Today we streamlined and fixed up the whole workflow when a new species is added to a sighting i.e. a species that doesn't exist in the database. We've blown away the old "Uncategorised species" category now and you can add a newly suggested species directly into the correct category now from the sighting page. Heaps better.
I'm surprised it didn't pick up Marbled Xenica here but instead went for Bright-eyed Brown, albeit with a low probability. Any thoughts @AaronClausen on this one or just all part of the learning process for our silicon friend?
Yep very interesting @MatthewFrawley - best defense here is to confirm it manually and now this exact image gets added into the model so you're nicely filling in it's effective gaps.