From L.L. Bean Mainframes to AI Safety: A Founder’s Story

By: Stephen Calhoun, CEO

I’ve been called 'too early' on the Internet of Things (1999) and wearable fitness gaming (2011). Today, I’m building SASI because the history of technology tells me that when a system becomes this powerful, the most valuable piece isn't the engine, it's the brakes.

Below is the story of how a kid from a Maine server room ended up building the safety middleware that the FDA, EU Act, and insurance underwriters are now calling for.

Roots in the Server Room

Born in Okinawa in 1971, my mother was a nurse and my father chose the computer programming path. After returning home to Maine, my father worked as a programmer for a few banks and then eventually settled into a programming position at L.L. Bean in Freeport, Maine, designing one of the first mail order databases. I remember playing on the keypunch machine, watching the giant dot matrix printers spin out multi-part documents into many bins to be distributed to the other programmers for review. The reel-to-reels spun up with magnetic tape hanging in giant loops falling and rising in the glass refrigerator size machines. The huge hard drives the size of a tire were hefted into the IBM readers. I learned how to look for the keypad number keys without dust to guess the code to enter the server rooms.

On Saturdays in 1979, my father would get some extra work done while I wandered the cubicle farms grabbing coins out of the coffee donation cup and feeding the vending machines to feed my sweet tooth. I would amble on over to the Apple II computer and load up some new games to play. I am not sure whether it was L.L. Bean’s or the other programmers, but there were plenty of games on 5.25” floppy disks to play with. There was Olympic Decathlon featuring Bruce Jenner, Lemonade Stand, and the cult classic that still exists today, Wizardry. Of course they had Wizardry, as I remember going to Dungeons & Dragons parties at my father's boss's house where many of the programmers would gather on a Saturday night to roll the dice or occasionally play poker.

The Maine Tech Frontier

Fast forward to 1988: I took typing class as an elective for the heck of it, not knowing it would be one of the most important skills I learned. I played around with programming low-res graphics on the Apple and enjoyed popping off the cover to see where the cords connected and which slots the floppy drive and video cards plugged into. My dad ran an Apple distribution store out of Portland, Maine, called Harper Computer, which eventually became Coastal Computers. Maine was actually a national leader in putting computers in classrooms, and MacSchool was the backbone of that movement. It was one of the first school administration software packages. When my dad left Harpers, he took over the MacSchool distribution and for years supported 250 schools through the state.

We used to go to the MacWorld Expos in Boston where I was allowed to find the best game to take home. My father would always buy one share of Apple stock, which he still has in his safe deposit box today. We didn’t see Steve Jobs as John Sculley, the Pepsi executive, had taken over in 1985. Little did I know I would find a mentor that was an executive at Apple during those formidable years. From 1988-1990, I was the accounting, word processing, and database expert for a pre-owned boat sales company. When I wasn’t repairing engines, pressure washing, or trailering the boats, I was inside on the Macintosh creating the inventory lists and descriptions for the boats for sale while storing everything on a 20MB external hard drive, the Apple Hard Disk 20SC.

From Architecture to Mega Yachts

I graduated in 1989 from high school and moved to Colorado in 1990 where I was running ski shops in Denver and Vail. I had taken an architecture class in high school, and when ski season was over, I was lost and not sure about my future. My roommate introduced me to a geeky friend of his who mentioned that Denver Institute of Technology was offering associate degrees in Architecture with a focus on AutoCAD. That started my dive into the PC world working with AutoCAD V10 on a DOS computer. Our instructor was a renowned architect and loved to leave the classroom and turn off the breaker for all of the computers to see if we were all remembering to save our work on a regular basis. Since I was one of his star students, I ended up working on blueprints for the homes he was designing.

That degree led to my dive into every aspect of AutoCAD drafting imaginable, including CAD on a Unix Sparc system. I ended up working on structural, electrical, mechanical, piping, and civil blueprints, including the Big Dig tunnel project in Boston for a few months. I had always been addicted to learning, and I sucked in everything I could process about each of those engineering focuses. I ran the tech support department at a large company called Inacom, and eventually moved to Florida to work on the private mega yachts, TV show Below Deck style, where my computer skills were quickly taken advantage of. I was the token nerd for the captain to come to, so I was always placed in charge of maintaining the PCs and keeping GPS databases updated. While playing Doom and Warcraft with my roommates in Maine, I had learned how to set up simple coaxial cable networks and these skills translated to my running Cat5 cables to link the engineering and wheelhouse computers.

