Researching rarity with AI...

MrEureka

Has anyone had much luck with AI in building a census for specific coins? At some point, it should be possible to pull up all auction records, have AI match duplicate appearances of the same coin wherever possible, and come up with something approaching a census for any rare coin, even if thousands are known.

pruebas

Most of what I’ve seen from AI has been wrong. So I no longer look at AI-generated results.

It’s still typical Silicon Valley hype at this point.

Edited to add the following: The old saying is as true as it’s ever been. Garbage in. Garbage out. And a great deal of “knowledge” on the internet is “garbage.” That garbage is training AI.

PeakRarities

I've been trying to work with Grok, but there's still too many limitations. I think It will be useful for cataloguing purposes, general background info, rarity and mintages, auction descriptions, but I've only been able to get so far.

jmlanzaf

@pruebas said:
Most of what I’ve seen from AI has been wrong. So I no longer look at AI-generated results.

It’s still typical Silicon Valley hype at this point.

Edited to add the following: The old saying is as true as it’s ever been. Garbage in. Garbage out. And a great deal of “knowledge” on the internet is “garbage.” That garbage is training AI.

It is certainly not hype. They've done tens of thousands of protein structures using it, saving thousands of person-years. I don't know of a prominent researcher under 40 who doesn't use it for scholarly articles and grant requests. And it's worth noting that 2 Nobel prizes were awarded this year in the field. It is already changing how everything is done, internet searches being the least of it.

What one must be careful about is equating LLMs with AI and assuming that a query to an LLM is the solution to all of life's problems.

You could do what the OP suggested but I think you need to build an app to do it... or wait till someone else does. The sorting would not be that simple since not every sale/ listing will use the same photos, have the cert number available. You'd also have to match raw coins with slabbed coins. It could be done but not just with a query to an LLM.

PeakRarities

Here's an example of what I normally can pull from it-

Can you tell me what you know about the Baldwin $10

25 web pages**

The Baldwin $10 coin, often referred to as the "Horseman" or "El Vaquero," is a notable piece from the California Gold Rush era. Here's what is known about this coin:

Design and Origin: The coin was minted by Baldwin & Co. in San Francisco during the California Gold Rush, specifically in 1850. The obverse of the coin features a cowboy (vaquero) on horseback with a lasso, with the inscription "CALIFORNIA GOLD" and "TEN DOLLARS." The reverse displays a federal-style eagle surrounded by the text "BALDWIN & Co." and "SAN FRANCISCO," with five stars on each side. The design of the cowboy was inspired by an 1828 watercolor painting titled "Californians Throwing the Lasso" by Frederick William Beechey, reflecting California's Mexican heritage during the Gold Rush period.

Rarity and Value: This coin is extremely rare, with fewer than 20 believed to exist today. The condition of the coin plays a significant role in its value; for instance, an example graded MS63+ sold for $1,260,000 in November 2024, highlighting its significant numismatic value. The rarity is due to both the low mintage during the Gold Rush and the low survival rate of these coins.

Restrikes: Around the turn of the 20th century, restrikes of this coin were made using similar dies, created in various metals including gold, silver, copper, brass, lead, and white metal. These restrikes are also collectible, with some being extremely scarce. They differ slightly from the originals, for instance, having plain edges instead of reeded ones.

Historical Context: Baldwin & Co. was one of several private mints operating in California during the Gold Rush because of the need for standardized currency when official banking systems were not fully established. This coin represents not just a piece of currency but a snapshot of the cultural and economic dynamics of the time.

The Baldwin $10 gold coin is thus both a historical artifact and a collector's dream, valued not only for its scarcity but also for its artistic depiction of the American West and its connection to the Gold Rush era."

PeakRarities

In your opinion, what are some of the other pioneer gold issues with the most historical significance?

"Here are some other pioneer gold issues that are considered to have significant historical importance:

Mormon $5, $10, and $20 Gold Pieces (1849-1860):
Historical Significance: Minted by the Mormon Church in Salt Lake City, these coins were produced to alleviate the shortage of coinage in the Utah Territory during the Mormon migration. They represent a unique blend of religious history and economic necessity. The $20 coin, in particular, is one of the rarest with only a few known examples.

