You Can’t Move Fast and Break the Grid
What happens when an AI-first mindset meets critical national infrastructure?
David is joined by Berian James, a machine learning and AI product leader whose career runs from payments at Square to supply chain transformation at Maersk, energy trading, and now Tem, where he’s helping rebuild how UK businesses and generators buy and sell energy.
Berian makes the case that “move fast and break things” simply doesn’t apply when the energy grid is at stake. He explains why risk management, done well, is an engine for growth rather than a cost to be controlled, and how pricing fixed contracts three years out depends on deep regulatory and domain knowledge, not just clever models. David presses him on a question every buyer of AI should be asking: if your provider runs on agents, how can you trust today’s price in two and a half years?
They also tackle the jobs question head-on. With graduate roles down by a third since ChatGPT launched, Berian argues that AI-native graduates may actually have a head start, and that today’s shift in engineering is the next leap in abstraction, much like the move from assembly to compilers.
