Hi guys,
Just checked the COVID-19 updates page on PSA, and among the changes they are implementing, there was a sentence that I don’t remember seeing the last time I read it, which states the following:
“The third initiative is automation through robotics. PSA has made an investment in special machinery that will allow us to expedite the Receiving and Research of trading card submissions. We expect this technology to be installed this summer with real production benefits being seen before the end of the calendar year. PSA will share more details on our automation efforts as the project develops.”
What I like about it is that they actually state it will be implemented this summer. At the same time, they are also expanding their workforce, so despite the current backlog I am incredibly optimistic about this update as logging orders in the system and researching cards takes a significant percentage of each submission’s grading time.
Does anyone else share my optimism, or do you feel like this will not have that much of an impact?
I don’t collect graded cards but up until this year I was getting my masters degree in AI and machine learning so I’m curious mostly from an implemention perspective what they’re doing and how.
Probably something where it can read what the card is and input the info directly into the computer. Like scans a Shadowless Gyarados against others that have been graded before to make the process of ‘research’ and entering data quicker.
This is what I am thinking. Almost the Google image search type function that scans card against a data base and enter all the information for that card
Yeah that would make sense and serve to make their grading far more consistent! I’m glad to see they are at least attempting to remedy the concerns of inconsistency regarding their grading standards.
Yeah, that would be my assumption as well. Even if it shortens the process by just one week per order, that should already go a long way towards ‘unclogging’ their current queue.
I mean, we all tend to complain about waiting times, but all those backlogged orders represent pending revenue for PSA as they only charge us after grading. I’m no business expert, but it’s probably safe to say that having hundreds of thousands, if not millions of dollars in ‘dead money’ sitting there for 6+ months is likely to push them towards optimizing their systems and increasing efficiency where possible.
AI grading is much more of a software problem than a hardware problem. If this robotics machine includes generating image data of each card as it comes in though, the data it produces is the the first critical step of going down the path of machine grading.
@kaldoverde, this could mean a million things, although I guess scanning the front and back of a card and double checking across a database of pre-existing defined categories (to help speed up the process) is probably pretty easy. It’ll be cool if it could give a pre-grade of centering and surface based on the scan to help graders to expedite the process (like the AI says it should be 9, and your job is just to quickly confirm it), since centering at least should be pretty easy to define.
Although, if they reduce the amount of human error in the grading process, PSA 10s will probably be harder to come by I guess.
In all honesty, even if it is just automation in entering orders and identifying cards, it should already bring processing times down quite a bit.
I do agree that for centering purposes, computers will do it much faster and much better but regarding the cosmetic aspects of grading, I personally would prefer to have those reviewed by humans instead.