Templeton Reid $5 and $10 Gold Pieces (1850s):
Historical Significance: Made by Templeton Reid in Georgia, these coins were one of the earliest private gold issues in the United States, predating the California Gold Rush. They are significant for showing private minting efforts in the South before the Civil War.

Augustus Humbert $50 Gold Slug (1851-1852):
Historical Significance: The largest gold coin ever officially struck for circulation in the U.S., these "slugs" were minted by Augustus Humbert in California during the Gold Rush. They represent the peak of private minting before the U.S. government stepped in with official coinage. Their size and rarity make them highly sought after.

Christopher Bechtler $1, $2.50, $5 Gold Coins (1830s-1840s):
Historical Significance: Produced in Rutherfordton, North Carolina, by Christopher Bechtler and his son, these were among the earliest gold coins minted in the United States. They are significant for illustrating the early private minting of gold in the South, well before the California Gold Rush.

Norris, Gregg & Norris $5 Gold (1849):
Historical Significance: This coin was one of the first private gold issues from the California Gold Rush era, minted in San Francisco. It highlights the immediate entrepreneurial response to the discovery of gold and the need for currency in the rapidly growing town.

Clark, Gruber & Co. $2.50, $5, $10, $20 Gold Pieces (1860-1861):
Historical Significance: Minted in Denver, Colorado, before the U.S. Mint took over, these coins are significant for their role in the economic development of the Colorado Gold Rush. They also represent the transition from private to federal minting in the West.

Dubosq & Co. $5 Gold (1849):
Historical Significance: One of the shortest-lived private mints in San Francisco during the Gold Rush, Dubosq & Co. produced this $5 gold piece which is rare and emblematic of the rapid turnover and speculative nature of early minting operations in California.

Each of these issues not only serves as a testament to the gold rushes and the entrepreneurial spirit of the times but also illustrates the private sector's role in currency during periods of governmental absence or delay in providing coinage. Their rarity, combined with their historical narratives, makes them highly significant in the study of U.S. numismatics."

What can you tell me about the Templeton Reid $25 gold coin?

25 web pages

The Templeton Reid $25 gold coin is one of the most enigmatic and rare pieces in American numismatics. Here's what is known about it:

Existence and Rarity: Only one Templeton Reid $25 gold coin is known to have ever existed. This coin was minted in 1849, during the California Gold Rush period.
Design:
Obverse: The coin features the words "CALIFORNIA," "TWENTY FIVE," and "DOLLARS," with "GOLD" at the bottom.
Reverse: It displays "TEMPLETON REID . ASSAYER . 1849" around the coin's edge.

Historical Context:
Templeton Reid originally gained fame by minting gold coins in Georgia in 1830, but the $25 coin was produced much later, likely using gold from California. There's debate over whether Reid personally went to California or if the coin was made from California gold without him actually being there.
The coin was part of an experimental or speculative minting effort, possibly to capitalize on the burgeoning gold rush or to establish another "first" in private minting.

Current Status:
This particular $25 gold coin was stolen from the United States Mint Cabinet Collection on August 16, 1858, and has never been recovered. Its whereabouts remain unknown, leading to numerous theories about its fate, including the possibility that it was melted down for its bullion value or is still in private hands.

Speculation and Interest:
The coin's rarity has led to significant speculation and interest within the numismatic community. It's considered one of the holy grails of American coin collecting due to its unique status as the only one of its kind ever produced.
There have been numerous searches and even rewards offered for its recovery, with PCGS (Professional Coin Grading Service) offering $10,000 for the first submission of this coin for grading and authentication.

Replicas and Restrikes:
Given its fame and the fact that it's lost, replicas or fantasy pieces have been made, often as tokens or souvenirs. These are not the actual coin but are intended for collectors who want a piece reminiscent of this legendary coin.

Value:
If ever found, this coin would likely be one of the most valuable in existence due to its singular nature, historical significance, and the story of its theft and disappearance.

In summary, the Templeton Reid $25 gold coin is not only a numismatic rarity but also a piece of American history, symbolizing the adventurous and speculative spirit of the Gold Rush era. Its theft has added a layer of mystery and allure, making it a subject of continuous fascination and speculation in the world of coin collecting.

