Artificial Information -the new How To Authority???

[alexiskai]

Originally Posted by johnneilson

AI uses nothing more than the data published on the web.
That being the case, with all the old wives tales and mis-information from videos, I feel sorry for new folks who start into this hobby.

John

Because they'll be getting their poor practices handed down from videos instead of from the random preferences of guys in their club?

[BRENT in 10-uh-C]

Originally Posted by alexiskai

Because they'll be getting their poor practices handed down from videos instead of from the random preferences of guys in their club?

Colin, either way much of the info published is still wrong.

[johnneilson]

Going back to the days of bulletin boards, you have to understand that the information posted is worth exactly what you paid for it.

There are some very good clubs and individual posts that are accurate, others not so much.

A friend from work has been playing with AI for a while and we found some very interesting information. Most of it nonsense, at least from a technical standpoint on subjects that are not "trending" at a particular time.

Computers will always be nothing more than a tool subject to the data input.

J

[jrelliott]

I helped develop some of the computers AI is using. One of the applications which we worked in conjunction with Mayo clinic was in the area of diagnostics for patients. The doctor can input all the symptoms and the computer will look through all the available information stored in books, journals etc. and try to give the doctor the best educated guess. This is just one application where it is really helpful. As more queries are made and more accurate diagnostics are achieved, then it becomes a valuable tool. If I had a unusual ailment, would want to go to Mayo for an answer.
It is just sad there is not a way to tap into the knowledge of all the experts who have worked on Model A's or other vehicles and could put that into a pool of knowledge that could be tapped via AI to give us answers to our problems. That is what is nice about this forum you can find experts.

[Hitman]

Originally Posted by jrelliott

I helped develop some of the computers AI is using. One of the applications which we worked in conjunction with Mayo clinic was in the area of diagnostics for patients. The doctor can input all the symptoms and the computer will look through all the available information stored in books, journals etc. and try to give the doctor the best educated guess. This is just one application where it is really helpful. As more queries are made and more accurate diagnostics are achieved, then it becomes a valuable tool. If I had a unusual ailment, would want to go to Mayo for an answer.
It is just sad there is not a way to tap into the knowledge of all the experts who have worked on Model A's or other vehicles and could put that into a pool of knowledge that could be tapped via AI to give us answers to our problems. That is what is nice about this forum you can find experts.

Your points are valid. Most comments here relate to questions being asked of AI, much like they are of google. Your results will vary based on the inputs!

AI is impressive when you play behind the scenes with bots and other things.

Most here are asking 2023 AI questions relating to very detailed information from nearly 100 years ago and expect it to differentiate between adjusting valves on a 1931 V16 Cadillac and a 1931 Ford. AI can’t distinguish between the two, and neither can any forum participant!

Don’t forget, ChatGPT hasn’t been updated in six months.

[Jerry in Shasta]

Sponsored Links (Register now to hide all advertisements)

Like the old saying GI-GO GARBAGE IN GARBAGE OUT

[David in San Antonio]

AI attempt to diagnose cancer from xrays picked 100% of examples with cancer but also a large number without. Turned out the pictures showing cancerous tumors all had a ruler in the picture to show the size of the tumor, and the AI selected all the pictures with rulers. This was corrected prior to further testing.
Another example: Very large financial customer service company AI monitored the thousands of daily phone calls for key words associated with sales. For good measure it also picked out offensive language. Within one week it showed an unbelievable count of clients and phone representatives discussing nudity! Turned out the calls were about "annuities" which wasn't in the stored vocabulary. Got fixed.

[alexiskai]

Quick update on my AI project now that Code Interpreter has been released to OpenAI Plus users. I had hoped that the Code Intepreter function would be able to parse PDFs of books and extract text, but so far I haven't been able to get it to do that. However, GPT-4 is able to write Python scripts for me that will do that on my workstation.

So my first task has been to convert images of The Restorer magazine to text files. As you may know, MAFCA sells the entire back catalog of The Restorer from 1956-2006 on a flash drive, but it's all images – not searchable. I was able to get the AI to write a script that converts each issue of the magazine from images to a single text file. Of course, all of the photos are stripped out, but for the moment I'm just trying to create a corpus that will be suitable to train the AI.

My next phase will be to work through the service bulletins, which I also have as PDF already. Then on to other books, like the JS, the Les Andrews blue book, and the McRee engine rebuilding guide. These I'll have to scan myself.

Once I finish those, I'll start figuring out how to process what I have into a database to train the AI on.

All of this content will live on my workstation, so not even a shadow of copyright issues yet. We'll cross that bridge when we get there.

[P.S.]

Originally Posted by johnneilson

AI uses nothing more than the data published on the web.

John

Yikes! That means that AI is basing its conclusions on opinions written by people. Even worse, mostly people who drink that beer in blue cans.

[Karl Wescott]

To pick nits...


This conversation has been about "THE MODEL A". The assumption is Ford 1928-31 Model A. Or should it be Ford 1927-32 Model A? What about Ford 1903 Model A? Or what about the hundreds of other manufactures, automobile or otherwise, who had a Model A?


OK so I deliberately confused the question. An AI will be initially easily confused until it has enough information and relevant database to parse a relevant answer.


Hypothetically, Q: "How many blades on an original Model A"? A: Please clarify, are we talking about a Ford 1903 or 1928-31, a 1914 Frankenburner, or something else?


Now a risk I just took is some AI might really think there was such a thing as a 1914 Frankenburner... Learning to recognize satire is HARD.


The BEST solution would be for the clubs to embrace AI, but have a proprietary and protected and vetted database, such as alexiskai is attempting.

[Jim Huseby]

Honestly, this thread contains answers for me that I wasn't expecting. Like many of us, I like old cars because they take me back to the years of my innocence and what I think was a more honest and less complicated and deceptive era. Though I was an electronics technician in the Air Force, I am contemptuous of computers in automobiles and contemptuous of the present philosophy that I must alter my human logic to use my cell phone or computer. I am alive to be open to all possibilities for advancement, including AI, but determined to grow toward the fulfillment of knowledge and of heart and soul.
The bright hope I see is the human souls and intellects that contributed to this thread.
Your humanity, as continually demonstrated by sharing info and anecdotes, and your intelligence is a safe haven for me, though I am frequently challenged beyond my intellectual capacity. Many of us have been fordbarners for a long time and can discern between bonafide and bogus knowledge and personalities. On ford barn, I have encountered some of the most remarkable sources of real intelligence and character.

[David in San Antonio]

Jim, Thumbs up. “…Contemptuous of the philosophy that I must alter my human logic to use my cell phone or computer.” Well said. The current rash of touch screen controlled car dashboards has stopped me from replacing my 2016 car. In my career I was responsible for marketing and sales of highly technical investment products. Left to their own devices the software programmers, investment managers, actuaries, administrators and the rest of the home office would have made the products impossible to understand or use except by their insider peers. No client wants to learn a whole new vocabulary and decipher procedures just to earn the privilege of buying the product. I went to war against jargon and acronyms for example. Plain English, short sentences. The basic test was whether my elderly mother could understand it. Intelligent, interested, but not an industry insider.
KISS (keep it simple, stupid) was a hackneyed cliche, but seems to have been forgotten.

[jb-ob]

Jim & Dave, 2X, jb

[mrraford]

Wait for the Terminator to arrive and ask him.... after you loan him a pair of pants......

[alexiskai]

Initial tests looking good! This was a test run using only a couple of sources.