OpenAI and Anthropic Models Claim Refutation of the Exponential Time Hypothesis

Skynet Chance (+0.06%): Models producing 'alien-looking' novel mathematics that experts struggle to interpret points to a growing gap between AI reasoning and human oversight. Such algorithmic advances also carry potential cryptographic implications that could undermine security infrastructure.

Skynet Date (-1 days): Frontier systems now resolving decades-old foundational problems signals that superhuman research capability is arriving faster than expected. Faster AI-driven research also compresses the time available to develop safeguards.

AGI Progress (+0.06%): Independently refuting a central, long-standing conjecture in computer science is strong evidence of creative, frontier-level reasoning, a key hallmark of general intelligence. Two separate labs reaching the result suggests the capability is broad across the frontier rather than a one-off.

AGI Date (-1 days): Rapid succession of proof releases and competitive replication indicates that AI-driven scientific discovery is accelerating. Such discoveries could feed back into faster algorithms and research, pulling AGI timelines forward.