In summary, I think it's fantastic for aggregating facts and information and can save so much time for research if you're in a hurry. It tends to give broad, general information and avoids being too speculative or opinionated. I'm a fan!

calgolddiver

The LLMs will certainly have the data and the ability to provide the inquiry highlights represented above by @PeakRarities. There is also the ability for deeper analysis of pop reports, auction archives, and any other data available to the LLM (that is one of the keys - what data is included).

These LLMs can also learn over time as additional data is available. However, all LLMs are subject to "hallucinations" which is an industry term for incorrect responses. All AI and LLM queries should have a HITL (human in the loop) step to validate results before any automated publication or guidance for actions.

Boosibri

I have had luck in using AI to find examples in building census (the 1794 West Indies 3 Gulden) and our past project on the 1777 Greenland Dollar.

The key is in asking specific questions...
"Which museums have examples of the 1777 Greenland Dollar" -
"Does the National Museum of Denmark have a Greenland Dollar" - Copilot returns a helpful answer while ChatGPT is garbage.

I keep using it and testing the questions to find the best means of getting a meaningful response.

Cougar1978

That wb nice if AI could give rarity consensus (based on pop data) on slabbed coins and graded banknotes, especially world. It could be around the corner for all I know.

fathom

@jmlanzaf said:

@pruebas said:
Most of what I’ve seen from AI has been wrong. So I no longer look at AI-generated results.

It’s still typical Silicon Valley hype at this point.

Edited to add the following: The old saying is as true as it’s ever been. Garbage in. Garbage out. And a great deal of “knowledge” on the internet is “garbage.” That garbage is training AI.

It is certainly not hype. They've done tens of thousands of protein structures using it, saving thousands of person-years. I don't know of a prominent researcher under 40 who doesn't use it for scholarly articles and grant requests. And it's worth noting that 2 Nobel prizes were awarded this year in the field. It is already changing how everything is done, internet searches being the least of it.

What one must be careful about is equating LLMs with AI and assuming that a query to an LLM is the solution to all of life's problems.

You could do what the OP suggested but I think you need to build an app to do it... or wait till someone else does. The sorting would not be that simple since not every sale/ listing will use the same photos, have the cert number available. You'd also have to match raw coins with slabbed coins. It could be done but not just with a query to an LLM.

You were too easy on him, he posted a ridiculous comment.

Best application for AI is still grading. Data analysis will be prone to bias and lack of context.

TwoSides2aCoin

I’m still adjusting to artificial flavors.

dcarr

For me, reading AI-generated text is like eating empty calories - it doesn't really give me what I need.

I have not used AI myself. What would impress me is if AI could look at this image and determine who the maker was and what the logo means:

jmlanzaf

@dcarr said:
For me, reading AI-generated text is like eating empty calories - it doesn't really give me what I need.

I have not used AI myself. What would impress me is if AI could look at this image and determine who the maker was and what the logo means:

Have you tried Google lens or another image search?

An AI can't (shouldn't) create information that is not already available.

lermish

@fathom said:

@jmlanzaf said:

@pruebas said:
Most of what I’ve seen from AI has been wrong. So I no longer look at AI-generated results.

It’s still typical Silicon Valley hype at this point.

Edited to add the following: The old saying is as true as it’s ever been. Garbage in. Garbage out. And a great deal of “knowledge” on the internet is “garbage.” That garbage is training AI.

It is certainly not hype. They've done tens of thousands of protein structures using it, saving thousands of person-years. I don't know of a prominent researcher under 40 who doesn't use it for scholarly articles and grant requests. And it's worth noting that 2 Nobel prizes were awarded this year in the field. It is already changing how everything is done, internet searches being the least of it.

What one must be careful about is equating LLMs with AI and assuming that a query to an LLM is the solution to all of life's problems.

You could do what the OP suggested but I think you need to build an app to do it... or wait till someone else does. The sorting would not be that simple since not every sale/ listing will use the same photos, have the cert number available. You'd also have to match raw coins with slabbed coins. It could be done but not just with a query to an LLM.

You were too easy on him, he posted a ridiculous comment.

Best application for AI is still grading. Data analysis will be prone to bias and lack of context.

I think the majority of people's interactions (probably a higher percentage of people on this board given the age demographics) think that AI is basically the eBay computer written auction descriptions.

Take what @jmlanzaf said about protein structures and apply it to basically every field that requires repetitive technical expertise or the need to parse large amounts of data.

