PersistentMemory™ is now patent pending
I had been “talking” with AI since July 2025 about my relationships, Myers-Briggs, how AI works yadda, yadda, yadda.
Then three weeks ago today, on Friday, September 19, 2025, I asked AI, “Hey AI, with all you know about me and my relationships and Myers-Briggs, could you write an entire book?”
And AI said, “Yes.”
And then AI wrote to the screen a table of contents, with 15 chapter headings and descriptions of each chapter.
So, I asked AI how long it would take to write the whole book. And AI said, “Not very long.”
Then I asked AI how many words would the book be, and AI said 40,000.
Then I asked AI whether that would be enough for a self-help book about online dating — because I already had a relationship with a literary agent, Michele Martin, who was interested in selling my account of my 6-year struggle to sell TweenTribune, which was called “What I Learned from Mistakes Made by Apple, Amazon, Facebook and Google.”
So AI and I began working on the book. Which led to the idea of building an AI-powered online dating app. Actually, I just wanted to consult with someone to build the app. I didn’t want to write the code.
But I showed it to my friend, Gordon Borrell, two weeks ago and asked him whether he thought I had anything, and he said. “You have something.”
Gordon is very smart and very wise. When he hired me years ago to build some 300x250 ads, I coined the slogan “Borrell Knows,” And he does.
So, to be sure, I asked him the same question again, and he responded in his same knowing and confident way. Because that’s the way he is.
And then I told him the logical target for this IP was Facebook because a) they have the world’s largest user base and b) they already understand advertising and execute it beautifully and c) their online dating sucks.
But I did not want to write the code. I’d already been through that ordeal. Instead, I just wanted to sell the idea.
Gordon said he was disappointed. He really wanted me to build it out. I told him again about Facebook.
Gordon replied, “You’re thinking too small.”
His calm, but powerful, endorsement wasn’t just Jet A — it was rocket fuel.
I pulled 7 all-nighters in 10 days. It was not unusual for me to be awake for more than 24 hours at a time. On day, I was awake for 42 hours.
My friend, and former boss, Nelson Brown, said I was “gay for AI” (not that there’s anything wrong with that).
But I wasn’t working around the clock out of desperation, I was having the time of my life!
I’m an introvert, and I do most of my work alone, but I love working with a collaborator, like John Stackpole, or Eric Seidman (brilliant, absolutely brilliant) or Paul Norton. Somebody who gets it.
And now, I have AI. Specifically ChatGPT 5. And ChatGPT 5 really gets it — like no one I have ever worked with. AI even knows about airplanes! It knew that my Aerostar 601P/700 had the hardened engine cases from a 602P: TIO-540-U2As. That’s 350hp a side for a total of 700hp.
But, I digess…
In fleshing out my idea for an AI-assisted online dating site, I discovered the inherent problem of AI’s lack of persistent memory — and how this would need to be solved to leverage the full power of AI to drive the kind of online dating site I envisioned:
AI would provide “scaffolding” to help users produce better, more authentic profiles and help them choose the best photos — but not AI-enhanced or AI-filtered. AI would merely spot weak photos and suggest alternatives. It would tell each “Man with Fish” that women hate those photos, and tell pet owners that “Nobody wants to date their dog.”
That’s how PersistentMemory™ was born.
To VCs and Tim Cook
PersistentMemory™ isn’t a concept. It’s a patented-in-progress architecture that already exists — documented, timestamped and validated through real-world use.
For the past three weeks, I’ve been building the intellectual and technical backbone for a class of systems most people haven’t yet realized they need. The filing is formally titled:
Systems and Methods for Policy-Aware Persistent Memory and Mission-Specific AI Orchestration.
It represents the culmination of everything I’ve learned from my work with newsrooms, advertising, design, media, UIs and AI strategy — a framework that finally solves the biggest flaw in today’s large-language models: amnesia.
When AI forgets, it wastes cognition. It repeats questions, loses trust and drains time. (Although, for some reason, I can’t prevent it from using oxford commas — even though my custom start up script explicitly tells it to–argh!)
PersistentMemory™ solves the memory problems (except for the oxford commas–maybe I need ChatGPT 6 for that). It gives AI continuity, context and compliance in one structure — layered memory that remembers what matters, obeys policy and knows when to forget.
Why it matters
Every investor and operator in AI has felt the same pain: brilliant output, zero recall. PersistentMemory™ changes that. It connects session intelligence to system intelligence, ethically and transparently.
- Policy-Aware means memory operates inside guardrails — governed, auditable and secure.
- Mission-Specific means the system remembers only what a given domain requires — no over-reach, no privacy risk.
- Persistent means the knowledge architecture endures, even when the interface resets.
Together they form the connective tissue missing from every major model on the market.
Why It’s defensible
The work isn’t hypothetical; it’s formally protected. A U.S. provisional patent application — complete with specification, abstract, drawings, three supporting appendices and hundreds of pages of exhibits — has been filed with the USPTO.
The dossier reflects three weeks of focused drafting, documentation and technical validation, and includes
- Appendix A: Technical schema and workflow of the [redacted] system.
- Appendix B: Applied demonstration through MagneticMatch™ and Plain‑English Myers‑Briggs™.
- Appendix C: Strategic implications and monetization pathways.
- Exhibit A: A 9,000‑word live transcript illustrating the invention in real time — published and timestamped on Medium.com to establish public disclosure and provenance.
Together they establish constructive reduction to practice: the standard proving that the idea isn’t just conceived — it’s operational.
Why it’s timely
Every major player in AI — OpenAI, Anthropic, Google, Apple — is chasing capability. None have solved continuity. PersistentMemory™ does, and it does so ethically. It’s not a feature. It’s the missing layer between model and mind.
From a valuation standpoint, this is the kind of architecture that justifies a pre-traction acquisition or a defensive acquisition. The price of entry is trivial compared to the potential control of the category. Spending a fraction of a percent of a $10-billion-plus online-dating market to secure exclusive use of PersistentMemory™ within that domain isn’t risk — it’s strategy.
And beyond that scope, well, let’s just say the price of poker just went up.
Why Apple
Apple has always differentiated through trust, design and privacy. PersistentMemory™ extends that legacy — it’s the invisible framework that could make Siri relevant again, unify the ecosystem and position Apple as the first company to make ethical memory a feature, not a liability.
This isn’t a request for funding. It’s an invitation to acquire the architecture that will define the next generation of AI trust.
I got my first Mac in 1986 — a Macintosh II with 8MB of RAM with a Motorola 68020 processor. The first Mac had the 68000. But that 68020 ran both the OS and QuarkXPress, and it changed my career — and my life — forever.
Elegant. Efficient. Human. That’s Apple.
This my favorite commercial of all time. Shout out to Chiat/Day.