When a new hire boarded these 150 foot+ yachts, they always had a hard time with our maintenance and cleaning supply locations. The chef wasn’t sure when the stewardesses had purchased the latest produce and meat in the huge walk-in freezers, and the chief engineer was daunted with maintenance logs for twin 800hp diesel engines, multiple generators, tenders and jet-skis. I had an "a-ha" moment and approached a friend who had programmed an accounting software that could handle different currencies as the yachts were always in multiple countries. I told him about my idea for a software system that would include his software package, but also had inventory, maintenance logs, owner/guest requirements, crew management and other features. I was able to get a $10,000 investment from one of the yacht captains to start building the software, but since my friend/partner had the programming experience, it seemed best that he work on it while I started creating the database of supplies, movies and engine parts for the yacht.

That was my first big lesson in entrepreneurship. My partner was too busy to work on the software and when the captain requested to start using it, my partner was flown over to the Feadship yacht construction yard in the Netherlands to work on the software exclusively while I was left behind. I learned that you always need a contract with your partner when you start a business, as the business I helped create is still running today and is one of the foremost in the yachting industry. I had lost out on the opportunity and went to work for MicroAge providing in-company tech support for the Clerk of Courts of Palm Beach and Office Depot headquarters in Boca Raton. I was one of the only contractors they trusted to be allowed alone in the Office Depot server rooms to move the computers from rack to rack while they were building their new server building close to the fiber optic internet hub in Atlanta. I received my Microsoft Certified System Engineer certification in 1999 based on the Windows NT server system.

Predicting the Future

After that, I was working for some architects, helping my musician friend with mp3s and webcams, playing with my new high speed DSL connection, and writing a business plan for an online used book website. I had another "a-ha" moment and set out to create a business plan for a company that would focus on providing customers a website that allowed them to find local ‘pros’ or contractors for their needs. It was a novel concept, as my business plan outlined a future many years ahead of its time where everyone would have a PC and an internet connection, they would pay their utility bills online, see their bank accounts on a website, and I also spoke about the future of the IoT tech where refrigerators would have LCD displays with scanners for inventory controls and you’d be able to set up schedules to start your coffeemaker and lights from a software system. I still have that business plan with a rusty staple and I found the ability to see into the future of technology.

In 2000, the yachts were installing new plasma televisions with Crestron control for the video and audio on these ugly touchscreens. My mom was living in Tiburon and I decided to move to California as I had the computer and technology skillset, Silicon Valley was nearby, and I wanted to pursue an interest in home technology systems. Jumping back a bit, I grew up as an only child in a religious home. My grandfather and father were Baptist ministers in Maine and my father attended BJU in South Carolina where he received his degree in teaching and holding hands with a lady was not allowed on this religious campus. I was super smart, taking apart the lawnmower and putting it back together, and installing windows into the garage by myself at the age of 12. My standardized test scores were always extremely high, but my religious mother and father believed that everyone was equal, so I was not considered a ‘special’ kid and was never guided in how to explore my capabilities or how to have confidence in myself.

The Silicon Valley Validation

My home automation career in California took me to various companies in various roles, but I was always the computer expert in charge of the servers, printers and networks when not in the field installing control systems in the uber-mansions in Marin and Sonoma county. Again, I sucked up everything I could and jumped from job to job as I either got bored or the business owner couldn’t see my capabilities and my career would stagnate with them. In 2009, I created my own home technology company called Technical Visionaries so that I could work for myself, as my father had done for many years before his retirement. In 2013, I designed, engineered, installed and programmed one of the largest home tech systems I had built for a venture capitalist based out of San Francisco. I mentioned this homeowner as my mentor earlier in my writing. After the first home I worked on was lost in the 2017 fires in Santa Rosa and Sonoma, and his DocuSign investment had just gone to an IPO, we set out to design an amazing audio, video and control network that would be extremely easy to use. He didn’t want any Crestron or control system as the goal was to have the TVs and Sonos surround speakers controlled with just the TV remote and all of the other systems on the property would be run through their individual apps.

Over the course of the next year in that Sonoma county home, I pulled miles of cable, installed TVs, and lots of speakers. I installed 5 network racks, 15 network switches, many Sonos amps, Lutron lighting keypads, and Nest thermostats. He would have an idea and I would make it happen, including the app that would turn on the property water sprinklers from his home in San Francisco should he see a possible fire approaching on the many cameras on the property. I built a system for opening the deadbolt and the barn door for their farm animals so they didn’t have to walk out first thing in the morning or to close the doors when the donkeys came in for the night. It was all done on iPhones and iPads. I was due for a vacation in Maine and the furniture was arriving and the time had come for me to start installing the TVs, network plates, and programming the lighting system. I had less than 9 days, so I ended up working from 10am to 3am every day to ensure that they could enjoy their new home before I left for Maine. It was exhausting and it was during this time that I received a text from him saying “You are the best”. I had never had much self confidence or faith in myself and receiving that text from an executive that worked with thousands of tech employees at Apple and Cisco in the most prosperous of years made me realize that I was really good at what I did.