AI is already in use and is changing our lives and economy (probably for the worse IMO but that is a much longer and unrelated conversation). Anyone who thinks this is only hype is either not paying attention or doesn't know what they are talking about.

EDIT: Here is an AI response to pruebas' criticism:

ChatGPT said:
ChatGPT
"AI isn't perfect, but dismissing it outright ignores its real advancements in fields like medicine, finance, and education. Garbage in, garbage out applies to any tool—it's up to us to use AI critically and responsibly, not to throw the baby out with the bathwater."

You said:
let's start over.
ChatGPT said:
ChatGPT
Sure! Here's a concise and pithy response:

"AI isn’t flawless, but it’s far more than just hype. While garbage data can lead to bad outputs, modern AI excels when used with high-quality input and critical oversight. It's a tool—not a magic wand—and dismissing it entirely overlooks its real potential."

You said:
I want a short, pithy, and mildly sarcastic response
ChatGPT said:
ChatGPT
"Ah, yes, because dismissing all AI based on bad examples is clearly the pinnacle of critical thinking. Tools are only as good as how we use them—AI included."

You said:
i would like a dismissive and rude response
ChatGPT said:
ChatGPT
"Got it—AI's the problem, not the user's inability to understand how to use it. Makes perfect sense."

jmlanzaf

@lermish said:

@fathom said:

@jmlanzaf said:

@pruebas said:
Most of what I’ve seen from AI has been wrong. So I no longer look at AI-generated results.

It’s still typical Silicon Valley hype at this point.

Edited to add the following: The old saying is as true as it’s ever been. Garbage in. Garbage out. And a great deal of “knowledge” on the internet is “garbage.” That garbage is training AI.

It is certainly not hype. They've done tens of thousands of protein structures using it, saving thousands of person-years. I don't know of a prominent researcher under 40 who doesn't use it for scholarly articles and grant requests. And it's worth noting that 2 Nobel prizes were awarded this year in the field. It is already changing how everything is done, internet searches being the least of it.

What one must be careful about is equating LLMs with AI and assuming that a query to an LLM is the solution to all of life's problems.

You could do what the OP suggested but I think you need to build an app to do it... or wait till someone else does. The sorting would not be that simple since not every sale/ listing will use the same photos, have the cert number available. You'd also have to match raw coins with slabbed coins. It could be done but not just with a query to an LLM.

You were too easy on him, he posted a ridiculous comment.

Best application for AI is still grading. Data analysis will be prone to bias and lack of context.

I think the majority of people's interactions (probably a higher percentage of people on this board given the age demographics) think that AI is basically the eBay computer written auction descriptions.

Take what @jmlanzaf said about protein structures and apply it to basically every field that requires repetitive technical expertise or the need to parse large amounts of data.

AI is already in use and is changing our lives and economy (probably for the worse IMO but that is a much longer and unrelated conversation). Anyone who thinks this is only hype is either not paying attention or doesn't know what they are talking about.

EDIT: Here is an AI response to pruebas' criticism:

ChatGPT said:
ChatGPT
"AI isn't perfect, but dismissing it outright ignores its real advancements in fields like medicine, finance, and education. Garbage in, garbage out applies to any tool—it's up to us to use AI critically and responsibly, not to throw the baby out with the bathwater."

You said:
let's start over.
ChatGPT said:
ChatGPT
Sure! Here's a concise and pithy response:

"AI isn’t flawless, but it’s far more than just hype. While garbage data can lead to bad outputs, modern AI excels when used with high-quality input and critical oversight. It's a tool—not a magic wand—and dismissing it entirely overlooks its real potential."

You said:
I want a short, pithy, and mildly sarcastic response
ChatGPT said:
ChatGPT
"Ah, yes, because dismissing all AI based on bad examples is clearly the pinnacle of critical thinking. Tools are only as good as how we use them—AI included."

You said:
i would like a dismissive and rude response
ChatGPT said:
ChatGPT
"Got it—AI's the problem, not the user's inability to understand how to use it. Makes perfect sense."

Lol. Perfect!

One of my students needed help and asked ChatGPT to "explain molecular bonding to me assuming I'm an idiot". He said the explanation was really helpful.

On a more serious note, I have a colleague who does biomedical engineering research. His group contains graduate students who are either engineers and biologists. He has ChatGPT explain the engineering papers to the biologists and vice versa. He fears by it and had actually been paying out of pocket for 2 years.

