Saturday morning arrived wrappedin fog so thick you couldn’t see the valley. At breakfast, the dining room buzzed with typical conference energy—tech executives comparing notes over coffee, venture partners exchanging cards, engineers debatingthe latest frameworks. Morrison, Liu, and Castellano worked the room separately, having casual conversations that seemed perfectly innocent.
Liu’s session started right on time in a conference room with about twenty attendees armed with notebooks and tablets.
“Today, we explore AI applications in complex signal processing,” he announced, his presentation slides professionally bland. “Pattern recognition in noisy environments.”
What followed was a masterclass in walking the line.
Liu discussed oceanographic data analysis and acoustic behavior through various mediums—all publicly available science. But his examples kept nudging toward military applications. He explained sound patterns in underwater environments using textbook equations, yet chose specific depth ranges and temperature gradients that anyone with submarine warfare knowledge would recognize as optimal for detection avoidance. He discussed radar signal processing using civilian airport examples, while his “blind spot” discussions would resonate with anyone familiar with defense installations.
“Consider this theoretical problem,” he said, sketching a detection grid. “If you needed to identify objects in a complex acoustic environment with these specific conditions…” He listed parameters that seemed random individually but would spark recognition in anyone who’d worked with certain systems.
The genius was in the ambiguity. Everything he said could be defended as an academic discussion. But for those with clearances—and I counted at least six in the room—the implications were unmistakable. He was suggesting research directions, highlighting which problems were interesting, guiding where curious minds might explore.
People took notes. Asked follow-up questions about theoretical scenarios that were anything but. Liu answered everything with plausible deniability built in.
“Of course,” he added after explaining optimal frequencies for avoiding certain detection methods, “these are just mathematical models. Real-world applications would require specific knowledge I certainly don’t possess.”
A few people exchanged glances—some confused, others comprehending.
During the break,Brenna and I stepped onto a balcony overlooking the gardens.
“He’s teaching them how to think about classified systems,” I said quietly. “Without technically revealing anything classified.”
“Giving them the road map without drawing the actual map.”
“Everyone in that room knows exactly what he’s really saying.”
“But proving it in court would be nearly impossible.”
Lunch wasin the main dining room, with Morrison holding court. He told stories about Silicon Valley’s evolution, each with subtle undertones about outdated regulations stifling innovation. He never said anything explicitly illegal, but his message was clear: the real innovators found creative ways around obstacles.
“The best entrepreneurs I know,” Morrison said, “understand that competitive advantage comes from seeing opportunities others miss. Sometimes that means international partnerships others are too cautious to pursue.”
Castellano’s afternoon session focused on “Global Investment Strategies in Emerging Tech Markets.”
His presentation covered legitimate international finance, but with interesting emphases. He explained how consulting agreements could be structured for maximum flexibility. How international partnerships could benefit from certain payment structures. How to evaluate whether foreign partners were “reliable” for sensitive collaborations.
“Due diligence is critical,” he explained, showing a slide about vetting international partners. “You need to know who you’re really working with. Their backgrounds, their other partnerships, their long-term interests.”
It sounded like standard business advice. It was also a primer on operational security.
“What about export controls?” someone asked.
Castellano smiled. “There are legal ways to structure partnerships that comply with regulations while still enabling meaningful collaboration. The key is understanding what’s truly restricted versus what people assume is restricted. Often, the perception of limitation is worse than the reality.”
Saturday evening broughtcocktails on the terrace. The fog had lifted, revealing the valley in golden light. The cocktail hour had a different energy than Friday night. People clustered in small groups, having intense but quiet conversations. Morrison, Liu, and Castellano held separate courts, each surrounded by interested parties asking careful questions.
I watched Morrison work. He had a talent for reading people—identifying who was genuinely interested versus who was just curious. The interested ones got his full attention. The curious got charming deflection.
“Atticus, Bronwyn,” Morrison appeared beside us with fresh drinks. “Enjoying the summit?”
“It’s been enlightening,” Brenna replied. “Liu’s session hit close to home. My husband’s security work deals with similar challenges—finding patterns in complex data.”
“Ah, then you understand the complexities.” Morrison studied us. “The real innovations are happening at the intersections—where civilian technology meets, shall we say, more specialized applications. The key is finding the right partners who share your vision for what’s possible.”
“And how does one find these partners?” I asked.
Morrison smiled. “They usually find you if you’re working on something interesting enough. But these conversations are better had in private. Perhaps we could schedule something after the conference ends? I have a few thoughts about the security sector that might interest you.”