Pipedia Dunhill Dating Guide V2

Log in

SmokingPipes.com Updates

96 Fresh Peterson Pipes
24 Fresh Tsuge Pipes
36 Fresh Savinelli Pipes
12 Fresh Dunhill Pipes
3 Fresh Li Zhesong Pipes

PipesMagazine Approved Sponsor

PipesMagazine Approved Sponsor

PipesMagazine Approved Sponsor

PipesMagazine Approved Sponsor

PipesMagazine Approved Sponsor

F

forciori

Guest
This guide represents its most definitive version to date, having been thoroughly revised and expanded with new additions from Loring's guide, data received directly from the factory, and invaluable information shared by fellow enthusiasts. The result of this continuous effort is a comprehensive compendium intended to encompass nearly all that is known on this particular subject.

Despite our earnest efforts to ensure the reliability of the information presented, the sheer volume of data and the extensive nature of this work mean that errors may still be present. Even within the established literature on the subject, one can find conflicting data when comparing one guide or document to another. To cite an example, Loring asserted that the logo enclosed within an ellipse first appeared in 1995. However, I have since received images that provide evidence of its application as early as 1990. Could this be a stamping error on the items I received? Perhaps. In the interest of intellectual honesty and respect for all sources, such discrepancies are preserved alongside new evidence.

Should you possess any new information, or wish to point out any divergence, error, or imprecision, I encourage you to contact me. My details are provided on the page.

A Note on the Effective Use of the AI Tool

I wish to emphasize that for optimal performance, the AI tool must be queried with precise and complete data, such as the full patent or the entire stamping. The tool's functionality is exclusively restricted to date codes, including nuanced details like underlining or the use of possessives.

The most effective approach is to provide the tool with the same comprehensive information you would offer a human expert for analysis. This clarification is necessary because an observation of the anonymous logs revealed queries that fall outside the tool's scope, including questions about lighter patents, shape numbers, and incomplete patents.

This is, of course, an optional resource.


Thanks =)
 

ssjones

Moderator
Staff member
May 11, 2011
20,039
15,798
Covington, Louisiana
postimg.cc
Interesting, kind of clunky (didn't recognize the nomenclature photo links I submitted.
But, subsequent info from me has this at 1964. which is the date I had determined. I guess that was an easy one.
Dunhill_747_RB_FIN (9).jpgDunhill_747_RB_FIN (10).jpgDunhill_747_RB_FIN (11).jpg
 
  • Like
Reactions: forciori
F

forciori

Guest
Interesting, kind of clunky (didn't recognize the nomenclature photo links I submitted.
But, subsequent info from me has this at 1964. which is the date I had determined. I guess that was an easy one.
View attachment 410081View attachment 410082View attachment 410083
Hi Jones,
Thanks for flagging this.

I saw the error in yesterday's logs (though I hadn't realized it was yours) — it was certainly an unexpected one.
I've already adjusted the AI's instructions and confirmed the fix with a similar query. Your feedback is key to this fine-tuning process.

Apologies for the disruption. It should be working correctly now.
 
F

forciori

Guest
I consider the response to be satisfactory. For such a quick answer, the amount of quality information it provides is incredible.

I still need to implement enhancements so the model can anticipate this type of scenario, as my last question demonstrated.

It currently relies on obvious interpretations, which requires the prompt to be very explicit. I will continue to refine it to become increasingly precise and intuitive.

Of course, there may be errors, but the model will continue to adapt with use and should soon be operating at its full potential.
 

ssjones

Moderator
Staff member
May 11, 2011
20,039
15,798
Covington, Louisiana
postimg.cc
The queries did seem like a logical progression.
Initially, I thought it was asking for pictures of the stamping, like we would here. I put in URL's to several photos of the stamping from a web photo hosting site. The program of course didn't recognize the photo, so started asking the relevant questions above.
George posted photos of a Dunhill a few weeks ago, where the stamping wasn't consistent with typical nomenclature. It would be interesting to feed it that data.
 
  • Like
Reactions: forciori
F

forciori

Guest
The queries did seem like a logical progression.
Initially, I thought it was asking for pictures of the stamping, like we would here. I put in URL's to several photos of the stamping from a web photo hosting site. The program of course didn't recognize the photo, so started asking the relevant questions above.
George posted photos of a Dunhill a few weeks ago, where the stamping wasn't consistent with typical nomenclature. It would be interesting to feed it that data.
Imagine giving this guide to a child and expecting them to interpret information non-literally in a completely new context — and ultimately, to be able to discuss the topic with you on equal footing. It's a lengthy educational process.

We still have limitations. It's a free platform and, therefore, has its restrictions. So, this "educational process" requires a great deal of dedication and patience. It might be possible to host the model directly on the Pipedia server — something that is currently under discussion with Scott. That would grant us incredible new capabilities. We'll see.

I plan to implement something similar on the Dunhill main page, and if that's successful, perhaps for all of Pipedia.
But make no mistake, even as a technology enthusiast, I have concerns about this tool's application. Will people stop appreciating the full scope of information on the portal, including its semiotic aspects? Perhaps. But I can tell you from observation that content consumption on these pages is low. As a rule, people are reluctant to read. In the end, the Dunhill page, for example, is rarely read. Even the guide is seldom applied. It's not uncommon to find people asking how to date something even after consulting the guide. Naturally and instinctively, we follow the path of least resistance. That said, I hold a faint hope that artificial intelligence can at least present some of the denser content with less effort required from the user. We'll see.

Your suggestions are interesting, and I will work on them as well. Thanks, Jones.
 
  • Like
Reactions: jguss and ssjones
F

forciori

Guest
Jones, I took the stamps from the pipe you suggested, and without needing to make any adjustments to the instructions, the result, it seems, was accurate. 1000171904.jpg1000171905.jpg
 
  • Like
Reactions: jguss and ssjones
F

forciori

Guest
I've just realized I've worked myself out of a job. I've become a paperweight and am no longer any use for dating pipes.

As Einstein would say, a mouse wouldn't build a mousetrap. Shame on me.
 
  • Haha
Reactions: jguss