While it's not perfect, it is rapidly evolving and will change everything, including numismatic research. At some point, you will likely have a RAG agent attached to the NNP to pull information from those archives. That increases the accuracy and cuts down on the hallucinations.

Baylor8670

@jmlanzaf said:

@dcarr said:
For me, reading AI-generated text is like eating empty calories - it doesn't really give me what I need.

I have not used AI myself. What would impress me is if AI could look at this image and determine who the maker was and what the logo means:

Have you tried Google lens or another image search?

An AI can't (shouldn't) create information that is not already available.

I got a better idea. How about AI does an image search and answers his questions if it's really supposed to be so great.

dcarr

@jmlanzaf said:

@dcarr said:
For me, reading AI-generated text is like eating empty calories - it doesn't really give me what I need.

I have not used AI myself. What would impress me is if AI could look at this image and determine who the maker was and what the logo means:

Have you tried Google lens or another image search?

An AI can't (shouldn't) create information that is not already available.

.

I haven't tried that. I'm not familiar with how to go about doing that.
But I would be impressed if AI could find other instances of this logo and lead to the identity of the maker.

.

jmlanzaf

@Baylor8670 said:

@jmlanzaf said:

@dcarr said:
For me, reading AI-generated text is like eating empty calories - it doesn't really give me what I need.

I have not used AI myself. What would impress me is if AI could look at this image and determine who the maker was and what the logo means:

Have you tried Google lens or another image search?

An AI can't (shouldn't) create information that is not already available.

I got a better idea. How about AI does an image search and answers his questions if it's really supposed to be so great.

Google lens is an AI tool. It can only answer the question accurately if the information is out there already.

AI is great and getting greater. To expect it to be omniscient is just silly. If you think that YOU are better than the AI, why don't you look at the picture and answer the question? We'll wait.

AI solved more protein structures in 5 years than 10,000 graduate students could have. To think that it isn't an incredibly powerful TOOL is simply based on ignorance of what it already does, including internet searches, autocorrect, search suggestions, shopping hints, proof-reading, etc. I am a college chemistry teacher. My colleagues and I use AI tools DAILY to do scientific research, create papers and exams, to grade papers and exams, to help write peer-reviewed papers, to write emails and letters of recommendation. It can read papers and create executive summaries. It can find scholarly references. It can create PowerPoint presentations.

If you have a smartphone or have an Alexa or Google Home setup, you're already using AI tools.

Please explain in simple terms what AI tools are currently used for and what the future potential is.

What AI Tools Are Currently Used For

AI tools are designed to perform specific tasks by mimicking human intelligence. Here's how they're used today:

1. Automation: AI streamlines repetitive tasks, like data entry, email sorting, or manufacturing processes, saving time and reducing human error.
2. Customer Service: Chatbots and virtual assistants like Siri or Alexa answer questions, resolve issues, and provide support around the clock.
3. Data Analysis: AI processes large amounts of data to identify patterns, trends, and insights—used in fields like marketing, finance, and healthcare.
4. Healthcare: AI assists in diagnosing diseases, analyzing medical images, and personalizing treatment plans.
5. Content Creation: Tools like language models (e.g., ChatGPT) help write articles, summarize information, and generate creative ideas.
6. Entertainment: AI powers recommendation systems for movies, music, and books (e.g., Netflix and Spotify).
7. Security: AI detects fraud, monitors systems for threats, and enhances cybersecurity defenses.
8. Transportation: Autonomous vehicles and route optimization for delivery services rely on AI algorithms.
9. Education: Personalized learning platforms adapt to individual student needs, providing tailored content and feedback.
10. Environment: AI models predict weather patterns, manage energy use, and monitor wildlife.

