The Tesoro AI Podcast

Frenos: How Local AI Models Are Powering the Future of Cybersecurity

Darius Gant Podcast Cybersecurity, tech opportunities

In the world of cybersecurity, speed and precision are everything. A single vulnerability can disrupt an entire city’s power grid or halt production in a refinery. Yet, traditional security models have struggled to keep up with the growing complexity of modern systems.

On the Tesoro AI Podcast, host Darius Gant spoke with Harry Thomas, founder of Frenos, an AI-native operational technology (OT) security platform. What followed was a deep look into how artificial intelligence is redefining the way critical infrastructure defends itself against attacks.

From Nursing to Hacking: The Unlikely Journey of a Founder

Before launching Frenos, Harry Thomas’s career took an unconventional path. He began as a nurse specializing in Alzheimer’s care before transitioning into cybersecurity. Burned out by long hours in healthcare, he returned to school to pursue his passion for computers.

“I started as a penetration tester for financial institutions in Boston,” Thomas explained. “That evolved into working for companies like Security Matters and later Dragos, where we used machine learning for anomaly detection.”

His time at AWS Security proved pivotal. There, he led research on user behavior analytics, using AI models to detect when hackers impersonated employees. The experience helped him see AI’s potential to shift cybersecurity from reactive defense to predictive strategy.

The Birth of Frenos: AI as the Ultimate Security Analyst

The idea behind Frenos emerged from a recurring problem in operational technology: the overwhelming flood of security alerts. “In OT, there’s too much noise. Teams don’t know what to prioritize,” Thomas said.

Frenos uses machine learning and large language models (LLMs) to transform this chaos into clarity. The system builds a digital twin of a company’s network and unleashes a reasoning agent—an AI that acts like a hacker—inside the simulation. The agent identifies weaknesses, chains vulnerabilities together, and predicts how an adversary might exploit them.

“We can simulate how a hacker could move through a network without ever touching the live environment,” Thomas explained. “It’s completely safe, but incredibly realistic.”

This approach allows Frenos to provide proactive security, showing organizations where attacks could happen before they do. Instead of relying on humans to sift through thousands of alerts, Frenos uses AI to surface the few that matter most.

Why Local AI Models Matter

While most AI startups rely on cloud-based LLMs, Frenos took a different route. The platform is fully on-premise, meaning all processing happens locally on customer hardware. This design ensures that no sensitive data ever leaves the organization.

“The Frenos platform isn’t a wrapper around OpenAI or Anthropic,” Thomas clarified. “We run local models with about eight billion parameters. They’re smaller, but faster and safer.”

Local models also address growing concerns about data privacy, compliance, and government access to cloud data. For industries that manage critical infrastructure or personal information—like utilities, healthcare, and public agencies—this local-first approach is becoming essential.

Running models locally comes with challenges, especially in cost and optimization. Frenos fine-tunes each model for specific cybersecurity tasks, ensuring high accuracy despite smaller model size. “If you put our local model up against a larger general-purpose model for our specific use case, ours will win every time,” Thomas said.

The Economics of AI Infrastructure

Building Frenos wasn’t cheap. Training its models cost hundreds of thousands of dollars in experimentation and compute resources. “Training an AI model is not for the faint of heart,” Thomas admitted. “It’s like watching paint dry while burning money.”

He emphasized that startups entering AI need to understand the financial implications. Unlike traditional software, AI systems incur ongoing expenses for API calls, GPU usage, and model retraining. “Every API call adds up fast,” he said. “We chose to build locally so that we could scale more cost-effectively.”

This approach helps clients avoid the sky-high costs of cloud inference, where a single API-heavy workflow can reach tens of thousands of dollars a month.

The Build-or-Buy Dilemma

Many companies approach Frenos thinking they can build a similar platform internally. According to Thomas, that’s easier said than done.

“If you go down the local model route, you’ll need developers who understand model context protocols, retrieval systems, and tool integration,” he explained. “Speed becomes the biggest issue. We can deliver results in under an hour, while an internal system might take days.”

Frenos achieves this speed by combining multiple neural networks—machine learning for rapid computation and LLMs for contextual reasoning. The result is a hybrid approach that balances accuracy and efficiency.

