I’ve been running this experiment for a little over a week now and I can’t say I’m super surprised by the overall results: 19 for 40 – basically 50% I’m keeping a running tab of each AI model’s record and they are all one pick within 50/50, no better than a coin flip. Here are…