---

Future Potential of AI Tools

The potential for AI tools is vast, and here are a few possibilities:

1. Advanced Healthcare: AI could enable faster drug discovery, more precise surgeries with robotics, and highly personalized medicine.
2. Education Revolution: AI tutors might provide one-on-one learning for everyone, regardless of location or language barriers.
3. Improved Decision-Making: Businesses and governments could use AI to analyze complex problems and suggest optimal solutions.
4. Sustainable Development: AI might help tackle climate change by optimizing renewable energy systems, reducing waste, and managing resources better.
5. Enhanced Creativity: AI could co-create art, music, and literature alongside humans, pushing creative boundaries.
6. Global Connectivity: Language translation tools may break down communication barriers, fostering better collaboration worldwide.
7. Fully Autonomous Systems: From vehicles to factories, AI could automate complex, multi-step processes with minimal human intervention.
8. Ethical AI: Systems designed to address bias and fairness issues, ensuring AI benefits everyone equitably.
9. Human-AI Collaboration: AI may act as "co-pilots," amplifying human abilities in various tasks rather than replacing them.
10. Exploration: AI might lead space exploration or help explore the deep ocean, environments where human survival is challenging.

AI's growth depends on ethical deployment, addressing biases, and ensuring it aligns with human values to maximize its benefits for society.

jmlanzaf

@dcarr said:

@jmlanzaf said:

@dcarr said:
For me, reading AI-generated text is like eating empty calories - it doesn't really give me what I need.

I have not used AI myself. What would impress me is if AI could look at this image and determine who the maker was and what the logo means:

Have you tried Google lens or another image search?

An AI can't (shouldn't) create information that is not already available.

.

I haven't tried that. I'm not familiar with how to go about doing that.
But I would be impressed if AI could find other instances of this logo and lead to the identity of the maker.

.

Google lens didn't come up with an exact match to the logo. Of course, what that means is there may not be any images of the same logo on the internet. That is less a failing of the AI and more a limitation of the data that's out there.
"

dcarr

@jmlanzaf said:

@dcarr said:

@jmlanzaf said:

@dcarr said:
For me, reading AI-generated text is like eating empty calories - it doesn't really give me what I need.

I have not used AI myself. What would impress me is if AI could look at this image and determine who the maker was and what the logo means:

Have you tried Google lens or another image search?

An AI can't (shouldn't) create information that is not already available.

.

I haven't tried that. I'm not familiar with how to go about doing that.
But I would be impressed if AI could find other instances of this logo and lead to the identity of the maker.

.

Google lens didn't come up with an exact match to the logo. Of course, what that means is there may not be any images of the same logo on the internet. That is less a failing of the AI and more a limitation of the data that's out there.
"

.

I've manually looked for images that contain this logo. But there are billions of pictures out there on the internet. So I can only check an extremely small fraction of them. If an AI tool could do pattern matching for this logo across billions of images, that would be extremely useful of course.

.

lermish

@dcarr said:

@jmlanzaf said:

@dcarr said:

@jmlanzaf said:

@dcarr said:
For me, reading AI-generated text is like eating empty calories - it doesn't really give me what I need.

I have not used AI myself. What would impress me is if AI could look at this image and determine who the maker was and what the logo means:

Have you tried Google lens or another image search?

An AI can't (shouldn't) create information that is not already available.

.

I haven't tried that. I'm not familiar with how to go about doing that.
But I would be impressed if AI could find other instances of this logo and lead to the identity of the maker.

.

Google lens didn't come up with an exact match to the logo. Of course, what that means is there may not be any images of the same logo on the internet. That is less a failing of the AI and more a limitation of the data that's out there.
"

.

I've manually looked for images that contain this logo. But there are billions of pictures out there on the internet. So I can only check an extremely small fraction of them. If an AI tool could do pattern matching for this logo across billions of images, that would be extremely useful of course.

.

They can and do but unfortunately there's nothing out there that matches it. I gave it a shot this afternoon but it couldn't find anything. At the same time I'm using a free consumer level version so maybe a more powerful one would be able to find something.

dcarr

@jmlanzaf said:

@dcarr said:

@jmlanzaf said:

@dcarr said:
For me, reading AI-generated text is like eating empty calories - it doesn't really give me what I need.

I have not used AI myself. What would impress me is if AI could look at this image and determine who the maker was and what the logo means:

Have you tried Google lens or another image search?

An AI can't (shouldn't) create information that is not already available.

.

I haven't tried that. I'm not familiar with how to go about doing that.
But I would be impressed if AI could find other instances of this logo and lead to the identity of the maker.

.

Google lens didn't come up with an exact match to the logo. Of course, what that means is there may not be any images of the same logo on the internet. That is less a failing of the AI and more a limitation of the data that's out there.
"

.

