AI experts and investors examine accelerating recursive self-improvement, election-forecasting agents that rival top humans, enterprise shifts away from open-source models, and the urgent need for better AI governance.