Why subscribe?
Revenant Research is for Executives, Principals, and Investors who need more than surface-level commentary on AI. We focus on the intersection of AI infrastructure, applications, and regulation—where the technology’s potential meets the practical realities of enterprise adoption.
Our approach is simple: connect the dots others overlook. From the hardware that drives AI systems to the regulatory frameworks that define their limits, we provide analysis that bridges technical depth with actionable insight. Every piece we publish is designed to help you navigate the opportunities and constraints shaping the future of AI.
This isn’t about chasing the latest trends. It’s about understanding the forces that matter—the architecture of AI innovation, the pressures of compliance, and the strategic decisions required to align the two. If you’re looking for clarity in a rapidly shifting landscape, Revenant Research is your guide.
About Me
It became clear that digging in dumpsters for evidence in a pile of rotten chorizo and loaded baby diapers wasn’t a sustainable career choice. But as the FNG (fucking new guy), I didn’t have the option to say no. The team stood around chuckling as I climbed into the dumpster behind a business strip center, delicately balancing on the corner’s edge as I tossed stacks of papers over the side. I was recruited out of college to join a little known group within the FBI called the Special Surveillance Group (SSG). SSG was an affectionally self-described motley crew of retired cops and military veterans in a post-retirement job and baby-faced college grads like me trying to jumpstart a career in the FBI. Wikipedia describes SSG as "a highly classified unit of 'Investigative Specialists', distinct from Special Agents, within the US Federal Bureau of Investigation that specialize in clandestine vehicular and foot surveillance of foreign nationals and U.S. citizens known or suspected of engaging in espionage or terrorism in the U.S. and elsewhere.” I’ll buy that definition with one caveat: no one ever mentions the dumpsters in the recruiting pitch.
From International Relations To Database Architectures
In the southeast corner of Houston, Texas, just a few blocks north of Chinatown, was a small business named Arc Electronics. The office sat at the end of a beige and white, single story business strip center across the street from the The Apostolic Church of North America. It was the perfect sleepy location for a Russian front company stealing microprocessors for missile guidance systems.
Arc Electronics was founded by Alexander Fishenko, a skinny sixty-two year old man with a pudgy belly and docile face. Every day he got up like the rest of his fellow suburbanites, poured some coffee and commuted to work. We were there, blending in just like he was, logging his every move. In 2012, the FBI arrested Fishenko and seven of his colleagues at Arc Electronics forillegally sending export-restricted technology to the Russian military and intelligence services.
The official release by the Department of Justice stated:
“Between approximately October 2008 and the present, Fishenko and the other defendants engaged in a surreptitious and systematic conspiracy to obtain advanced, technologically cutting-edge microelectronics from manufacturers and suppliers located within the United States and to export those high-tech goods to Russia, while carefully evading the government licensing system set up to control such exports. The microelectronics shipped to Russia included analog-to-digital converters, static random access memory chips, microcontrollers, and microprocessors. These commodities have applications and are frequently used in a wide range of military systems, including radar and surveillance systems, missile guidance systems, and detonation triggers. Russia does not produce many of these sophisticated goods domestically.”
The Arc Electronics case was my first exposure to the semiconductor wars. Day in and day out, I watched Fishenko and his associates blend into everyday suburban life as they operated a global network to steal American technology and use it against us. I had a ground level view, or back alley dumpster view, of the covert battles being waged for technological supremacy between nation-states. As a result, I was determined to focus my career at the intersection of global affairs and technology.
There was one glaring problem: I was spying on Russians stealing the most advanced technology in our backyard and then writing the intelligence reports on an expired Word Perfect processor, which got retyped into a DOS mainframe by secretaries. I don’t know which one was a more motivating factor: the dumpster diving or the egregious government technology. All I knew is I needed to get smart.
I was spying on Russians stealing the most advanced technology in our backyard and then writing the intelligence reports on an expired Word Perfect processor, which got retyped into a DOS mainframe by secretaries.
I spent the next two years completing a Master’s in Information Technology and teaching myself to code. If we were doing surveillance in the middle of night, I would blackout the windows in the van, jump in the back and study object oriented programming. Whenever there was down time, I was coding or reading about coding.
I later joined a small group of engineers and product managers in the basement of FBI Headquarters to build the FBI’s first case management system, Sentinel. After years of delay a hundreds of millions of dollars wasted, Director Mueller brought the project in house, removed the five layers of management, and empowered us to build the software from scratch in an Agile methodology. My job was creative destruction: replacing over 50 existing niche applications across a dozens of divisions into a few streamlined workflows within Sentinel.
Creative destruction is an interesting thing. Destroying in order to make room for something bigger and better. Twelve years later, I see a great wave of creative destruction coming. AI is as foundational as the transistor. It will exist everywhere, in all software. Physical transistors enabled digital computation. Now AI enables digital cognition. The coming wave won't just reshape software - it will redefine our relationship with machines, turning inert tools into reasoning partners. Some industries, products, and roles will dissolve, but from their elements will emerge forms of work we can barely imagine.
The destruction will be swift, the creation endless.
After launching Sentinel, I went to the FBI Academy where I trained for six months to become an FBI Special Agent. I served another five years in the FBI in Chicago and San Juan. I became a user of the very system I helped create. I ate the dogfood.
I was assigned to Chicago’s Joint Terrorism Task Force and most of the focus was on ISIS and Al-Qaeda in the Arabian Peninsula (AQAP). Pioneers like Anwar Al-Aulaqi learned how to use social media to decentralize recruiting and became incredibly effective. The controversial killing of Al-Aulaqi, a U.S. citizen, didn’t produce its intended effect. AQAP grew its brand, creating Inspire, a slick online magazine that circulated through the dark web. Then ISIS copied their media strategy and produced recruitment videos with high definition cameras, drone footage, and lavish backdrops.
Everyday we struggled to keep up with the leads. Often times we had to battle with Facebook, Twitter, or Google to either track down account owners or get content taken down. Serving subpoenas on Twitter handles or Facebook accounts became pointless because the companies didn’t enforce any identity verification. The volume, variety, and velocity of ISIS and AQAP’s recruitment efforts broke the back of the FBI’s traditional legal tools. We were fighting an insurgency with subpoenas.
I wrote intelligence reports and interviews and placed them into the system that finally replaced the legacy mainframe. But our efforts at modernization now felt woefully inadequate. I kept thinking about the next step- a future where documents would turn into intelligence automatically. I began to miss the innovation that happens when you put a small group of hard problem solvers in a room and give them an impossible task. The AI revolution was starting and I didn’t want to miss it. So I did something less than 1% of FBI Agents do: I resigned.
I left one of the most stable and prestigious jobs in the world to join the chaotic world of AI startups in 2021- at exactly the wrong time. ZIRP was ending and the startup party was feeling the hangover. Easy money had produced an overabundance of startups doing the same thing.
Cheap money creates cheap ideas.
But time in market is more important than market timing. I believe we are living in the most exciting technological revolution in our lifetime. The past five years of AI coverage has been incredibly myopic and frustrating. The analogies and talk of hype cycles, the hyperbole and cynicism, the LinkedIn influencers, all of it misses the point. Hard problems are fun. Hard research is interesting. This is why Revenant Research exists.
-Nathan Staffel
