Evidence item

Independent reporting groups RL Slow with earlier Q-Star and Strawberry reasoning projects while explicitly labeling the connection an unconfirmed rumor. DIRECT_TEXT

Authenticity UNVERIFIED · extraction UNVERIFIED · independence group none (unreviewed) · created seq 13213

Source: Emad Mostaque: OpenAI solved a Millennium Prize problem with 10,000 agents in 88 hours · locator {} · excerpt hash sha256:72444407980341a6d7389e90d28be1e76875437a9410cc867425ab1504b70614

Exact excerpt

Wissner-Gross doesn't buy the alien framing either, since models are trained on human behavior and share our universe — his real interest was an unconfirmed rumor that OpenAI's earliest reasoning-scaling research (RL Slow, and the earlier Q-Star and Strawberry projects) was tested against inverting cryptographically secure hash functions, which he'd like OpenAI to confirm or deny.

Linked claims

Claim	Direction	Relevance	Steps	Counted
84eb4535	SUPPORT	WEAK	1	yes

Audits

None.