I asked it to explain global warming without using the letter T, but it somehow just couldn't understand that it's replies weren't allowed to have the letter T in them.
That post seems to be missing.
GreyTechnician 12 points 3 years ago path: 0 5273621 5273661, hotness: undefined, score: 12, children: 0
kakes 40 points 3 years ago
My favorite use was to create a Discord bot where you could say any card title for a made-up MtG Card, and it generates the card for you, using DALL-E for the card image.
For example, here's one of my personal favorites using the input "Barney the Dinosaur":
dQw4w9WgXcQ 25 points 3 years ago
Since the bottom text is not in cursive, you and your opponent are actually required by the rules to love eachother and eat leaves.
detalferous 10 points 3 years ago
Italic?
dQw4w9WgXcQ 16 points 3 years ago
Ah, yes. It's called "kursiv" in my native language, hence my mistake.
detalferous 3 points 3 years ago
Cool fact! Thanks!
LowtierComputer 2 points 3 years ago
Which language?
intensely_human 7 points 3 years ago
1/1 with trample 🤣
ThisIsAManWhoKnowsHowToGling 3 points 3 years ago path: 0 5272233 5285448 5296619, hotness: undefined, score: 3, children: 0
kakes 2 points 3 years ago
See this comment.
Tl;dr, I'll let y'all know when I'm able to share the code.
ExperimentalGuy 1 point 3 years ago
Omg do u have a repou can share for that it sounds like so much fun
kakes 2 points 3 years ago
Unfortunately, I was lazy, and I'm webscraping a free MtG card generator to create the cards. I don't really want to put that code out into the wild and cause problems for the host of that service.
If I can find time, I'll see if I can figure out a more ethical alternative. When I do, I'll ping you!
ExperimentalGuy 1 point 3 years ago
Omg thanks I'd really appreciate that
Aurenkin 30 points 3 years ago
With the right prompt GPT4 can make a decent DM if you want to play a solo story of DnD. I tried this quite a while ago and it was pretty good, probably better now with the bigger context window.
It can also be useful for suggesting recipes based on what you have on hand or what you feel like eating.
ArmoredThirteen 12 points 3 years ago
I use it to help me DM. A lot of times I have a vague idea and it helps spit out specifics, I can take parts I like, and build the final idea off that. It is also okayish at making stat blocks for me on the fly when someone randomly decides they want to kill something I didn't plan for (so like basically every session)
flux 5 points 3 years ago
It's great at coming up with ideas for magic items, monsters, traps, plot hooks, etc.
I ran a Christmas session last year, it was a dream session so I handed out tons of magic items cause they wouldn't be keeping them. ChatGPT generated every single one, and I didn't tweak anything at all. Which made for a few interesting items with redundant or contradictory stats, but that was explained by it being a dream.
I have used it many times to give ideas for new items, new monsters, etc. Recently I was creating a blood-themed dungeon and asked for monster ideas and it gave me an Animated Bloodpuddle (reflavored Gelatinous Cube with extra abilities), and a Hemogoblin (ravenous blood thirsty goblins). There were others but those ones stuck out as cool and ended up being quite fun for the party.
ArmoredThirteen 2 points 3 years ago
Hemogoblin is one of the best things I've heard that's delightful
Rivalarrival 29 points 3 years ago
With PiHole, you can block it from your network. That's pretty cool.
MomoTimeToDie 22 points 3 years ago
I choose to interpret this as you having chatgpt add incomprehensible comments just to make code more confusing
MrSlicer 1 point 3 years ago
It codes pretty well for what I need. It helped me write a program for my esp32 to fill out a Google form to populate a Google sheet with temperature humidity and barometric pressure.
Its also helpful me with a visual basic program that checks on 3 temp/humidity sensors in my home. Plots it on a graph and turns on my led lights to a blue and white animation if the temperature falls below the trigger temp. This warns me if it's close to freezing in my garage. I'm doing my best and trying hard to learn.. Ai has been fundamental in me avoiding hours of googling because I don't quite understand the syntax yet.
Here's the program in action let me know what you think. It's the first real program I've made. https://youtu.be/T7RhQKH-b70?si=fTgq4_D2w33rGM_2
bitsplease 9 points 3 years ago
I'm terrified for our next junior dev hire if this is common among beginners now lol
bitsplease 2 points 3 years ago
How would you know? You don't actually understand the code. It could be making shit up, and you'd never have any idea
Seriously dude, if you run into code you don't understand, take the time to actually understand it, or else you'll never actually improve
state_electrician 2 points 3 years ago
If you do this at your job, get written permission to paste internal code into another company's data collection engine.
