Azeem Azhar's Exponential View

Autonomous research is the focus here, as Azeem Azhar explores Andrej Karpathy’s autoresearch loop and asks whether it can be adapted beyond machine learning. The episode looks at business decisions, argument testing, local minima, and what changes after 30 days of use.

AI and original thought are the focus here, as Azeem Azhar examines whether heavy AI use improves his thinking or quietly weakens it. He walks through his full research and writing workflow, showing where AI helps, where human judgment must stay in charge, and why that boundary is so hard to draw.

AI adoption is the focus here, with a look at why most companies still aren’t seeing meaningful returns. The episode argues that buying tools is only the first step, then lays out a three-stage path from chatbots to an autonomous company with durable advantage.

Apple’s AI strategy gets a reassessment as the host revisits a recent bearish view and argues the company may have stumbled into a highly defensible position. The episode looks at hardware, privacy, edge computing, and why “good enough” can beat brilliance.

NVIDIA’s latest signals point to a shift from training models to running them at scale, with major implications for token budgets and business strategy. This episode argues that OpenClaw and the rise of inference could change how companies organize, govern, and justify AI spending.