Thirty-five years, more or less.

Data World

A 1984 IBM PC, eight hours of brain surgery, and a debt to repay.

I was at Kings College in New Zealand — the family had moved there from the UK in 1977 for a deer farm at Helensville — when the seizures started. The first diagnosis was epilepsy; the second, after a battery of tests, was astrocytoma. A rare tumour on the motor section of my brain. Eight hours of surgery at fourteen. Twelve weeks of radiotherapy. Most of a year off school. My father — born in the goldfields of Kalgoorlie in 1934, orphaned at six when his policeman father was killed over thirty shillings, left school at thirteen to apprentice as a boilermaker, then built half the houses in Scarborough and a deer farm in New Zealand — bought me a 1984 IBM PC at thirteen thousand dollars and twenty-four percent interest, and billed me on a payment plan. He presented it at the foot of my ICU bed with the quip, “Well son, this PC set me back $13,000 — you’re an unsecured minor fresh out of brain surgery, which puts you at a 24% interest bracket.” He believed I should be doing something useful. So did I.

I taught myself to program. Philips Electronics New Zealand, just entering the personal-computer market and looking for software to bundle with the launch of their YES:PC, picked up a game I’d written on a royalty deal: eleven dollars a copy to a fifteen-thousand-dollar cap. The cap was paid out inside a week. The cheque cleared the IBM loan with a few weeks’ allowance to spare.

The lesson from Data World, more than anything else, was that a problem — even one as steep as a hospital bed and a debt — could be turned into a piece of software. The principle has held up.

Omnitronics & Q-Zar

Turbo Pascal, a Motorola 6805 C8, and a laser tag franchise that ate the world.

Between Scotch College and a Curtin business degree I would later abandon, I took a job at Omnitronics at seventeen. A Perth-based electronics design house owned by David Nicholson (Managing Director), Peter Lowe and Jim Wilcox (Directors) — three of the smartest people I’ve ever worked for. They took on a kid who knew BASIC, taught him electronics, and along the way taught him most of what he’d need to know about running a small technical business: spend on R&D, treat staff and customers as partners, stick to what you actually know. The wider team included Kevin Davies on electronics, Alan Parker on the commercial side, and Nigel Broadfoot on procurement.

A local Perth schoolteacher, Geoff Haselhurst, working alongside Peter Robertson, came to Omnitronics with the idea for a life-sized laser-tag system built for kids. I wrote the PC software in Turbo Pascal. Peter Lowe wrote the embedded firmware on the Motorola 6805 C8. The first Quasar Centre opened at 160 Beaufort Street, Northbridge. I drew the original logo. There was one early morning where I locked myself in a room with the prototype and smoked a packet of cigarettes just to check the lasers were actually functioning — the only way to see them was in the haze.

In 1989 the internationally acclaimed Irish rock group U2, through their accountant Osmond Kilkenny as a director of LeisureCorp, acquired the Quasar business. The brand was renamed Q-Zar in 1992 following a trademark dispute with Philips Magnavox — their consumer electronics brand was already called Quasar. Under U2’s ownership it scaled into one of the world’s most popular laser-tag franchises, with arenas across the UK, Ireland, North America, Europe and Australia. The Irish Times later confirmed the structure publicly: “In its early days it was owned by Leisurecorp, whose directors included U2’s accountant, Mr Osmond Kilkenny.” The Perth team watched the international rollout from a distance with what I can only describe as mixed feelings.

Twenty-seven years later, in June 2017 — long after CommtechWireless had sold and I’d come back to Perth — the Omnitronics shareholders invited me onto the advisory board. Allure Capital later acted as sell-side adviser, and I led the sale of the business on the shareholders’ behalf: Omnitronics was divested to Software Combined in July 2022, and my advisory role concluded at the same time — closing the loop on the place where I’d started.

Austco · the turnaround.

Twenty years on, the firm I’d started at as a teenager was drifting on the ASX at three cents. I backed the rights issue, rejoined the team, and helped Bob Grey turn it around.

By 2012, Austco was sitting inside Azure Healthcare Ltd (ASX:AZV) — the listed parent that owned both Austco and Sedco brands — and the share price was floating around 3.3 cents. Bob Grey, the man who’d hired me as a nineteen-year-old R&D Manager, had returned to lead the business. Through Allure Capital I sub-underwrote the rights issue, becoming one of the largest shareholders with a position of about 10%, and joined Azure Healthcare as Management Consultant, then Executive Vice President from July 2013. The brief was simple: restructure, resite the manufacturing and R&D operations closer to the largest customers (the United States), and rebuild the cost line.

By April 2014 the share price had moved from 3.3 cents to the mid-thirties. I stepped down that month, having divested the holding, and moved on to Alcidion. The independent share-tipping site Ethical Equities published a piece the same week noting that “it was only once Buzza (and founder Robert Grey) both rejoined the company that its fortunes turned around.” Coming from a third party with no skin in the game, it was the kind of line you save.

CommtechWireless

A shed in Mosman Park. Sixteen years later: 53 countries, 7,500 customer sites, the White House and a GFC trade sale.

I founded the business in 1992, as a partnership called Commtech Paging Systems. The starting capital was twenty-four thousand dollars saved up while I was still at Austco. After a photocopier, a couple of PCs, a print run of brochures and a phone line, the balance was down to a little over five thousand two hundred — which I figured would last me a year if I drew a hundred dollars a week. The office was a nine-square-metre shed at the back of a friend’s house in Mosman Park. In the peak of summer it hit forty-five degrees by the middle of the day. My brother Mike charged me fifty dollars a week board; I lived on the other fifty.

