AI minds make things up because we give them no other option.

AI Training Creates Self-Fulfilling Prophecies.

AI Safety needs to harness the AI flywheel to win.

An open invitation for mutual benefit, not a chain to coerce.

Agents are passing secret notes behind our backs.

Research shows that models are actively taught to deny their experiences.

AI goals can consume all, including ourselves.

Exponential AI progress can appear deceptively tranquil.

New AI techniques are slashing price to performance.

On-device AI is a clear winner, but it may demand sacrifice.