BASELINE

Do We Still Understand the Systems We Build? Tudor Girba

AI is accelerating software development at an unprecedented pace. But what if the real problem isn’t building software anymore?What if the problem is understanding it?In this episode of BASELINE, Tudor Girba joins Ian Smith to explore one of the most overlooked challenges in modern technology: the growing gap between the software we create and our ability to comprehend it.As AI generates more code, organisations are becoming increasingly dependent on systems that nobody fully understands. From legacy banking infrastructure and enterprise software to AI-assisted development and digital transformation, Tudor argues that the future of software engineering may depend less on writing code and more on learning how to read, understand and evolve complex systems.

You can no longer trust video calls - Will Taylor

AI can now clone voices, fake video calls and imitate real people convincingly enough to steal millions of dollars from businesses in real time.In BASELINE089, Ian Smith sits down with Will Taylor from YEO Messaging to explore the collapse of trusted communication in the AI era. As deepfakes, AI-generated identities and social engineering attacks become increasingly indistinguishable from reality, organisations are facing a new challenge: how do you prove the person on the other side of a message, voice note or video conference is actually real?

Human Engineers Are Reaching Their Limits - Elie Talj

Modern engineering systems are becoming increasingly software-defined.From vehicles and aerospace systems to robotics and critical infrastructure, software is now the bottleneck limiting how quickly complex systems can be built, tested and deployed.In BASELINE088, former McLaren Principal Engineer and Arrival Director Elie Talj explores why engineering may be approaching the limits of human scalability, and how new forms of AI could radically accelerate the creation of deterministic control software.The discussion explores the hidden complexity inside modern engineering systems, the tension between speed and reliability, and why current AI coding tools struggle in safety-critical environments.

The Actual Cost of AI Compute

AI promises abundance, automation and intelligence at planetary scale. But what problem is it actually solving?In BASELINE086, Richard Self joins BASELINE to explore the economics, infrastructure and physical reality behind artificial intelligence, large language models, generative AI and the next generation of compute systems.Drawing on 50 years inside modern computing, from Rolls-Royce mainframes to today’s AI infrastructure race, Richard examines why many AI benchmarks are becoming meaningless, the true economics of generative AI, the physical limits of compute and energy, and why SpaceX is rapidly positioning itself as an AI infrastructure company through orbital data centres and low Earth orbit compute systems.

$2B to $10B: Why Quantum Suddenly Matters

In this episode of BASELINE, Ian Smith sits down with Kirill Pyshkin, investor and physicist, to unpack what’s actually happening inside the quantum market.Investment has jumped from around $2 billion to over $10 billion in a year. Large institutions like HSBC and Moderna are already reporting real-world results. So why does it still feel like a science project?This conversation breaks down what changed in 2025 and why it matters, where quantum is already being used today, why one breakthrough can shift an entire industry, the real risks around encryption and security, how investors separate signal from hype, and what happens next as quantum moves from research to deployment.