Finding the Right AI Talent

When asked about the kind of people it takes to build a company like Frenos, Thomas highlighted two core traits: adaptability and curiosity.

“I care more about how fast someone learns than what they already know,” he said. “If I give you five research papers, can you summarize them for me by the end of the week? That tells me more than your resume.”

Frenos separates its AI team into LLM specialists and machine learning engineers, recognizing the distinct skill sets each role requires. Ideal candidates have both technical proficiency and domain expertise in cybersecurity. “It’s easier to train a cybersecurity engineer in AI than the other way around,” Thomas noted.

Balancing Innovation and Regulation

As AI regulations tighten worldwide, Frenos’s on-premise architecture gives it a competitive edge. Clients retain full control over their data, avoiding compliance risks tied to external model providers.

“The platform doesn’t send information anywhere. Everything is processed locally,” Thomas said. “That’s how we help clients who handle sensitive data stay compliant and confident.”

He believes that as privacy laws evolve, local AI solutions will become the default for critical industries.

Looking Ahead: The Future Beyond Transformers

Thomas closed the interview with a bold prediction. “Within two years, I think transformer-based large language models will be obsolete. Someone will develop a new architecture that’s smaller, faster, and more efficient,” he said.

As global companies explore new frameworks for AI, Frenos’s innovation demonstrates that bigger isn’t always better. By combining local intelligence with domain-specific expertise, the company is charting a path toward smarter, safer, and more sustainable cybersecurity.

Conclusion

The story of Frenos illustrates a broader shift in AI development. The future isn’t just about building bigger models; it’s about building smarter systems that fit the realities of business, regulation, and security.

From healthcare to hacking to entrepreneurship, Harry Thomas’s journey mirrors the evolution of AI itself—cross-disciplinary, fast-moving, and focused on solving real-world problems.

Frenos shows that when AI meets purpose, it can transform not just technology, but the way industries protect the infrastructure we all depend on.

Breaking Barriers in Venture Capital: Austin Clements on Building Slauson & Co. and Funding the Future
Darius Gant Podcast

When Austin Clements co-founded Slauson & Co., he wasn’t just launching another venture capital firm. He was rewriting the narrative of who gets funded, who gets a seat at the table, and how inclusive wealth creation can reshape innovation in America.

On the Build Tech Stack Equity, Austin sat down with Darius Gant to discuss his path from South Los Angeles to becoming one of the most thoughtful emerging managers in venture capital. What followed was an honest, behind-the-scenes look at the business of funding innovation, the importance of timing, and how AI is reshaping the world founders now build in.

From Gardena to Growth: The Making of a Venture Capitalist

Austin’s story begins in South Los Angeles, where curiosity about technology met a desire to understand wealth. After studying at Morehouse College, he returned home to work in finance, learning what separates those who build wealth from those who maintain it. But technology was always calling.

In the late 1990s, long before the world saw the power of e-commerce, Austin built websites for small businesses. “I was explaining to people why they needed a website when most didn’t think they’d ever want one,” he recalled. The experience taught him the power of technology to expand opportunity and visibility for underserved entrepreneurs.

Eventually, Austin discovered venture capital while studying for his CFA. Drawn by the mix of finance and innovation, he immersed himself in the writings of investors like Fred Wilson, Brad Feld, and Mark Suster, who became informal mentors through their blogs. The more he read, the clearer his path became.

“I realized that venture capital was exactly where I belonged,” he said. “It combined the parts of investing I loved with the future-building side of technology.”

After earning his MBA from NYU, Clements joined TenOneTen Ventures, where he learned the craft of early-stage investing. Those years gave him the foundation for something bigger.

Founding Slauson & Co.: Optimism, Naivety, and the Power of Timing

Austin teamed up with longtime friend Ajay “AJ” Thomas, who had helped build Queensbridge Venture Partners with Nas. Both had made their way into venture capital independently, but they shared one vision: to create a firm that reflected the communities they came from and the founders they wanted to serve.

