Proof that Border Patrol Agents Have No Sense of Humor

Southwest Assoc. of Buffalo Soldiers

It was sometime near the end of the 1990s. I’d finished meeting with the network-security folks at Ft. Huachuca (“Wah-CHOO-kah”) in Sierra Vista, AZ and checked off all the items on our to-do list.

I found myself with a free afternoon and an opportunity to see a part of the country I’d probably never visit again, so I went exploring. First stop: climbing the dried riverbed to Coronado Cave.

National Park Service

🙶You don’t need to worry much about rattlers, as long as you can see your hands and feet.🙷  — Ranger at the welcome center

Next, I drove up to Coronado Overlook to see the spectacular views all the way into Mexico. Then, for some reason, I decided to drive the only road south, toward the border.

Phoenix Magazine

 

There were no other cars on the road for miles. At some point I realized that route might not have been the cleverest decision, so I turned around and headed back toward my hotel. Eventually I saw a car on the side of the road in the distance, lights flashing. An official-looking gentleman stepped out as I approached. He motioned for me to pull over. A border-patrol agent.

He asked what I was doing. I told him I’d had business at the Fort and now was just being a tourist. He wanted to know if there was anything in the trunk. “I don’t know; it’s a rental and I haven’t looked. Do you want to check?” He did, and of course it was empty. He asked for the paperwork as a final check. I commented as I handed it to him, “Wow, you guys are really serious out here.”

He raised an eyebrow. “What do you mean?”

“Well, back east, our idea of border security is just keeping out the folks from Pennsylvania.”

I don’t think he even cracked a smile.

He shook his head, pointed at the road ahead, and said, “Go. Just...go.”

Is it plagiarism or aLLMost plagiarism?

From a discussion on a Linkein post by Aron Brand:

The question arose as to whether LLMs store "representations" of their training data, and if so, why is it not plagiarism to use those representations as they respond to users' prompts?

I think that's a very nuanced and insightful question that comes down, I suppose, to the definition of "stored representations."

At first blush, it seems obvious: Let's say an LLM ingests the post you're reading at this moment. For the sake of argument, assume it's not here on a blog, but in a book that I've published and copyrighted. Of course I've included the standard notice that "no part may be stored, transmitted, reproduced, etc. without written permission." The owner of the LLM has bought and owns a copy of my book.

Later, you ask it about my opinions on LLMs and plagiarism, and it summarizes what I've written. I allege that it has "stolen" the content and used it unlawfully, without my permission.

Has the LLM "stored" my content?

An AI Apology

I'd been working late into the night on part of a large personal project. It was progressing slowly but steadily since early morning and I was learning tons. (Translation: I was spending more time fixing than creating, trying various approaches to accomplish each step, and discovering clever new ways to make mistakes. You know, the kind where you don't know whether it'll be harder to fix it, or to figure out why it didn't work in the first place.) To be fair, I don't have a lot of experience in this particular area, so I was exploring alternatives and poking at the options that were available. Given the luxury of time, I've found this strategy to provide a far deeper understanding than just getting it right the first time and moving on. 

 

Are You More Creative Than an LLM?

Edited from my comment on a LinkedIn post by Roger Dooley, who I don't know.

Roger wrote about an article in "Nature" that describes Centaur , an LLM that predicts human behavior with uncanny accuracy. It reminded me of another project that's similar, but in an opposite way.

A team of researchers created an interesting experiment that tries to measure one aspect of creativity. From www.datcreativity.com/about:

🙶The Divergent Association Task is a quick measure of verbal creativity and divergent thinking, the ability to generate diverse solutions to open-ended problems.🙷

So it's sort of like the project Roger described, but in reverse: Instead of asking an LLM to predict human decisions, in a sense it analyzes our ability to be unpredictable. It asks for 10 common nouns, chooses the first 7 that it deems 🙶acceptable,🙷 and estimates how different they are from each other (emphasis added):

🙶The average score is 78, and most people score between 74 and 82. The lowest score was 24 and the highest was 96 in our published sample. Although the scores can theoretically range from 0 to 200, in practice they range from 6 to around 110 after millions of responses online.🙷

For fun, I thought it might be interesting to compare a human's results to a few LLMs'. Guess who won? 😀️

  🤖 Gemini scored 67.98, higher than 5.90% of the people who completed the task.

🤖 ChatGPT scored 79.1, higher than 54.23% of the people who completed the task.

  🧑 I scored 86.8, higher than 91.52% of the people who completed the task.

Given this highly scientific (ha!) sample size of 1, it seems we humans are still better than LLMs at being human. If you choose to take the test, I'd love to know your score—please post it in the comments!