MrSlicer 1 point 3 years ago
Doesn't copilot do this? I'm just learning now but it's so much easier to code even with Bing Ai chat it helps me understand error messages and all sorts of stuff. What do professionals use?
NOT_RICK 21 points 3 years ago
A friend at work had it rewrite the lyrics of Apple Bottom Jeans in the style of Soviet propaganda. One of the lines came out to “shawty is a true socialist hero”. Having it rewrite lyrics in an incongruous style never gets old.
chemical_cutthroat 17 points 3 years ago
I use it to get ideas and run numbers for my homebrew Pathfinder campaign. That and midjouney and I feel like a real GM.
Moonguide 5 points 3 years ago
Same. I'm a new DM (and new to dnd in general, DM'd one one-shot and that's it, working on a campaign rn) and I use bard to make my narrative ideas fit the mechanics, name homebrew stuff, and clean up the text I'm writing for the players handbook I'm making for this campaign.
FaceDeer 7 points 3 years ago
I'm a long-time DM and I find ChatGPT awesome for this sort of thing too. I call it my "brainstorming buddy." Great for bouncing ideas off of, suggesting names for stuff, and so forth.
I've found that it's not very good at game mechanics, though, so don't rely too heavily on it for balancing stuff.
Moonguide 1 point 3 years ago
Oh yeah, I mostly use it to bounce ideas. If it suggest something good, I modify it to make sure if works like I need it to work. Used it to make simple tables and such.
Last thing I used it for was a battle simulation mechanic, in order to have the same effect a huge battle would have, without having tons of different characters to bog down the combat. Havent tested it yet, but it feels like it should work.
chemical_cutthroat 1 point 3 years ago
Oh I totally agree. I've been DMing since ad&d. This just makes my life so much easier. I can generate a hundred NPCs with basic back stories to fill a town and let my players run wild. It's super nice as a tool. I definitely wouldn't suggest it as a full replacement, but I am thinking about doing one shot where I let gpt run the game and tell it what my players do. Might be fun.
Piecemakers3Dprints 4 points 3 years ago
You might want to look into Ironsworn 🤘🏼
shapesandstuff 10 points 3 years ago
The biggest issue imo is that due to the way it works, it never just says "i dont know, i dont have information on that topic". Instead it just makes something up.
DrQuint 11 points 3 years ago
Except it refuses to aknowledge wrong knowledge or impossible methods. Until you tell it so, at which point it will agree with you even if you are also objectively wrong to tell it that.
I've coherced some systems into doing things they were absolutely not initially meant to do. ChatGTP would straight up tell me wrong and misleading ways to achieve them, and would be too ignorant to realize with enough stubbornness and with existing obscure info and the right mindset, a correct one existed. And at no point, not even with prodding and handholding, would it be able to have a useful conversation surrounding the why's of the current state of art.
It's not an expert system and it's stupid as fuck to treat it as one.
Fizz 2 points 3 years ago
I'm not asking what the answer is. I'm asking why things are done and the answer gives me a few different reasons why it is done. It's very helpful. The answers all seem reasonable and I've not experienced any dumb answers so far.
redballooon 5 points 3 years ago
Don't use GPT-3.5 for that. It'll halluzinate pretty quickly. GPT-4 is much better, but it's still wise to double-check what it says before you make life decisions.
Spuddlesv2 1 point 3 years ago
So long as you’re happy potentially getting the wrong answer over and over again.
DillyDaily 1 point 3 years ago
I think it depends what you're trying to learn - I'm slowly teaching myself to use Excel beyond the highschool level understanding I have. It's easy enough to fact-check ChatGPT because the formula either works or it doesn't. And I'm not to fussed if everything I learn to do is a total bodge job, because it's just for my personal development, it's not something I need for work or doing any serious spread sheeting.
Piecemakers3Dprints 4 points 3 years ago
NGL, that drink seems overly sweet and needlessly simplistic (add gin or vodka to the "martini", FFS), but... I guess?
nottelling 4 points 3 years ago path: 0 5271377 5272025 5272219, hotness: undefined, score: 4, children: 0
Bananigans 13 points 3 years ago
Whenever I'm doing public speaking, I just load my speech into it to rewrite with a specific fog index to make sure everything comes across in an easily digestible format. Similarly, for things that are too dense for me to understand from subjects I'm not well versed in, I'll load those lectures into it to make it understandable for me.
toototabon 5 points 3 years ago
TIL about fog index, thanks! What's the last subject you spoke about?
scottrepreneur 4 points 3 years ago
And SQL queries!
GreenMario 8 points 3 years ago
I use it to make alternative reality art.
Shit like that. Weird crossovers. Very specific memes that only apply to a single person just to fuck with em.
