Some praise for @alex and friends. Thanks to the new model (and more attentive observers?), I’m seeing fewer mistakes in Penstemon this spring. Especially in Colorado, with Penstemon virens, which has 15 “other species commonly misidentified as this species”.
Also, I was able to test it successfully with a Macrotera bee. I’m a little worried that it doesn’t know the difference between the bees and the globemallow plants they are often found on, but humans should know the difference in the “visually similar” menu.