SomeGuy
Hobbyist
The problem is it's typically people telling the AI to generate the content, maybe doing one skim read, then posting it as their own work...
Though YouTube has got something close they make me answer on each video I upload:
-
Scam Alert. Members are reminded to NOT send money to buy anything. Don't buy things remote and have it shipped - go get it yourself, pay in person, and take your equipment with you. Scammers have burned people on this forum. Urgency, secrecy, excuses, selling for friend, newish members, FUD, are RED FLAGS. A video conference call is not adequate assurance. Face to face interactions are required. Please report suspicions to the forum admins. Stay Safe - anyone can get scammed.
-
Several Regions have held meetups already, but others are being planned or are evaluating the interest. The Calgary Area Meetup is set for Saturday July 12th at 10am. The signup thread is here! Arbutus has also explored interest in a Fraser Valley meetup but it seems members either missed his thread or had other plans. Let him know if you are interested in a meetup later in the year by posting here! Slowpoke is trying to pull together an Ottawa area meetup later this summer. No date has been selected yet, so let him know if you are interested here! We are not aware of any other meetups being planned this year. If you are interested in doing something in your area, let everyone know and make it happen! Meetups are a great way to make new machining friends and get hands on help in your area. Don’t be shy, sign up and come, or plan your own meetup!
Anyone using ChatGTP?
- Thread starter Mcgyver
- Start date
The problem is it's typically people telling the AI to generate the content, maybe doing one skim read, then posting it as their own work...
Though YouTube has got something close they make me answer on each video I upload:
Pretty good actually! I'd sign off on that totally.
I confess that I'm a bit surprised to be targeted though..... Don't others have my sense of humour? @140mower comes pretty close if you ask me. I'm usually smiling before I even read his posts.
So I tried on chatGPT "make a picture of an old white man with a white beards sitting in a rocking chair on a ranch style house porch holding a big fly swater" and got....absolutely nothing! It kept claiming it was busy, for about 2 days, then finally said your image is ready and here it is! and showed nothing at all. I told it nothing was there and it said I'll regenerate it. same results. maybe because I'm using Safari? I just tried again using gemini. it said I don't make images yet. I asked some questions and got a bunch of material back and that seems useful if the answers are correct.
Chatgpt now says "I can’t generate images right now."
Chatgpt now says "I can’t generate images right now."
Last edited:
This one made me laugh ....I posted recently a 3D printed saw rack using rounds bits as a lock. When thinking about it, I was search around for ideas and asked it "drawings please for a dowel based hand saw rack"
This is what I got. So yeah, its slick, but a long way from perfect
This is what I got. So yeah, its slick, but a long way from perfect
can we replace fly swatter with scattergun? because that's my nursing home goal
get off my lawn!! perfect!
I'm going to have a hard time getting anything done tomorrow at work. I asked for machinists dirty limmerick....
His Monarch 10EE ran tight,
With a bed that would gleam in the light.
He’d chase every thread,
With precision in bed,
And swear she purred louder at night.
He’d fine-feed with passion and flair,
While the tailstock just begged for some care.
He’d retract with a grin,
Then go plunging back in,
“This isn’t just turning — it’s rare air!”
His Monarch 10EE ran tight,
With a bed that would gleam in the light.
He’d chase every thread,
With precision in bed,
And swear she purred louder at night.
He’d fine-feed with passion and flair,
While the tailstock just begged for some care.
He’d retract with a grin,
Then go plunging back in,
“This isn’t just turning — it’s rare air!”
As far as the art stuff is concerned, I find AI remotely interesting.
As far as answering questions is concerned, I have little to no interest. It simply cannot be trusted. If I have to hard check every answer against reference material to see if it's correct, I may as well do that in the first place.
As far as having a discussion, debate, argument? Nope. Language models "read" text & form answers based on what they find. If what they absorb has a particular ideological bias or political slant then that will (and does) appear in the responses. I find arguing w/ ill-informed, ignorant idiots in real life frustrating enough as it is. I don't need it from an LLM, too.
As far as answering questions is concerned, I have little to no interest. It simply cannot be trusted. If I have to hard check every answer against reference material to see if it's correct, I may as well do that in the first place.
As far as having a discussion, debate, argument? Nope. Language models "read" text & form answers based on what they find. If what they absorb has a particular ideological bias or political slant then that will (and does) appear in the responses. I find arguing w/ ill-informed, ignorant idiots in real life frustrating enough as it is. I don't need it from an LLM, too.
I am currently part of a project creating a chatbot to assist technicians in mechanical, electrical , electronic and computer troubleshooting, maintenance and associated procedures on machines that I share responsibility for and ultimately will partially replace me when I retire.
Still in the discovery phase as they call it but the prompts and resulting answer can be challenging and look to be the bulk of the work for us, knowing all the answers and understanding the question puts me way ahead at the moment but 24/7 access without calling someone is a definite advantage for it as it gets built. We are trying to use all of our documentation, videos, schematics, procedures but I am not sure if it will all work together. The concern I have is that they will not receive any formal training and try and rely solely on it but in the end a steady flow of pension cheques will ease my mind. The whole thing will be only accessible internally and will be the flagship for future opportunities within different departments in the company.