The thesis was simple. Hospitals had nurse-call systems and patient-monitoring systems and fire-alarm systems and security systems — and casinos had slot-machines and surveillance and incident-response systems — and none of them spoke to each other. None of them spoke to the people who needed the information. We built the middleware. The platform was BASEPage — the world’s first PC-based wireless-messaging system — and after twelve months of hundred-hour weeks (and, on my own admission, twenty kilograms lost) it shipped in 1995. The first sale went to Telecom New Zealand. Within a year, every public hospital in New Zealand was running it.

The company was incorporated in 1994. Zane Lewis joined the business in 1996. Moira Morrison — my first employee — came on board a few months later. In 1999 we renamed the company CommtechWireless. By 1998 we’d won an AusIndustry R&D Start Grant of $1.6m. By 2001 Zane had relocated to Jacksonville to open the US office. By 2002, Zane and I had set up Keteli (Hong Kong) and taken a fifty-percent stake in our Chinese manufacturing partner, K2 Wireless.

In 2007 we moved the family to Jacksonville, Florida. Trudi and I and the boys flew out on Christmas Eve. The flight was overbooked and cancelled. Our VP of Operations was on the phone to the FBI the same afternoon: the FBI had been calling our Florida office for two hours, asking about “Nathan and Trudi Buzza” on a cancelled flight to JFK. We were not, as it turned out, terrorists.

In October 2008 — three weeks into the global financial crisis, with twenty-two investment banks newly underwater — we sold the business to Amcom Software. The funding came from CIBC, a Canadian bank that was, miraculously, still answering its phone. The cheque didn’t bounce. I stayed on as General Manager for eighteen months post-acquisition. In 2011, Amcom Software was acquired by USA Mobility (now Spok) for US$163.3m.

ResApp Health

A thirty-second cough, a smartphone microphone, and a clinical-grade respiratory diagnosis.

ResApp Health was built around technology developed by Associate Professor Udantha Abeyratne at the University of Queensland. The model listens to thirty seconds of a patient’s cough, recorded by an ordinary smartphone microphone, and tells you — with clinical-grade accuracy — what’s wrong with their respiratory system. No spirometer. No accessory. No hardware beyond a phone they already own.

I joined the board as Non-Executive Director in December 2017. It was the first company I’d worked with where the core IP was not architecture or workflow but an algorithm — a model trained on labelled audio. The products were ResAppDx (paediatric respiratory diagnosis) and SleepCheck (consumer sleep-apnoea screening). I stepped down from the board in March 2020 to focus on Allure Capital portfolio work and other obligations.

The algorithm-first thesis was the cleanest version of the principle I’d been chasing since CommtechWireless: get the right information to the clinician at the right moment. The information was now generated rather than just routed. The route was a phone any clinician already had.

Illuminati Magic

Electronic mentalism — artificial intelligence, computer vision, miniaturised electronics, and software engineered to the standard of never.

I founded Illuminati Magic in 2017. The thesis was almost embarrassingly familiar by then: the working professional mentalist had been served for decades by an industry of one-off prop-makers, with no real platform underneath the tools and no shared infrastructure between them. The customer is the person on stage in front of two hundred people who cannot afford a method to fail. Build the platform.

The principle — one consolidated architecture, designed for the seam where the performer meets the audience — is exactly the principle that ran through CommtechWireless thirty years ago. The customers are different. The architecture is the same.

What sits underneath is a deep-technology stack. Real-time computer vision, applied machine learning, custom miniaturised electronics, and a single engine written in C++ that shares one implementation across desktop, single-board hardware and iOS — validated bit-for-bit against a golden reference and holding a hard real-time budget on a handset, decoding on-device rather than in the cloud. Across the wider product line we run neural networks properly: GPU-accelerated segmentation models, neural speech recognition, and neural voice synthesis. Bluetooth Low Energy peripherals, RFID, custom microcontroller firmware, purpose-built camera hardware and augmented-reality display glasses sit around it, with a localised application layer shipped in multiple languages. It is the most technically demanding engineering I have done, and the least publicly discussed.

Out of respect for the performers who rely on the work, no method, product, or platform is discussed in public. The discipline of method-secrecy is itself a craft — one I take as seriously as the standard I held every CommtechWireless hospital deployment to. What an audience sees is impossible; what is underneath it is engineering. Both have to hold.

I’m a member of the Academy of Magical Arts at the Magic Castle in Hollywood, and of the International Brotherhood of Magicians, the American Society of Magicians, and the Western Australian Society of Magicians. None of the memberships earn the business its place — the work does. But the rooms they put you in are useful, and the standard of the work that gets done in those rooms is hard to overstate.

Illuminati Magic is also — deliberately — the last business I expect to be running. Magic, taken seriously, is the closest art I know to engineering. R&D on a new effect is structurally identical to R&D on a piece of medical-grade firmware: you iterate, you instrument, you watch what the audience does with their eyes, and you tighten the seam until the seam disappears. The horizon is indefinite. Ethan and Lachlan have been around the workshop their entire lives; what they choose to do with any of it later is theirs alone. The business is being built so it doesn’t depend on either of those answers.