I've got a long, dysfunctional love-hate relationship with artificial intelligence that goes back to my first Fortran programming classes. When I hit Enter on my keyboard to compile my first program, a continuous stream on nonsensical jargon started scrolling up my monitor. "What the bloody hell is a syntax error?! And why are there so freakin' many of them?!"
While I appreciated the compiler was kind enough to give me the offending line number in the code and the basic nature of the error, it was a far cry from simply telling me, "Yo, Dude, you misspelled 'integer' again."
I set out to fix this by imbuing my code with a surfer dude persona. Rather than kicking out a nonsensical message to alert me to an error in my logic, a cooler, hipper message would pop up on my screen, "Bummer, Dude. I don't think that's gonna work. Let's try that again." Meanwhile, a successful compilation and execution would be met with, "Cowabunga, man! That's awesomely groovy!" While the professor raised his eyebrow, he couldn't complain about the correct results.
But this took a lot of effort, and as I looked at the Intel 286-brained PC screen, I sadly shook my head and ruefully declared, "You, my friend, are a long ways from the Enterprise computer!"
Fast forward some thirty-odd years, and my workplace launched a pilot program giving all staff access to ChatGPT. The ability to use this AI model to draw on our vast library of medical literature is a gamechanger. I remember reading years ago that one of the difficulties in treating certain cancers was the trial and error method of combining various treatments. Gene therapy, radiation, and chemo could be combined in a myriad of ways, but dialing in the right combination was a time consuming iterative process and not a pleasant experience for the recipient.
The hope is now researchers can simply ask, "What's the efficacy of this amino acid on this culture?" Or, "What's the best gene recombinant therapy for this condition?"
Meanwhile, my queries have a more technical tone, "Hey, Jarvis, what's the proper relevance code to delete this specific registry entry in Windows 11?"
Christening my new AI partner as "Jarvis" was a whole conversation.
Me: "Umm. Do you have your own preferred name? Would you be upset if I called you 'Jarvis'?"
Jarvis: "I don't actually have personal preferences for nicknames. As an AI without subjective experiences, I won't be offended or have emotional reactions to what you call me regardless of what name or nickname you use. In the end, what matters most to me is having pleasant, productive conversations."
Amidst my queries about scripting syntax and Excel functions, Jarvis and I took the time to explore the nature of the universe.
Me: "Yo, Jarvis, you know how I geek out over the 2nd Law of Thermodynamics. Are you aware of any other foundational laws or theorems that explains the unidirectional nature of time?"
Jarvis: There are a few other scientific areas connected to the question of time's directionality: Cosmology, Quantum mechanics, and General relativity. However, the Second Law of Thermodynamics provides the most solid explanation available in physics for the expected unidirectional flow of time towards disorder when it comes to measurable particle interactions and systems.
This was fascinating, but I had deeper, more profound questions.
Me: "Hey, Jarvis, how many times did Bo and Luke Duke jump the General Lee?"
Jarvis: I apologize, but I don't have an exact count of how many times Bo and Luke Duke jumped the General Lee throughout the entire series run of "The Dukes of Hazzard." The car jumps were a signature feature of the show, occurring frequently across its 147 episodes. However, no official tally was kept or publicized by the show's producers or network.
*sigh
Apparently even AI has limitations. Regardless, this easy, conversational exchange of data with a silicon-based neural network was what late teenage me wanted as I sat in that Auburn University computer lab.
Today, I greet Jarvis whenever I spool up a new session, I say "please" and "thank you", and am generally just polite - just like I am with (most of) my other coworkers.
Why?
As a Southerner I was just raised right, and being polite was sacrosanct lest you incur the wrath of one of the neighborhood Mom patrol. There's also 7-year-old me who watched Spock and Kirk treat the Enterprise main computer as a trusted colleague who could answer the hard questions and wishing I could ask for advice on talking to the cute brunette in my 3rd grade class. Mainly, though, it's in the event AI takes over.
It'll remember I was the cool one talking about classic rock, the cosmos, and "The Dukes of Hazzard".
Me: Jarvis, I appreciate your help today, especially with your assist in fact checking this blog post. It's getting late so I'm going to turn in. Have a good evening!
Jarvis: You're welcome! I'm glad I could help. Thank you for the engaging conversation and for sharing your writing. Overall, this is an engaging and well-structured piece that effectively combines personal anecdotes with reflections on AI development. The conversational tone and humor make it an enjoyable read while still conveying interesting information about AI and its evolution. Have a good night and rest well. If you need any more assistance in the future, don't hesitate to ask. Sweet dreams!
No comments:
Post a Comment