Still in the discovery phase as they call it but the prompts and resulting answer can be challenging and look to be the bulk of the work for us, knowing all the answers and understanding the question puts me way ahead at the moment but 24/7 access without calling someone is a definite advantage for it as it gets built. We are trying to use all of our documentation, videos, schematics, procedures but I am not sure if it will all work together. The concern I have is that they will not receive any formal training and try and rely solely on it but in the end a steady flow of pension cheques will ease my mind. The whole thing will be only accessible internally and will be the flagship for future opportunities within different departments in the company.
Oh Mothers of Gods!
How come in my imagination, I see the mother of all life-sucking, Soul-draining offspring of the voice-mail hell 'AI" "Assistants, that so many of the large companies have before you can actually talk to a live human?
I spent enough of my life correcting Aircraft Maintenance Manuals, of 50-plus years old airframes, of misprints, typo's, or simply bad information, I kinda feel how that is going to play out longer term...
Suggest retirement at a far enough distance from the worksite, that the guy that finds your name on the parts of the chatbot that just made him scream and tear his hair out, has a chance to cool down, if the self driving Uber, can actually find your house, eh?
SomeGuy
Hobbyist
Oh Mothers of Gods!
How come in my imagination, I see the mother of all life-sucking, Soul-draining offspring of the voice-mail hell 'AI" "Assistants, that so many of the large companies have before you can actually talk to a live human?
I spent enough of my life correcting Aircraft Maintenance Manuals, of 50-plus years old airframes, of misprints, typo's, or simply bad information, I kinda feel how that is going to play out longer term...
Suggest retirement at a far enough distance from the worksite, that the guy that finds your name on the parts of the chatbot that just made him scream and tear his hair out, has a chance to cool down, if the self driving Uber, can actually find your house, eh?
This is going to be the problem, people are going to lose the ability to create and correct documentation, and the AI won't know any better than what it is fed...it's eventually going to result in a huge loss of knowledge and increase in stupidity and we might not be able to find our way out of it. People worry about AI getting smarter than us, but it might actually be it results in all of us getting dumber.
You don't think we're there already, before AI? For example, what percentage learn machining now on you tube vs buying a book? Zero QA on you tube. With a book, the author wouldn't be considered by the publisher if he didn't have credentials and then everything is studied and vetted before release. What percentage of people seek out a the truth from credible authors and authorities vs finding someone espousing what already suits their biases? Idiocracy arrived before there was AI imo.
Nothing's changed in the sense that it's still a responsibility to verify and figure out who's/what is credible enough to trust.
I've been using it this week for a bunch software and IT stuff. Parsing old contact manage systems (program long, just have a database), setting up docker, setting up SQL recovering old data, writing python scrips to parse and integrating it into a new open source program via api etc. I know it's credible because it works. While there are guys here who could do it in their sleep, I'm not one of them so it's take what was impossible (give time constraints) and made it happen. A real game changer for me.
Last edited:
SomeGuy
Hobbyist
Fair enough...though the difference with these AI things is that it can string together complete nonsense, it's effectively just a bunch of probabilities about what word comes next. At least with a dumb person there's still a coherant thought they are trying to convey (even if it's wrong) and typically you'll have a bunch of people commenting on videos if someone completely screws up, crowd sourcing that QA that might even be more effective than peer review. With AI there's nothing to validate the answer it provides.
Your example with software and what you're working on...what happens when the current generation of developers can only develop with AI tools, they can't do it from scratch or MSDN or whatever (like I do)...who is feeding into the AI then? Does the AI stop getting any better? And if people are so reliant on the AI to be able to do anything to begin with, do we start regressing in our intelligence?
I don't disagree with you, just think we were already screwed and that AI (A1 as the head of US education calls it lol) has its good points. Jimmy Kimmel went around handing out bottles of A1 telling kids it was to help them in school lol.
Education Secretary Linda McMahon confuses AI with A1, sauce brand capitalizes on blunder
Secretary of Education Linda McMahon confused AI with A1 while speaking during the ASU+GSV Summit, and the sauce brand capitalized on the moment.
www.usatoday.com
100%.
Instead of asking a neighbour, an online forum, a local expert...just any person for help and advice.... folks are turning towards AI slop machines. This tech accelerates our already fragmented society of individualized consumption further.
Too "busy" to call your grandparents to check up on them? Just use AI:
Introducing: Joy Calls - AI Phone Calls to Support Your Aging Parents
AI Companion for Seniors that can make several calls per day, reducing loneliness and increasing peace of mind for family members concerned about their aging parents.
onscreeninc.com
AI erases the labour that produced all this easy access "knowledge". Who gets to profit from this? Definitely not the novelists, artists, engineers, machinists, mechanics, etc.
It would probably be faster asking AI my simple metalworking questions than to sort through everyone's responses here... heck AI has probably scrubbed all the posts here. I should just ask ChatGPT to look through this forum and find my answer instead of reading this place myself or starting posts... but I'm intentionally not doing this.
I opened up ChatGPT once just to know what it's all about. I've never gone back since.