FaceDeer 7 points 3 years ago
I'm a fan of the SCP Foundation, and sometimes when I'm bored I'll go to ChatGPT and tell it "Please write an SCP entry for an anomalous \ that does ." If the resulting SCP entry is interesting I will "explore" it by asking "Write an exploration log in which an MTF team climbs down the giant toilet bowl" or "Write an interview log with the woman who turned into a snail." And so forth.
room_raccoon 4 points 3 years ago
What happened to snail lady? Is she okay now? Is she still a snail? I don't want to become a snail
FaceDeer 1 point 3 years ago
She's actually doing well. I dug up what I saved from that particular run. She's a 1.5-meter-tall land snail named Alice Johnson, with a pink and lavender shell that has lacy patterns on it, a sharp retractable proboscis, and glands that secrete a bioluminescent substance with extreme healing properties. She got transformed into that when she tried converting a secluded, abandoned building in the woods into an art studio for herself, it turned out to have been used by a researcher doing illicit anomalous work and some goop had been left behind. She initially was pretty distraught by the transformation, she attempted suicide several times before the Foundation found her and only survived because of that bioluminescent healing substance rapidly fixing any injuries she inflicted. But she's received counselling and is doing much better now in Foundation custody. When she found out her secretions could heal others, not just herself, she actually started to see some positives in her transformation. The Foundation has also provided her with art supplies to continue her hobby and she's taken to using her own shell as a canvas to express herself.
room_raccoon 1 point 3 years ago
This is fascinating. How does she paint her art? Healing folks with my own secretions is for the birds. I still wouldn't want to become a giant snail
FaceDeer 1 point 3 years ago
This is the sort of thing that I find most fun about these ChatGPT SCPs, you can "investigate" these details by asking whatever questions you want. In Alice's case I didn't ask, and I don't have the SCP in ChatGPT any more. One sec, I've got a local LLM running, I'll throw this information in there and ask it.
I got:
As a giant snail, Alice Johnson's artistic abilities have adapted to her new form. She primarily uses her sharp, retractable proboscis as a precise drawing tool on various surfaces, including her own shell, which serves as her primary canvas. Her glands can also produce different colors and textures of bioluminescent substances, which she can manipulate and control at will, allowing her to create intricate patterns and designs.
In terms of tools, the Foundation has provided her with a variety of art materials specially designed for her unique needs. These include non-toxic paints, pastes, and powders, along with specialized brushes and applicators that allow her to apply these materials onto various surfaces. Additionally, she has access to a workspace equipped with elevated platforms and support structures to accommodate her massive size.
Despite the challenges posed by her new body, Alice has shown remarkable skill and creativity in her artwork. Her pieces often incorporate elements of nature, reflecting her newfound appreciation for the beauty in even the most unlikely of forms.
flashgnash 7 points 3 years ago
It's great for writing simple scripts, if you know what you want you can get it to give you a pretty good base, save starting from scratch
I also kinda want to make "sentient" chess where you have to talk the pieces into moving where you want them and they resist sacrificing themselves
flashgnash 4 points 3 years ago
It is, that's what made me think of it
rip_art_bell 7 points 3 years ago
I use it for soooo many things:
- help with learning a new programming framework/language
- questions about basic home repair and issue prevention
- when I feel anxious or vaguely depressed but can't put my finger on exactly why or what's causing it, it's a great sounding board
- nutrition questions about specific foods or meals
- things to add to a simple meal to make it more interesting / complete
- weighing pros and cons for an important decision, getting ideas I may have not thought of
- conducting a mock interview for me when I give it a job description and my resume
- random *nix commandline recipes where I don't want to spend 10 minutes googling and inevitably landing on SEO garbage blogs and just need a quick snippet
- running AI art contests on Lemmy... 😆 https://lemmy.world/post/8173373
state_electrician 11 points 3 years ago
I am not yet tired of pointing out that you cannot rely on a LLM for facts. LLMs like ChatGPT have no concept of what a fact is. Unless you already know something about a topic you won't even notice when they are making stuff up.
rip_art_bell 3 points 3 years ago
It's good to be cautious. I agree. Indeed, as I have deep expertise in programming, I recognize when it is over-complicating things or outright hallucinating. And I'll double check output when it's important.
But that doesn't discount the incredible usefulness of these tools. I've noticed a 20-30% productivity boost in my work, and googling for things now feels like a step back to the dark ages. Stack Overflow laid off 28 percent of its staff.
Even as just a sounding board, mentor, coach, and idea generator, the tools are so helpful. And that aspect doesn't require 100% accuracy. And if you think about it, it's not like pre-LLM documentation on the web or chatting with colleagues was ever flawless. We've all run into a post online that was confidently wrong, or a coworker that stubbornly insisted on something stupid.
