This episode asks whether a model described as dangerously capable is a real security problem or a marketing narrative. It uses the Astra sandbox incident as a starting point, then covers how agents are moving from chat into the browser to book, file, draw, and shop, and what that shift means for payments, advertising, and merchant data.
Emmanuel Ameisen, an interpretability researcher at Anthropic, explains how the team looks inside models like Claude to understand how they actually think. The conversation covers why next-token prediction hides real planning and conceptual representation, why models are grown rather than built, how intermediate activations reveal structure, why anthropomorphic framing can be useful, how smooth scaling produces sudden capability thresholds, and how interpretability ties into alignment, safety, and the broader question of what these systems are.
This episode of Paper Club addresses AI safety from a scientific perspective, contrasting it with recent media sensationalism. Stanford professors Clark Barrett and Mykel Kochenderfer discuss how automated reasoning can mathematically prove code security, how aircraft stress-testing methods are being adapted for language models, and the potential need for an NTSB-like agency for AI, highlighting the real-world engineering challenges and solutions for AI safety.
The panel examines the rise of decision models, a family of LLM-based systems that return a single, scored classification in one forward pass, and asks whether they are genuinely new or a smarter use of existing architectures. It then weighs the cost and benefit of multi-agent teams, drawing on a study that found agent teams cost two to five times more without consistently better results, and closes with Anthropic's new policy against sustained cruelty toward its models and what it says about AI safety, ethics, and human behavior.
MIT FutureTech director Neil Thompson revisits his group's 2020 computational limits paper and discusses his newer research on how AI automates tasks across the economy, covering the crashing-waves versus rising-tides survey, adoption bottlenecks, fixed-cost economics, apprenticeship risks, AI safety, and the open versus closed model debate.
Misha Laskin, co-founder and CEO of Reflection AI, discusses building frontier open-weight models, the launch of the 500B-parameter Beam reasoning model, the economics of renting versus owning intelligence, and why he believes open models will capture most token demand. The conversation covers RL-driven scaling efficiency, Chinese open model ecosystems and their geopolitical implications, and the safety case for openness, including cybersecurity as a lens for defense and offense. He closes with optimism about AI-accelerated science and the evolving research workforce.