Launching a venture fund wasn’t easy. “Like most entrepreneurial journeys, it started with optimism and naivete,” Austin laughed. “We thought everyone would line up to give us money.”

Instead, fundraising proved brutal. After leaving his job, Austin had a year’s worth of savings to make it work. He spent it cold-calling over 300 potential limited partners (LPs), often hearing polite rejections. Just as the savings were running out, 2020 changed everything.

The dual shocks of the COVID-19 pandemic and the George Floyd protests reshaped the conversation around inclusion in finance. Suddenly, investors were looking for credible, diverse fund managers who could bring new perspectives. “We didn’t change our story, but the world finally started listening,” Austin said.

The firm’s first fund ballooned from an initial $15 million target to $75 million, backed in part by legendary investor Ron Conway. The lesson, Austin noted, was simple: “Timing is everything. Sometimes you just have to stay in the game long enough to get lucky.”

A New Kind of Investment Thesis

Slauson & Co. focuses on pre-seed and seed investments across both B2B and consumer technology. On the B2B side, the firm targets tools that empower the next generation of small business owners. On the consumer side, they invest in what Austin calls “architects of the human experience” founders who deeply understand cultural and behavioral shifts shaping society.

“We look for lived experience as a competitive advantage,” he explained. “That means we’re backing people who understand the world they’re building for because they’ve lived it.”

Slauson’s checks typically range from $1 to $3 million, often as the first institutional capital into a company. Austin looks for founders who can identify macro shifts and align them with unique insights a framework he credits to the book Pattern Breakers by Mike Maples Jr.

“Your idea might pivot,” he said, “but your insight should stay constant. That’s what separates great founders from the rest.”

The Power of Storytelling in the Startup Era

Austin believes one of the biggest shifts in entrepreneurship is how founders connect with their audiences. “A few years ago, everyone thought tech platforms should be neutral. Now, people want authenticity,” he said.

He encouraged founders to embrace vulnerability and purpose in their messaging. “Your story matters. People want to know who you are, what you stand for, and why you built what you built.”

From Tesla to ChatGPT, Austin noted that public perception of leaders now directly impacts business performance. “It’s not enough to say your product speaks for itself,” he said. “If you don’t get your story right, it’s hard to build a billion-dollar company.”

Investing in AI: Finding Opportunity Beyond the Hype

When the conversation turned to artificial intelligence, Austin was clear: while the LLM race may be over for new entrants, opportunity still exists at the application layer.

He pointed to Abby, an AI-powered therapy app in Slauson’s portfolio. The platform is designed specifically for mental health, addressing issues that general-purpose AI tools like ChatGPT cannot. “Abby is built with the guardrails, empathy, and research that a therapy-focused product needs,” he said.

Austin believes the next wave of AI innovation will come from specialized, purpose-built systems that solve real problems rather than chasing massive infrastructure models. “Find your lane and own it,” he advised founders. “You don’t have to compete with OpenAI to build something transformative.”

Redefining Access: The Friends & Family Accelerator

For founders without the privilege of early funding networks, Slauson & Co. launched the Friends & Family Accelerator. The program invests $300,000 into early-stage startups and provides six months of mentorship, exposure, and resources.

“Our goal is to support founders who might not have a friend or family member who can write a five-figure check,” Austin said. “We want to close that access gap.”

The accelerator helps participants refine their ideas, connect with seasoned founders, and prepare for seed-stage investment. “Sometimes all a founder needs is a little capital and a strong network to change their trajectory,” he added.

Looking Ahead: Where Venture and AI Collide

Austin sees the future of venture capital and AI converging around human insight. As automation scales, creativity, lived experience, and emotional intelligence will become defining factors in successful entrepreneurship.

“The founders who thrive in this era are those who combine technology with humanity,” he said. “We invest in people who not only see the shift coming but know how to shape it.”

Slauson & Co. isn’t just betting on ideas—it’s betting on the next generation of innovators who reflect the diversity, resilience, and creativity of the world they serve.

To hear her full, inspiring journey and learn more about her advice for non-technical founders, listen to the complete conversation on the Build Tech Stack Equity Podcast. Follow us for more stories of the incredible entrepreneurs and investors shaping the future of technology.