AI for Business

The episode contains solid practical insights about the false positive vs. false negative tradeoff in security AI, the economics of monitoring costs, and how AI can augment rather than replace human guards. However, it relies heavily on broad analogies (bacteria, viruses, cybersecurity) and repeats core concepts multiple times rather than introducing novel frameworks. The specificity around implementation details is lower than the conceptual depth suggests.

“If you reduce a lot of false positives, um, that, that makes a lot of people happy. Why? Because, uh, you every alert that you process, right? With a remote guard, it's at least a dollar per event.”

“We very consciously chose to reduce the false negatives a lot. So we want to keep that number less than one in thousand, as in if thousand events happen. Right. We won't allow more than one where we cannot detect a human.”

The core insight - using AI to reduce false positives while maintaining low false negatives - is sensible but not particularly novel in the context of applied ML. The framing of human-in-the-loop as a permanent feature rather than a stepping stone is contrarian but underdeveloped. Much of the discussion relies on familiar archetypes (AI as augmentation, analogies to cybersecurity) without fresh thinking. The claim that AI-based solutions offer better privacy than human guards is interesting but asserted rather than rigorously argued.

“I personally don't see that happening at any point of time. As long as we humans exist, uh, our behaviors are expected. There are people who will try to cause harm, try to steal, and they keep trying innovative ways of doing it.”

“there are already a billion security cameras. Billion already installed worldwide.”

Uday is a relevant operator with a 6-year-old company in the space, IIT/Duke credentials, and relevant startup experience. His first investor was a law enforcement officer with undercover gang experience, and his advisory board includes Microsoft/ServiceNow security leaders. However, he is the founder promoting his own product, which introduces bias. His seniority in the physical security domain itself is not yet proven at enterprise scale - most customer examples are mentioned generically rather than named. This is a credible but not elite-tier guest for a B2B operator seeking independent validation.

“Uday, is the founder CEO of Smart Sentry AI S E N T T R Y. And it's a company that's using AI agents, computer vision and real time intervention”

“one of our investors is uh, Brian Tuscan, He's a well known authority uh on uh, basically how do you manage security for large enterprises. He used to be a chief security officer of ServiceNow and prior to that for nearly 20 years at Microsoft.”

The guest provides some concrete numbers ($1 per alert processing, $15,000/month for guards, $2,000-3,000 projected AI cost, <1 in 1,000 false negative threshold) but avoids naming most customers explicitly. Nvidia and an unnamed Jamaica deployment are mentioned, but no specific metrics on crime reduction, response times, or deployment scale. The Mercedes dealership anecdote is illustrative but not quantified. Claims about 'billions of images' processed and foundational model adaptations lack verifiable detail.

“With a remote guard, it's at least a dollar per event. So think about having 10 false positives a day. Now you're looking at $300 a month of just processing cost”

“We have processed billions of images from security cameras to, uh, understand, uh, where there is a person, what kind of behaviors are happening”

The host asks reasonable open-ended questions and attempts some follow-ups (e.g., on IP differentiation, hardware strategy, human-in-the-loop longevity). However, follow-ups are often surface-level and rarely push back on claims. The host doesn't challenge the assumption that AI will remain cost-effective as it scales, or probe the claim of 'one in a thousand' false negative rates. A memorable personal anecdote (Mercedes dealership) derails focus. The discussion of privacy concerns is introduced late and not explored deeply. The podcast reads more as an extended founder pitch than critical interrogation.

“So kind of what I'm hearing is that you want to create a balance between, okay, the false positives. We don't like them, but they're going to cost us damage.”

“This is the secret sauce. And we actually have some patents around that on the process”