics: the tool that forgot how to be a tool

GA4 is the clearest example of what happens when a product team loses the plot.

The GA I used five years ago was fast, intuitive, and powerful.
The GA I used this week behaved like a maze built by a committee that never met each other:

settings hidden behind unlabeled panels
metrics renamed with no rationale
essential features buried or removed
navigation that defies logic
workflows that break every principle of information architecture

It took me half an hour to find something that used to be one click.

You don’t do that by accident.
You do that when architecture collapses.

LinkedIn advertising: The UI that actively prevents money

I spent hours trying to give LinkedIn money.
Not because I wanted to — because I needed to.

And do you know what happened?

LinkedIn made it almost impossible.

The UI:

switches between four different ad builders
hides targeting options behind invisible toggle states
disables features depending on which random sequence of clicks led you there
removes basic functionality if you enter the “wrong” builder
provides zero clues about where you are

It took me hours to find Company Name targeting — a feature that should be visible in three seconds.

I wasn’t confused.
The UI was broken.

If LinkedIn fixed its architecture, they’d make 20 percent more revenue immediately.

But they won’t — because the system is too big to fix from the inside.

MacOS and iOS: elegance replaced by incoherence

Apple used to stand for clarity.
Minimalism.
Predictability.

Now?

Try changing a notification rule.
Try finding a system preference.
Try doing anything involving iCloud, focus modes, or cross-device settings.

Everything is nested inside everything else.
Everything is abstracted.
Everything feels like it was designed by seven separate teams who never compared notes.

I’ve used Macs since before the iPhone existed.
Never in my life have they felt this messy.

The rot has spread everywhere.

What does all this mean?

It means this:

The world is drowning in incoherence.
The tools are failing.
The architecture is gone.

Big companies are no longer able to design systems — they just patch them.

And here’s the irony:
People think I’m building my AI architecture stack because I want to create complexity.

No.
I’m building it because everything around us has already become too complex for humans to manage.

AI needs guardrails because the systems humans built are already out of control.

PersistentMemory.
TIS.
IDK.
TE.
Context Window.
AI$

They’re not abstract inventions.
They’re solutions to the exact structural failures I just spent a week navigating — failures that did not exist five years ago.

I’m not crazy.
I’m not imagining things.
I’m not catastrophizing.

I’m just early.

And the world is catching up to the thing I saw years ago:
when systems lose coherence, collapse is inevitable unless you redesign the architecture from the ground up.

I’m not trying to control the future.
I’m trying to fix the present.