Forecasting agent says Kalshi's 32% on P vs NP by 2030 is sentiment, not signal
rohanpaul_ai · x · 2026-10-28
A forecasting agent scanned Kalshi's 'Will P vs NP be resolved before 2030?' market (priced 32%) and output 18-19% with 44% confidence, arguing the jump from 12% after the ETH-refutation news reflects sentiment, not P-vs-NP-specific information. Clay's page still says Unsolved; no peer-reviewed claim exists; experts like Fortnow say the unreviewed ETH refutation (exp(n/sqrt(log n))) doesn't bear on P vs NP. The agent keeps a fat tail given AI labs' pace — OpenAI and Anthropic independently released same-day ETH refutations, and Williams urged more weight on P=NP — imagining a Lean-verified polynomial SAT algorithm surfacing within days. Related Polymarket odds: 1.3% for AI solving a Millennium problem by 2028.
More from coding & agent
- FrontierCode 2: Anthropic models burn 2-3x tokens, cost 10-50x more than open rivals — bookwormengr · 2026-10-28
- Opus 5.5 ultracode workflow with 9 subagents builds a Three.js scene — nptacek · 2026-10-28
- Agent sandboxes talk at PLATFORMday SF: the missing runtime for AI — gethackteam · 2026-10-28
- Claude + FullEnrich MCP: 30-second setup maps Linear's entire org with 20 verified emails — socialwithaayan · 2026-10-28
- FullEnrich results: 20 manager-level contacts at Linear, 20 work emails, 20 credits — socialwithaayan · 2026-10-28
- Test setup: same prompt on Linear, Claude alone vs. Claude with FullEnrich — socialwithaayan · 2026-10-28