P(doom)

Several AI industry leaders, including Anthropic's Dario Amodei and Elon Musk, have recently advocated for "pacing the frontier" in AI development due to concerns about catastrophic risks, with some estimating a 10-25% probability of AI causing significant harm. The author agrees with the concerns about potential misuse—such as AI-enabled cyberattacks and botnets—but is skeptical that meaningful oversight is possible, noting that only two companies (Anthropic and OpenAI) currently matter in this space and questioning what "pacing" would actually entail.