Executive Summary

GRAYBELT Academy is a companion project to Reflex. While Reflex addresses long-standing gaps in the field of information security, this site explores how its underlying architecture could be applied to other areas — particularly those involving professional activity that isn’t currently captured in structured form.

I’ve worked in computing since 1980, through the development of modern protocols, networking, and the early foundations of what became today’s Internet and security practices. This site reflects how that experience may still be relevant in integrating emerging technologies with long-established systems.

GRAYBELT Academy documents ongoing work that extends beyond Reflex, with the possibility of new applications based on the same architectural principles. It also clarifies what a potential acquiring company would gain if both properties are brought together.

Introduction

My name is Craig Brown, and including my resume site, I maintain four different websites. I started working in tech around 1980 as a developer and have held technical roles throughout my career. Now, being past retirement age, I am organizing my intellectual property and presenting it to the world with the hope that it will advance technology in multiple ways. You can read more about me at my resume site: csbrown.com. I’m confident you’ll be surprised you don’t already know who I am. The previous generation of tech giants have at least met me once or twice. But my life has forced me to keep a low profile, and every so often, there’s even debate about whether I’m still alive.

I’m very excited about this new project. The past 12 years have been totally centered around a product called Reflex. Around 30 years ago, I created a company called Floater Corporation. The product it sold was also called Floater, a personal finance program—but not in the way you’ve seen before. I built the first commercially available artificial intelligence engine. If you visit the GRAYBELT Innovations homepage, you can see a video of Floater in action. Shortly after Floater’s demise, I specialized in security and became a founding member of the information security profession. I founded CISOware about 12 years ago to solve the persistent, unsolvable problems in the field.

Reflex was initially designed with input from friends and colleagues from the early days of information security. Over the years, I had continued evolving the AI engine I created for Floater. Adding it to the Reflex project wasn’t about preserving old tech—it had evolved through 20 years of development. And it was well ahead of the current AI companies, which didn’t even exist at the time.

Out of courtesy, I avoid referring to Reflex’s thinking component as AI. While I have every right to, since I helped define the term in public use 30 years ago, the word “AI” now carries connotations I find limiting—similar to how the term “hacker” has been distorted. Instead, I refer to it as “real intelligence” (RI) or “natural intelligence,” the same term I used when marketing Floater. Reflex is designed to be an expert in a narrow field, but in a way today’s LLMs (large language models) can never be. Reflex has access to information that is never shared between organizations—such as the details of security incidents and the knowledge learned during response. This information is considered too risky to disclose and is never included in the data that LLMs can scrape or access.

Reflex is far ahead of current technology. I worked seven days a week for over 10 years developing it. I didn’t even track how many innovations went into it until I was required to list them all in a provisional patent application.

You can read more about the product on its dedicated site, but in summary, Reflex understands human behavior. During an incident, it tracks everything in multiple dimensions. It quantifies the people involved, the skills that led to success or failure, and uses that to make real-time predictions during future incidents. If a responder is missing during a live incident, Reflex calculates how their absence might affect the outcome. After the incident, it facilitates a lessons-learned session with unprecedented statistics. That data is archived as read-only. Another platform can then examine multiple archives to draw broader conclusions.

The intelligence function of Reflex is true intelligence, but it does not understand language. It performs specific tasks and receives questions through code.

Now, let me introduce GRAYBELT Academy. Reflex creates a model based on hundreds of thousands of incidents. It understands the nuances of each step and misstep. Originally, the plan was to offer this model to information security companies. But with the rise of language-based AI, I now envision a hybrid model where Reflex’s natural intelligence assists a modern LLM. Unlike traditional AI, Reflex processes raw human activity—not edited summaries from books or articles.

I have stayed out of the mainstream LLM space on purpose. I predicted the current AI landscape over 10 years ago and wrote about it in Reflex’s original specification. Today’s AI companies are focused on creating the “smartest” AI by processing the same datasets in better ways. But true intelligence comes from data that isn’t shared—from failures and unrecorded thoughts. A doctor who cures a disease may be published, but the messy, iterative thought process that led there is lost to AI.

Reflex is more than a product. While it can stand alone, it was designed to collect and preserve the knowledge others ignore. It includes its own email and messaging systems and creates virtual forums that exist only during active incidents. All communication is stored in the archive.

That archived text isn’t used the way a modern AI would. Reflex searches it using fuzzy logic and pattern matching—more like a sophisticated search engine than a chatbot. It doesn’t try to “think” in the LLM sense.

The purpose of GRAYBELT Academy is to bring Reflex’s core ideas to other domains. Reflex stays focused on information security. GRAYBELT Academy, which has a source code license to some core Reflex components, will explore how these systems can archive and structure human activity in other fields. Its charter prevents it from competing in security, so it complements rather than overlaps.

If Reflex were acquired by an AI company, it could produce an unmatched model for information security. The knowledge it contains cannot be duplicated by any competitor.