-4Measuring how often an AI research agent fools itself: an open protocol (seeking feedback and funding)Igor·1w ago·2m readIgor·1w ago·2m read