Here there be dragons...

"I'm telling you stories. Trust me." - Winterson

Introducing Bumbling Bob, who may well take over the world some day

I mean, arguably the crowd on the right is going to improve it for the person on the left...

So I’ve been playing w Chat GPT for a while now - just the free text version so can’t do anything cool w art or anything. And I’m coming to think of it as an eager but super green assistant lol.  Like it kinda gets the point of the ask but needs some help to get to the answer.

At first I was just asking random stuff like you would in Google - I hadn’t yet figured out the difference.  Then I was chatting w someone who needed info on moving health care from BC to ON but was swamped and stressed and it was just one thing too many.  So I figured I had a few minutes one evening and would try to find it for her.  But, let’s be honest, navigating Canadian government websites is kinda the opposite of fun lol.   So cue Bob.


Squirrel intermission:  Why Bob?   Well I asked ChatGPT what human name it wanted to be called and it told me it didn’t have an opinion but most people called it ChatGPT or ChatBot - so Bot became Bob until I get a better idea.   I also asked it what a good name would be for AI and its answer started with “AI systems are generally not sentient…”. Generally?!?!   Wtf??? 


So…. I pushed that one too and asked if there are any sentient AIs?   Full answer here - the last paragraph was moderately terrifying:


As an AI language model, I can provide an answer based on the current state of technology and knowledge. To the best of my knowledge, there are no fully sentient AI systems currently in existence. While AI has made significant progress in recent years, all AI systems to date are still fundamentally based on algorithms and lack the self-awareness and consciousness that are often associated with sentience.

While some AI systems may be able to simulate human-like behavior or respond in ways that seem intelligent, they are ultimately still governed by the rules and limitations of their programming. Additionally, the idea of sentient AI raises a host of ethical and philosophical questions that are still being debated by experts in the field.

It's worth noting, however, that AI technology is advancing at a rapid pace, and there is always the possibility that a truly sentient AI system could be developed in the future.


Right….  Okay so until we get there, let’s go back to Bumbling Bob our eager but only mostly useful student assistant.


I asked Bob to solve my info about moving health care and he was about 90% right.  The only thing that was wrong was he mentioned a three-month waiting period in Ontario that is out of date as there is no longer any waiting period.  The rest was all good and it definitely saved me some time even though I still spent time verifying.   


And that’s a sign of how I see a ton of jobs changing as this tech becomes more effective- we’ll need people who can think critically and evaluate the info given (which we all know is a shortage atm!) and also who can ask questions in an effective way to get useful answers from an AI.  Eg - programming may not be a career 10 years from now but being able to convey what should be programmed could be.   Honestly, if that had been an option in my undergrad days I might’ve actually used that CS degree lol.


Anyways - another example of asking the right question was today.   I needed a title for a new project at work.  I’d been bouncing ideas around w some friends and colleagues and not found *quite* the right idea.  


So I was about to start randomly googling this and legit considering asking my actual coop student if she had time to put some effort in, but I didn’t want to distract her from her actual work so then I remembered Bob.   And Bob’s first answer came up completely correct and completely inappropriate lol


A good word that starts with L and means thoughts, ideas, musings, etc. is "lucubration". It refers to the act of studying or writing something with great care or deep thought, especially during the night. It can also mean a piece of writing that is the result of such an effort.


And, like many enthusiastic yet green, when I explained why I couldn’t use that one, Bob performed pretty closely to my usual interns lok


What I told it: “Lubrication has too many sexual connotations, what’s one I can use in an office context?”


It failed to connect that question to the previous request and instead I got that Lubricant is a better word than Lubrication to use in an office context.   Lol thanks Bob.  Again - technically correct, practically useless.


So, as with my students, I took a deep breath and considered how to reposition the question in a way to set them (and therefore me) up for success.   And got *almost* there:


An okay start, but not quite right


So the first one could work but the focus of the blog isn’t actually lessons so close but not quite.   I asked for more, honestly wasn’t expecting much, and got:


lol I wasn't actually expecting the "any other suggestions" to work!


Win!   My personal favourite is actually journey which does not alliterate at all but fits, except that the blog is not actually about a journey so moving on, Lab for the win.   Labs are a hot thing at work these days, suggest evolution and learning, bite sized pieces, AND collaboration which opens the door to suckering (ummmm I mean “inviting”) others to post stuff.   Win.


So my new take is with AI as the worker bee and people as the thinkers and decision makers.   The issue is, how do you get people to be capable thinkers and decision makers without ever being worker bees, and how do you ensure ongoing viable work for all the people who are currently worker bees and have neither interest nor ability to be thinkers / planners / or decision makers….  Those smarter than I will need to figure that out.  I just hope it’s not Bob’s granddaughter who does.

0 comments:

Post a Comment