I'm sure you didnt mean any harm but AI including Google's AI is not always factually correct, even if it cites its sources. Tbh, wikipedia is better than the AI explanation. Just be careful!! ^^
Never used a LLM, ever. I don't see the point. They just churn out gibberish in accordance to what it *thinks* sounds right. I've seen too much made up nonsense to trust them.
If it's just chat gpt, probably not uncommon. But if it's any AI, then a lot of people in the early days of AI, before they realized just how bad it is, used it because it was novel and funny just how ridiculous the responses were. So for that, a bit more uncommon than most realize.
I admit, I refuse to touch ChatGPT or other similar services for fear that it'll give me the wrong information. Too badly burned out by finding sites with the wrong info, so I've been hesitant
I use Copilot specifically for training myself. If it tells me a solution that doesn't work, I still learn how to do a function or use a feature that I might've not known how to use otherwise.
Good example is excel. Learned a lot by letting Copilot teach me.
And the best way I learn is with conversation. LLM is beside the point; it's a knowledgeable resource that has been built in a way that allows it to break things down into simple steps.
Perfect for my personal needs without needing to pay for a personal consultant.
It wouldn't hurt to be more sociable.
"Conversation" usually happens with people. And "personal consultant" is the last thing one would think of when learning a common skill.
You can speak for yourself, of course. Whatever works for you when learning. I'm learning the best this way; spot-on knowledge, can do it myself and poke at it from the give point forward. It just works.
And I aint paying for in-person courses in this economy.
You're taking the right approach. ChatGPT is trained to write things that look like the response you would get, not to actually give you information. Use actual sources, and evaluate them critically.
I had to fucking correct ChatGPT the moment I find out they wrote something wrong and it's just frustrating.
Ironically, the system that corpos want to use to "cut corners" (read: stop paying/treating creators well) end up being WAY more time-consuming (and energy wasting).
Today: man + woman looking at plants. Man says "i wouldnt know where to put them" woman says " I'll take some pictures and run them through chatGPT for ideas"
I wonder to myself, how I could have become a god if only I was not tempted by the fruit of curiosity. I fell on the sword when I searched up synonyms for elegance.
Domain expansion...... I touch grass, when activated it forces every sweat in a 300 mile radius to be so disgusted that their actively repulled away and fall like a goober
The only time I ever used an LLM was to run an essay through trying to figure out if it was written by chatgpt. Looking back it was kinda obvious bc of how much everything was jumbled together and repetitive in it.
π€¦ββοΈ Sorry, I didn't quite understand the verb "redraw."
I assume the original (the smoking issue) is yours.
The reason for my question is the signature. I wanted to suggest you sign it so I can share it without losing the credit, link, and whatever else you want.
Tbh I felt conflicted abt signage cause the redraw for the meme is fairly close to the original source material. My sona is enough of a signature for me but yeah I did redraw the entire thing by hand.
augh... the damn line between plagiarism and derivative work.
Hmm, let's see. I think the concept of a derivative work as a meme is clear, and on the other hand, you probably wouldn't have told me the way you did.
+
What about something like "Original: [?]; redrawn and reinterpreted by ToriPng"? "Original: [?]; redraw & reconcept by ToriPng"?
(I don't know how to say it better in english and country.)
I think for native English speakers the original caption makes clear enough that the redraw was done by the poster, and the lack of signature indicates that they don't particularly want explicit credit.
I recall complaining on a forum that I didn't have time to properly plan meals or a shopping list, and someone came months later and said to use chat gpt. Like I'd literally rather eat pasta the rest of my life than let a robot decide what I should make.
When AI was first becoming a thing, I thought it was neat and didnt realize how damaging it would become, so I did use things like characterai and at one point an image generator, but I am proud to say I've never touched chatgpt
Same with my writing friends, actually. They got excited about Midjourney when it first came out, then started to realize the baggage that came with it and dropped it almost immediately.
I used it a couple times for my physics homework back in high school before i even rlly knew what AI was, it got the answers wrong like half the time π
I've tested it to see what all the fuss was about. Any time it tried math you had a 50/50 chance of being given the right equation. Never actually used it to *do* something.
All it does is produce copious amounts of bullshit and people all over the internet are smearing its shit across the walls.
I can also proclaim this but I wish I could be able to say I haven't used AI at all, unfortunately I was actually a pioneer in ai by having extensively played AI Dungeon in the past, before AI "got good".
Comments
Good example is excel. Learned a lot by letting Copilot teach me.
Like, people have been learning to use excel for decades before the LLMs.
Perfect for my personal needs without needing to pay for a personal consultant.
"Conversation" usually happens with people. And "personal consultant" is the last thing one would think of when learning a common skill.
And I aint paying for in-person courses in this economy.
So instead of a system helping people, us people have to outsmart it and put more effort into writing.
It can generate a paragraph, but when you read it something does not add up and it just gives the illusion of "informative".
Ironically, the system that corpos want to use to "cut corners" (read: stop paying/treating creators well) end up being WAY more time-consuming (and energy wasting).
Me: π
My shop is open rn and I have no cons this month so I'm linking it here:
https://shop.toripng.com/
π€
It's yours?
I assume the original (the smoking issue) is yours.
The reason for my question is the signature. I wanted to suggest you sign it so I can share it without losing the credit, link, and whatever else you want.
Hmm, let's see. I think the concept of a derivative work as a meme is clear, and on the other hand, you probably wouldn't have told me the way you did.
+
(I don't know how to say it better in english and country.)
Not like i haven't at all, I've got the lore but stuff like lil docx stories and such i gotta come up with for once
It took me quite a bit to realize people don't accept justifications of any kind, but yea, sounds fair
As a joke, i asked where i could hide a dead body.
It answered. Also, iβm the 69th comment.
All it does is produce copious amounts of bullshit and people all over the internet are smearing its shit across the walls.