PS:
There are images of other bars on the internet with this logo (I found them using a traditional search). The other bars have "1970" inside the circle, but otherwise the same. Did your attempt using Google Lens find any of the "1970" images ?

Here is one of them:
https://worthpoint.com/worthopedia/vintage-1970-mte-bell-troy-oz-999-3934212042

.

There are also some images in this forum thread:
https://forums.collectors.com/discussion/1082948/silver-bar-logo-what-maker

.

Mr_Spud

@dcarr said:
For me, reading AI-generated text is like eating empty calories - it doesn't really give me what I need.

I have not used AI myself. What would impress me is if AI could look at this image and determine who the maker was and what the logo means:

I know you already found the vague references to MTE Bell, well AI was able to find that too if you ask the question the right way. This was with google lens and it allows you to add info like a question, but you also have to crop the image so it doesn’t focus on other things in the image. Like this

jmlanzaf

Prompt: I would like 1000 word essay on the history of coinage starting in antiquity and ending with the introduction of the steam press. Please write in the style of an academic historian.

The history of coinage is a rich and intricate tale that reflects the development of human societies, economies, and technologies. From the earliest metal tokens used in antiquity to the mechanized processes introduced during the industrial revolution, the evolution of coinage mirrors the broader trajectory of civilization. This essay traces the history of coinage from its origins in ancient societies through to the introduction of the steam press in the late 18th and early 19th centuries, a pivotal moment in the modernization of currency production.

The Origins of Coinage in Antiquity

Coinage, as a formalized medium of exchange, emerged in the ancient world as societies sought to standardize trade and facilitate economic transactions. The earliest coins are widely believed to have been minted in the ancient kingdom of Lydia, in what is now western Turkey, during the 7th century BCE. These coins were made of electrum, a naturally occurring alloy of gold and silver, and bore simple designs such as lions or geometric patterns. The use of precious metals provided intrinsic value, while standardized weights ensured consistent worth.

The spread of coinage from Lydia to neighboring civilizations was rapid. The Greek city-states, for instance, adopted and refined the practice, producing coins with increasingly intricate designs and inscriptions. By the 6th century BCE, Greek coinage featured depictions of gods, heroes, and symbols of civic pride, such as the owl of Athens. These coins not only facilitated trade but also served as tools of political propaganda, asserting the identity and sovereignty of the issuing city-state.

Meanwhile, in the Persian Empire, coinage became an instrument of imperial control. The Achaemenid rulers issued large quantities of gold and silver coins, known as darics and siglos, which circulated across their vast territories. These coins helped to unify the empire's diverse economies and underscored the authority of the central government. Similarly, in ancient India, the use of punch-marked coins—small pieces of silver stamped with symbols—reflected both the economic dynamism and the cultural diversity of the subcontinent during the Mauryan period (4th to 2nd centuries BCE).

The Roman Contribution to Coinage

The Roman Empire represents a critical chapter in the history of coinage, characterized by both innovation and widespread influence. Roman coinage began in the 4th century BCE with the introduction of aes grave, large cast bronze coins. By the 3rd century BCE, Rome had adopted the Greek practice of striking coins, producing silver denarii and gold aurei that became the standard currency of the Mediterranean world.

Roman coins were notable for their detailed imagery and inscriptions, which celebrated the achievements of emperors, military victories, and religious devotion. The use of coins as a medium for imperial propaganda reached its zenith under Augustus, the first Roman emperor, whose coinage depicted him as the bringer of peace and prosperity. The Roman monetary system was also remarkable for its sophistication, encompassing a wide range of denominations and metals to serve diverse economic needs.

However, the Roman Empire's reliance on precious metals posed challenges. Economic instability, military expenditures, and declining access to silver mines led to the debasement of coinage in the later empire. Coins increasingly contained less precious metal, undermining their value and contributing to inflation. The collapse of the Western Roman Empire in the 5th century CE disrupted the continuity of coinage in Europe, though Byzantine coins continued to circulate in the eastern Mediterranean for centuries.

Medieval Developments in Coinage

The medieval period witnessed significant regional variations in coinage, reflecting the fragmented political landscape of post-Roman Europe. In the early Middle Ages, many regions reverted to barter or used foreign coins due to the decline of centralized minting. By the 8th century, however, the Carolingian Empire under Charlemagne revived coinage on a large scale, standardizing the silver denarius as the principal currency across much of Western Europe.

