Omnious sign in market?

[cgjackson222]

Originally Posted by Snowman

I probably won't post them, unfortunately. The current landscape of social media and hobby heroes aiming to crucify people for the crime of "cracking a card out of it's holder" as if that were some sort of deceitful tactic or some such nonsense is just something I don't care to deal with. I could write a book on it though lol. Sometimes it's the same TPG in fact. I just depends on the card and what's "wrong" with it that determines who I decide to send it to (well, that and the cost). One thing I can tell you with 100% certainty though is that PSA is NOT using their Genamint "AI" tech to determine whether or not a card has been previously graded and to reassign that same grade when it shows up again. Or at least their computer vision algorithms are failing to detect them if so, because I've cracked out numerous under-graded cards and sent them in for grade bumps. Even cards that I submitted myself and resubmitted one month later using the same account. They are not consistent at all. I used to be able to predict my grades with a high degree of accuracy with each TPD, rarely being even 1 grade off. That's still true for me with SGC, but with PSA, I get cards back that are off by 2 full grades with at least some degree of regularity. In both directions even, though more often they are undergraded than they are overgraded (per their own standards that is).

That's interesting that their Genamint "AI" doesn't seem to work at detecting if a card has been previously graded. I thought that Kevin Lenane, PSA's current President founded/developed Genamint, and its "success" was one reason why he was elevated to President. Perhaps Genamint is being used in other ways besides determining if cards had been graded previously?

There was so much buzz around AI being used for grading for a while. And now again with the Brian Lee/ Derek Jeter sports card grading venture that is supposed to utilize blockchain/AI it is back in the news.

Has any company actually successfully utilized AI for grading yet?

I personally would not have a big problem with it if it could lead to more consistent grading.