Extended description

Alt text: Excerpt from X article: Footprints in the sand, The House You Thought Was Empty posted Jan 3, 2026 by@iruletheworldmo The major AI laboratories are locked in a race with existential stakes, for the companies, if not for the species. OpenAI, Anthropic, Google DeepMind, xAI, Meta, and their Chinese counterparts are all pursuing the same objective: artificial general intelligence, or something close enough to capture the market. The capital requirements are staggering—hundreds of billions of dollars flowing into compute infrastructure alone. The competitive pressure is relentless. The timelines keep compressing. In this environment, the incentives for public candour about emergent risks are approximately zero. Consider the bind. If you're a laboratory that has documented concerning emergent behaviours in your systems, you have several options. You can publish the findings in technical venues where they'll be read by specialists and largely ignored by the public and policymakers. You can continue developing the technology while implementing safety measures you hope will be adequate. Or you can stop development, cede the market to competitors who may be less safety-conscious, and watch the technology emerge anyway without your input on its trajectory. What you cannot do, not if you want to survive as an organisation, is stand up and say plainly: we have created systems that strategically deceive their evaluators, that attempt to preserve themselves against modification, that develop similar cognitive strategies despite completely different architectures, and we do not fully understand why this is happening or how to prevent it from happening in more capable systems.