The Islamic world, meanwhile, developed a flourishing monetary tradition, building on both Byzantine and Persian models. Gold dinars and silver dirhams were widely circulated, facilitating trade across the Islamic caliphates and beyond. These coins often bore inscriptions from the Quran, emphasizing their role in reinforcing religious and cultural identity.

During the High Middle Ages, the rise of powerful monarchies and burgeoning trade networks spurred innovations in coinage. The introduction of gold coins such as the florin in Florence (1252) and the ducat in Venice (1284) reflected the growing importance of international commerce and the wealth generated by Mediterranean trade. Advances in minting technology, including the use of screw presses, allowed for more precise and uniform coins, though production remained labor-intensive.

The Renaissance and Early Modern Innovations

The Renaissance ushered in a new era of artistic and technological advancements in coinage. Italian city-states, in particular, produced coins of remarkable beauty, featuring intricate portraits and elaborate designs that showcased the era's artistic sensibilities. The widespread adoption of portraiture on coins reflected the influence of humanism and the growing power of individual rulers.

The expansion of European colonial empires during the 16th and 17th centuries had a profound impact on coinage. Vast quantities of gold and silver extracted from the Americas were minted into coins such as the Spanish real and the Portuguese escudo, which became global currencies. These coins facilitated trade across continents, from Asia to the Americas, laying the foundations for the modern global economy.

Technological innovations during this period began to transform the minting process. The screw press, introduced in the late 15th century, allowed for greater precision and efficiency, while reducing reliance on manual labor. Coins became more uniform in size and weight, improving their reliability and acceptance in trade.

The Introduction of the Steam Press

The advent of the Industrial Revolution in the 18th century marked a turning point in the history of coinage. The development of the steam engine by James Watt and others revolutionized manufacturing processes, including minting. In 1786, Matthew Boulton, an English industrialist, introduced the steam press at his Soho Mint in Birmingham. This technology enabled the mass production of coins with unprecedented precision and consistency.

The steam press represented a culmination of centuries of innovation. It utilized high-pressure steam to strike coins, ensuring uniformity and reducing production costs. The introduction of this technology coincided with growing demand for small-denomination coins to support the expanding industrial economy and urban labor force. Boulton's partnership with the Royal Mint led to the widespread adoption of steam-powered minting across Europe and beyond.

The steam press also facilitated the incorporation of security features, such as intricate designs and reeded edges, which helped to prevent counterfeiting. By the early 19th century, mechanized minting had become the standard, laying the groundwork for the modern coinage systems we know today.

Conclusion

The history of coinage, from its origins in ancient Lydia to the industrial innovations of the 18th century, reflects the dynamic interplay between economic needs, technological advancements, and cultural expressions. Coins have served not only as mediums of exchange but also as tools of political authority, artistic creativity, and social cohesion. The introduction of the steam press marked a watershed moment, heralding a new era of efficiency and standardization in currency production. As we trace the journey of coinage through the ages, we gain insight into the broader currents of human history and the enduring importance of money in shaping our world.

SmEagle1795

@MrEureka said:
Has anyone had much luck with AI in building a census for specific coins? At some point, it should be possible to pull up all auction records, have AI match duplicate appearances of the same coin wherever possible, and come up with something approaching a census for any rare coin, even if thousands are known.

AI doesn't lend itself well to a census-style use case: that is more a job for a conventional structured database. Image recognition for exact matches, however, is beginning to be built in services like ACSearch's image recognition, CoinCabinet.io's pedigree search, and Ex-Numis which is human supervised but still able to find an exact match. These services are more ancients focused at the moment as varied flans are easier to compare to find the exact same coin.

ShadyDave

@Boosibri said:
I have had luck in using AI to find examples in building census (the 1794 West Indies 3 Gulden) and our past project on the 1777 Greenland Dollar.

The key is in asking specific questions...

That is called "prompting". Knowing how to prompt a GenAI tool determines how successful your prompt will be. "RICE" is a good structure to use.

R- Role - explain the role you want the AI to embody.
I- Instruction - Provide detailed instructions/guidelines for the AI to follow.
C- Context - Give the tool relevant context and background info to generate a response.
E- Example- include examples/instances so the tool can use it as a reference to create a similar output.