Chat and Me

 My initial foray into ChatGPT has resulted in mixed success. (Please note: I'm attempting to spin it positive.) 

We all know, I'm resistant to technology anyway, so the failure came as no shock to me. But I refuse to say "ITYS." Although I'm really tempted.

In my defense, I did attempt to use ChatGPT. And I've read a couple AI summaries. That, in and of itself, is a HUGE step. 

I can even see some value in AI for summarizing LLMs (Large Language Models) by noting patterns--although this time it failed miserably with nuance.

However, while the place it took me was not expected, the failure did not surprise me. Let's revisit my foray...

Chat: What would you like to know? 

Me: Show me easy fused glass projects. (Show. Not tell me about, or explain.)

Chat: [describes 4 techniques. No photos (which is what I was looking for). I've completed several fairly intricate beginner projects. I understand the basic process so I almost skipped this. However, I skimmed through and found #3 (Sgraffito) to be interesting: by the AI description it sounded like the grade school art project where you paint or magic marker colors on the page, cover your dried page with heavy black crayon, then scrape off line drawings to reveal colorful drawings etched into the black background. I wanted to see an example to make sure that was what chat meant.]

So, when chat asked...

Chat: Do you want to see a demonstration or example? 

Me: Yes, please. (Assuming he knew what he was asking me about.)

Chat: please clarify what you mean. 

Me: Show me examples of Sgraffito. (I wasn't this direct the first time because this is "chat"--when chatting with a real person, IRL, the person would understand I was referring to the last subject discussed--in this case, glass fused Sgraffito. AI did not react like a real person would. In my defense, I at least didn't say "show me #3." Honestly I didn't want to say "sgraffito" because I hoped to avoid trying to spell sgraffito.) 

Chat: A solo trip to Japan...

Me: *exasperated sigh*. That's not what I asked. 

Chat: Please clarify. 

Me: Please show me examples of sgraffito. (I should have specified fused glass sgraffito, but by now I wanted ChatGPT to fail. Just so I could say it failed. 🤭)

Chat: [proceeds to give degrees at which pottery is "baked." LOL. You and I know pottery is "fired." Evidently AI is not as smart are we are. 😉]

Me: No. Show me images of sgraffito (now I'm not about to specify "fused glass sgraffito" because I'm spiteful). 

He showed me pottery examples of sgraffito that honestly looked as if the designs could have simply been painted on the pottery. I was under impressed with his choices.

Conclusion: 

Just like. Yogi Bear is "smarter than the average bear," my fallible human friends and I appear to be much smarter than the average AI, at least on ChapGPT. At least my friends understand, and keep up with, my conversational thinking much better than ChatGPT

And I even showed ChatGPT more mercy than I show most humans--I didn't "squirrel" mid-thought and expect him to keep up. In other words I stayed on track ,

LOL. If I had done that, I might have broken him! 

Hummm. Breaking Chat might not be such a bad idea...