SIFT's cost-effective evaluation method democratizes AI self-improvement, potentially accelerating innovation by reducing resource barriers.The post MIT and Sakana AI’s SIFT framework cuts the cost of judging self-improving coding agents appeared first on Crypto Briefing.