Mr. Anzu would like to remind any college students stupid enough to use AI to write their essays that 1. The AI essay generator is lying to you, the professors CAN usually figure it out (and their software will give them a big warning that it may be AI generated to boot) 2. Don't be an idiot and simply turn it in AS-IS without going in and toss in in a typo here or there or deleting a word to make it more believable 3. Checks those sources because guess what, AI essay generators will cite themselves in the sources! You might want to delete that. 4. Read through the whole thing and verify the essay generator didn't hallucinate and start talking about a slightly different topic entirely.
Yeah he's going to have to have a difficult conversation with one of his students. Amazingly, only one so far this year. I guess the other ones have gone through and added the needed dumb mistakes and deleted the AI generator source listing so it's not so goddamn obvious.
Yeah, I deliberately ignore any A.I generated summary to a quick reference Google search (like a word definition or who is so-and-so) because it's so fucking unreliable. Hell, just this morning, Google proudly invented a brand-new word for me to "correct the spelling" of the perfectly acceptable (and correctly spelled) word that I had used. And AI chatbots are perfectly willing to make "facts" up in support of their statements when they're challenged by a human. We need to start using the term Artificial Stupidity more.
I honestly can't remember, it was this morning, and I already fixed it. It "corrected" an "un" prefix to an "in" prefix and wound up with something that looks like a real word but isn't, that's all I can remember.
Conducting 3rd party risk assessments for my company and almost every department wants some form of AI, whether it's ready for primetime or not. I always, always write in the security notes section that AI is not a replacement for a human, someone needs to check the work and make a decision. AI only has the information that it's been given with which to make a decision and it's usually incorrect.
You can use technology to do a lot of things without it necessarily being AI! Our web portal uses an API to grab stuff from our main app, display it in real time on the portal, and then pass back the info to the main app. Certain submissions can be auto-accepted if the parameters are correct. No AI necessary, just good old fashioned if/else statements in the code.
But about 90% of our cases are going to require human review, since so few of them are simple yes/no + here you go submissions.
I'm beginning to wonder if sufficient prompt engineering could give one unemployment benefits of $10,000 a month for however many years they keep the system in place ...
Government benefits is one area where you can pay absolutely no attention to the quality of your customer service, because where else they gonna go? Nevada could use Tarot cards (or dice? more appropriate) to sort through applications -- it would make about as much sense, and complaints can still be ignored.
Ob. soundtrack:
https://www.youtube.com/watch?v=th62Oo_exXE
Alan Sherman, "Automation"
Fantastic writing Marcie! Just fantastic. About a horrible subject.
How about we just let errybody have unemployment?
COMMMMMMMMMEEEEEEEESSSSSS!!!!
Why, that sounds like some kind of universal basic income!
Hmmmm....🧐
Ta, Marcie. Ai-ai-ai-ai-aye!
Allow me to introduce to the clusterf*ck that happened in Oz when that sort of thing was tried.
https://en.wikipedia.org/wiki/Robodebt_scheme
Mr. Anzu would like to remind any college students stupid enough to use AI to write their essays that 1. The AI essay generator is lying to you, the professors CAN usually figure it out (and their software will give them a big warning that it may be AI generated to boot) 2. Don't be an idiot and simply turn it in AS-IS without going in and toss in in a typo here or there or deleting a word to make it more believable 3. Checks those sources because guess what, AI essay generators will cite themselves in the sources! You might want to delete that. 4. Read through the whole thing and verify the essay generator didn't hallucinate and start talking about a slightly different topic entirely.
Yeah he's going to have to have a difficult conversation with one of his students. Amazingly, only one so far this year. I guess the other ones have gone through and added the needed dumb mistakes and deleted the AI generator source listing so it's not so goddamn obvious.
Nonsense! AI is undetectable! Everyone knows dogs are allowed to play professional basketball!
I too have seen the documentary Air Bud!
Yeah, I deliberately ignore any A.I generated summary to a quick reference Google search (like a word definition or who is so-and-so) because it's so fucking unreliable. Hell, just this morning, Google proudly invented a brand-new word for me to "correct the spelling" of the perfectly acceptable (and correctly spelled) word that I had used. And AI chatbots are perfectly willing to make "facts" up in support of their statements when they're challenged by a human. We need to start using the term Artificial Stupidity more.
Don't leave us hanging! What was the word?
I honestly can't remember, it was this morning, and I already fixed it. It "corrected" an "un" prefix to an "in" prefix and wound up with something that looks like a real word but isn't, that's all I can remember.
Probably "unpossible", which is a perfectly good word on this here dick joke and mommy blog.
Unpossible is too a word! Also, my spell check doesn't even blink at it.
Nicely prompted.
Do you mean YEATS?
Or did ChatGPT somehow discover a hitherto unknown but strikingly stylistically similar poet named YATES?
Rowdy Yates, the poet laureate of Missoula.
Sounds like you should sign up for some Old Glory Insurance!
Conducting 3rd party risk assessments for my company and almost every department wants some form of AI, whether it's ready for primetime or not. I always, always write in the security notes section that AI is not a replacement for a human, someone needs to check the work and make a decision. AI only has the information that it's been given with which to make a decision and it's usually incorrect.
You can use technology to do a lot of things without it necessarily being AI! Our web portal uses an API to grab stuff from our main app, display it in real time on the portal, and then pass back the info to the main app. Certain submissions can be auto-accepted if the parameters are correct. No AI necessary, just good old fashioned if/else statements in the code.
But about 90% of our cases are going to require human review, since so few of them are simple yes/no + here you go submissions.
Oh goodie! Should just use a roulette table.
Or a Magic 8-Ball.
AI is a total scam, the Lisa's Monorail of today. Go ahead and invest in that bubble, oh soon-to-be-shirtless ones!
Better than leaving it up to the GOP.
I'm beginning to wonder if sufficient prompt engineering could give one unemployment benefits of $10,000 a month for however many years they keep the system in place ...
Government benefits is one area where you can pay absolutely no attention to the quality of your customer service, because where else they gonna go? Nevada could use Tarot cards (or dice? more appropriate) to sort through applications -- it would make about as much sense, and complaints can still be ignored.