The Big Cat Race and SASI

My son was born in 2010 and in 2011 and while pushing him in the swing in the backyard, I realized that I wanted a way for my son to enjoy gaming while also being educated, getting exercise, and being outdoors. I wrote the business plan for GenZPlay where over the next few years I predicted that wristwatches would unfold to show a large computer screen. There would be sensors inside of every piece of sports equipment, including baseballs, basketballs, and more. There would be interactive fitness games with sensors to tap on the walls and floor. I also had designed a game that came out shortly later that was very similar to Pokémon GO. My main focus was The Big Cat Race, where children ran with a wearable movement tracker or smartphone and then saw themselves on the smartphone screen as a cartoon Lion, Tiger, or Cheetah that was racing against other big cats (or other children in a multiplayer environment). The kids loved the simple beta version during testing, and I networked in order to find an angel investor so that I could create the full version of the game with levels, parental-set rewards and educational stories, all revolving around outdoor adventure and fitness.

During this phase, I also built one of the world's first apps for a professional sports event called the “Rahlves Banzai Tour”, where Daron Rahlves, a renowned X-Games and Olympic skier, was setting up extremely challenging ski courses that competitors had to navigate at top speed. Similar to skier or boarder-cross but over cliffs, moguls and off-piste terrain. The app allowed spectators and competitors to see the 3-day schedule, know where to check in, see the race results quickly, and learn more about the race, Daron, and the sponsors. It was a hit and everybody was using it, and Daron allowed me to promote GenZPlay at the races. GenZPlay never took off. The full software wasn’t built, I never found an angel, and the crowdfunding campaign tanked. I was ahead of my time and eventually most every idea I had thought of came to fruition over the next decade. It was humbling but rewarding to see that I still had my visionary capabilities, but it was another startup failure.

Today

I have created and fully built SASI, a safety and governance middleware for AI apps and systems. It started as a project to create AI apps that were safe spaces for children and adults where boundaries were enforced and AI replies were conscientious. It eventually turned into a middleware when I realized, like Norton Antivirus and Stripe, a middleware or third-party layer was going to be required in order to ensure consumer safety, audit-ability, and insurability. The last 5 months has brought up many memories of being ahead of the curve, but I’ve stuck it out and found the best way to market is to contact enterprise medical technology executives and cybersecurity and E&O underwriters.

The ed-tech and mental health AI app developers could not see the need for a middleware and were trying to safeguard their apps with system prompts and MCPs. In November, the FDA mentioned that “Given that premarket enforcement has been in effect since 2023, the industry has been bracing itself for the post-market cybersecurity requirements. For example, safety testing company UL Solutions has a page dedicated on its website to answering FAQs on how best to navigate Section 524B, where companies treat security as a core engineering requirement from day one, opposed to a documentation afterthought” The concept of middleware is now being seen and we are currently one of the very few software packages that is designed for high-risk AI systems with more oversight, audit-ability, and control than IBM, Microsoft, or Guardrails systems.

The transition from experimental AI to regulated industrialization has exposed a massive liability gap. High-liability companies don't just need probabilistic filters; they need forensic-grade accountability. This is exactly what I have built. SASI (Symbolic AI Safety Intelligence) acts as the 'black box recorder' for high-risk AI, providing deterministic, tamper-evident proof of process for every decision.

Our momentum is already building:

We have a fully functional sandbox ready for demonstration, a B2B marketing plan in motion, and we are currently enrolled in SOC2 Phase 1 and pursuing ISO 42001 certification. We have also submitted our Statement of Interest to the FDA TEMPO pilot program and are coordinating a formal study with UMass to validate the system. Furthermore, we are disrupting the industry’s broken 'Token Tax' model by offering flat, predictable billing for governed decisions, a feature designed in collaboration with a cybersecurity underwriting consultant to ensure we meet the most rigorous audit standards.

My history has been defined by seeing these technological shifts long before they become the status quo. Today, the shift toward AI safety and insurability is no longer a prediction; it is an immediate requirement for enterprise adoption.

We are currently seeking a $500k angel round to scale our infrastructure and meet the demand of high-risk AI deployments. I am looking for partners who understand that in the gold rush of AI, the most enduring value lies in the safety and governance layers that make the industry sustainable.

I would love to walk you through the SASI architecture and discuss how we are positioning ourselves for the next decade of AI integration. Would you be open to a brief call in the coming weeks?