Why is PCGS priceguide so high on this 1822 large cent?

Discussion in 'US Coins Forum' started by lordmarcovan, Aug 16, 2025 at 10:19 AM.

  1. johnmilton

    johnmilton Well-Known Member

    AI is like any other computer program. If you put garbage in, you get garbage out. It is only bigger and more sophisticated.
     
  2. Avatar

    Guest User Guest



    to hide this ad.
  3. KBBPLL

    KBBPLL Well-Known Member

    I have no clue on these but it seems like certain scarce varieties might throw off the price guide and it's tough to toss a generic value on one. The finest known AU58 N-13 for example sold on Heritage in 2006 for $11,500. An N-6 AU58 sold in 2014 for $3818, and an N-4 for $2585. But those are the only three that have ever sold there for over $1700 so...

    I would add insidious to that list of adjectives. I tested ChatGPT again last night for "Canadian 1947 maple leaf dollar" and it came back with nonsense. I told it that it was nonsense and why, because on another forum someone said it's still being "trained", so I figured I would see what happened when I tried to train it. It was apologetic, told me I was right, but still returned wrong information. This ended after 4-5 attempts to train it, with "Let me try once again, this time drawing on the correct and factual details based on reliable numismatic knowledge:" and continued to insist that the maple leaf was added because of a "flaw in the 7" (total bunk). It's insidious because it acts so authoritative about nonsense, people trust it, and it will be manipulated in other areas for bad reasons.
     
    johnmilton and lordmarcovan like this.
  4. samclemens3991

    samclemens3991 Well-Known Member

    @KBBPLL . I commented on this is in a different post. My daughter has advanced degrees and ,as I relayed before, she attened a confernce that 3 of the creators of AI programs were at. All 3 said they made a major flaw when rushing to create a program for market.
    They told their programs that they should look for the "best possible answer" but then also instructed them that computers are flawless information sources and people are flawed emotional creatures. This means derivitive AI programs will take incorrect information from other AI programs as truthfull but discount human response.
    All 3 felt that within 5 years current programs would become useless because of accumulated wrong AI information or a corrective response was needed right now. All 3 also thought current programs would probably get junked and replaced. james
     
    Randy Abercrombie and KBBPLL like this.
  5. chrissy1955

    chrissy1955 Member

    Which begs the question; Is AI (ugh) going to be used to grade coins indiscriminately? Methinks this would be impossible due to AI is simply a whole lot of data piled up and then sorted for optimal use (I think). But where does that data come from................humans. So skip the middleman AI (ugh again) and just get an informed opinion from a reliable and respected source.
     
Draft saved Draft deleted

Share This Page