Options
Shoutout to Coinstudio.ai
tradedollarnut
Posts: 20,799 ✭✭✭✭✭
tradedollarnut
Posts: 20,799 ✭✭✭✭✭
Comments
Looks great. Same on I used for my infamous "AU50" Seated Quarter
Latin American Collection
Their 1892 25C PCGS AU58+ example looks pretty darn good...except LIBERTY is gibberish and TRUST is misspelled.
BST transactions: dbldie55, jayPem, 78saen, UltraHighRelief, nibanny, liefgold, FallGuy, lkeigwin, mbogoman, Sandman70gt, keets, joeykoins, ianrussell (@GC), EagleEye, ThePennyLady, GRANDAM, Ilikecolor, Gluggo, okiedude, Voyageur, LJenkins11, fastfreddie, ms70, pursuitofliberty, ZoidMeister,Coin Finder, GotTheBug, edwardjulio, Coinnmore, Nickpatton, Namvet69,...
More digital coin Frankensteins. Denticles wrong. Lettering wrong. "LIBERTY" wrong.
Looks nice to the Instagram crowd, I guess...
Superb collections are built more by omission than by addition.
Like any tool it is as good as the info given to it. From what I have seen so far it has some deficiencies, dark areas on photos seem to be one of those. Like others I've seen these look great at a glance, if you study them closely then the issues are somewhat obvious. But for those like myself who are somewhat photo challenged this is a useful tool.
My Collection of Old Holders
Never a slave to one plastic brand will I ever be.
They look nice from far away. But if you look at the coin, there are bizarre errors in the lettering and fields.
The 1909 s has two Ts in UNITED. And LIBERTY looks like it’s spelled “library”.
The 1909 proof has “UniteB”
The 1909 in MS has ghost-clashing on the cheek that’s probably not on the coin (maybe it is?)
Presumably they will get better in time.
Looks like a "Shutout for Coinstudio.ai" to me.
Ah - but I don’t care about the coin at all. It’s a holder collection
They are killer holder pictures.
I’m concerned about the abuse potential especially if you’re on a smartphone.
Hi! Will from Coinstudio here. Appreciate the shoutout.
I used to post here many years ago with the handle GoldenEyeNumismatics. Coinstudio has been a fun project I’ve been working on for a couple months. I’m still not sure if it’s a side project or an actual business.
I wanted to tackle coin photography because I used to buy and sell coins a lot on eBay (handle was yankeecoin). Most of my success there was because I got really good at taking coin pictures. I invested in cameras, lenses, lighting, stands, and software…. and then invested many hours of practice. With AI, it’s now much easier to bridge that equipment/skill gap.
Many folks are skeptical of using AI for coin photography. You are all correct in your doubts. AI can hallucinate details, embellish, lie, etc.
It’s also gotten to be quite capable, and getting better astoundingly fast.
Whether or not people use my tool, people will use AI to edit their coin pictures. So I’d rather they use a tool that’s purpose-built for coins and that has decent guardrailing built in. Instead of just throwing it into ChatGPT or Midjourney.
If any of you are curious about how Coinstudio works, I’m an open book. Ask away! Most of it is relatively simple… a few parts get much more complicated.
There's "editing coin pictures" and then, there's "editing coin pictures". Editing to adjust brightness or white balance or whatever to make the image appear on the screen more like it does in hand is not anything at all like editing it such that actual design elements are modified.
JMO, YMMV
In the case of AI, this does not apply. it's non deterministic... In most cases it's not meant to give back precision just estimates albeit very sophisticated. IMHO not a good use case rendering hyper specific aspects of coins which are always best seem in person by experts.
It'll get you in the stadium but never in the right seat. Its not the strength of the tech. My experience. Anyone who says different is likely trying to sell you something.
That’s a very broad brush assessing an advancing technology at a point in time. In some things it’ll get you exactly to the right seat and in other things, nowhere close. The pace of change is rapid, so the answer to the question today will be very different than the answer tomorrow.
My company is working with it now around improving security by identifying patterns, the absence of established patterns, and then the probability weighted scenarios that an adverse event is occurring to then take action before it occurs. These are all probabilities, but the technologies advanced to the point where you can say with 98+ percent confidence interval that’s something bad is about to happen, before it’s happened and then take action to prevent it. A year ago, the technology had not advanced to the point where this was likely possible, or it was such a great cost that it couldn’t be adopted at scale.
Point is, what seems to be a problem today will likely not be a problem in the future and experimentation is required for innovation.
Latin American Collection
Can I get you a tin foil hat to go.
Dude what the heck, all I said is this is a tool that can be useful for some and you throw out all this nonsense. I'm not selling anything nor am I buying anything.
My Collection of Old Holders
Never a slave to one plastic brand will I ever be.