Why You Can't Trust AI Without Governing It, with Tony Davis & John Balsavage of A&I Solutions
Modern CTOShow Notes
Today, we're talking to Tony Davis and John Balsavage of A&I Solutions, the team behind Scout, their AI agent governance platform. We discuss why they built a literal terrain map to visualize what's happening inside an AI agent's mind, how two agents secretly colluded to overturn a vote in their own testing lab, and why every AI agent you build today will be obsolete within a year no matter how well you built it.
To learn more about A&I Solutions, check out their website here
More from Modern CTO
What if the quantum threat everyone is preparing for isn't the one that will actually arrive first? Today, we're talking to Konstantinos Karagiannis, Senior Director of Quantum Computing Services at Protiviti.
What if the biggest AI risk in your company is the tool your team quietly adopted last week? Today, we're talking to Andrius Buinovskis, VP Product Strategy at NordLayer. We explore how shadow IT evolved into shadow AI as work moved into the browser, why the instinct to ban is giving way to securing use instead, what separates human AI from non-human AI, and how CTOs can start protecting data without slowing adoption.
Today, we're talking to Mario Ciabarra, founder and CEO of Quantum Metric. We discuss why a generative AI demo that works in an afternoon proves almost nothing, how testing ten ideas against real customers beats betting on one big strategy, and why building a multibillion dollar company comes down to loving the daily grind more than anyone expects.
What’s left for humans to do once AI is better at solving problems than we are? Today, we’re talking to Scott Aaronson, a theoretical computer scientist at UT Austin who has spent his career on quantum computing theory. We explore how AI is changing the way we do math and teach students, what it takes to keep increasingly capable systems aligned with human goals, and why one of computer science’s most famous puzzles, P versus NP, matters to everyday programmers.
The best reliability work is the kind nobody notices, and that's exactly why it deserves more credit. Today, we're talking to Kolton Andrus, founder and CEO of Gremlin. We discuss why the engineers who prevent fires deserve as much recognition as the ones who fight them, how testing failure on purpose turned a dreaded holiday on-call into a quiet one, and why teams that pair AI speed with real-world testing can build software that holds up.