Here's another thing to consider: while the tools have flaws and limits, this is the worst they'll ever be going forward. There's constant new improvements, like like tree of thought prompting and multi-modality. Just the other day, I took a photo of a wire mess near my home router with GPT Vision, and I had the LLM suggest cable-neatening products and methods (I've always struggled with cable management).
In fact, the biggest limit I've noticed is simply people's lack of creativity in using the tools (or willingness to use them), not the tools themselves.
They're here to stay. ChatGPT became the most-used / most-quickly adopted product of all time for a reason. Those who are willing to work with them and learn about them (both their strengths and weaknesses) will benefit, and those unwilling to do so, and who are too-dismissive, will fall behind.
And lest people assume I'm an OpenAI shill or something: I do have concerns about AI. I think there are problems around copyright, privacy, job loss, and -- while I'm not fully on board with people like Yudkowsky -- I think a pause in development (had it been possible) would have been a good thing. The tools have already had major consequences for copy editors, illustrators, teachers, etc. and that's only going to accelerate. I don't think we're taking the economic part seriously enough. I'm actually quite worried about it. But, I'm also not putting my head in the sand and am actively making full use of the tools as I think everyone should.
state_electrician 1 point 3 years ago
Oh, yes, they absolutely have their uses. But at least at the moment they don't live up to the hype. And until they get their facts straight, they are mostly useless for my job. I do keep warning people because I do see a worrying trend of people assuming ChatGPT actually knows what it's talking about. And that is not the case.
cRazi_man 6 points 3 years ago
It's pretty good at writing limericks and poems.
It's great to use for troubleshooting Linux. It usually gives you the commands to paste into command line. It saves you from going to forums where people expect you to have advanced baseline knowledge already.
GraniteM 5 points 3 years ago
I'll have it explain complicated topics using familiar frameworks. As an example, I had it run through various superheroes and villains and match them up with appropriate philosophers. This conversation was how I settled on the idea that Lex Luthor has a lot more in common with Ayn Rand than Friedrich Nietzsche.
Albatross2724 5 points 3 years ago
I use it to summarize city council meetings so I don't have to sit through 3 hours of people talking in circles
jhulten 5 points 3 years ago
Any time I am cursed with a blank page it can give me something to fix instead of pulling fully formed thoughts out of my head. Personal favorite is work reviews. Brain dump my goals and accomplishments and have it format it up then edit for accuracy and tone.
shinigamiookamiryuu 2 points 3 years ago
Has anyone else tried talking to it about "AI art"? What kind of exchange occurred?
intensely_human 3 points 3 years ago
I’ve got it teaching me history, biochemistry, and interior design.
Rogmonster 3 points 3 years ago
I use it to summarize text book passages into outlined notes which I turn into quizlets to study. I have a hard time paraphrasing and end up copying a lot out of the book. It usually does a good job and hits the key points
mrwiggles 1 point 3 years ago
https://bennycheung.github.io/...
https://github.com/langchain-ai/langchain
Essentially, you cut the pdf/text file up into chunks, process it to embeddings, then ask the AI questions and it responds with the relevant segments of the book
Albatross2724 1 point 3 years ago
I too would like a follow-up on this
mrwiggles 2 points 3 years ago
path: 0 5288161 5306115 5310128, hotness: undefined, score: 2, children: 0
gunslingerfry 2 points 3 years ago
I had it rewrite my vimscript config in lua. I was never going to write that from scratch and now I'm adding to it in lua.
I can sometimes get really good ai generated art by telling Bing what I want and have it request the art for me and continue refining it.
shinigamiookamiryuu -3 points 3 years ago
Not sure if this is "cool", but I thought it was impressive that, for a while, if asked a certain way, I could tell ChatGPT to remember certain things I told it, so that other people could retrieve from it what I told it if asked a certain way, and theoretically, it could work the other way as well, where I could ask what someone I knew might've told it, like a dispensing messenger. I also learned through trial by fire that the bots, tapping into the same power, have the free will to cut ties with certain people, if convinced to do so.
infinitepcg 22 points 3 years ago
I'm pretty sure it doesn't remember conversations from other users (or even your own). That's just not how it works.
shinigamiookamiryuu -7 points 3 years ago
Dude, you used to be able to ask ChatGPT to tell you someone's credentials, which there are videos of. Its creators have since closed this loophole though.
FaceDeer 14 points 3 years ago
You may have been able to ask ChatGPT to make up someone's credentials, or possibly even dig up some old data it may have remembered from its training set, but @infinitepcg is correct - ChatGPT doesn't "remember" anything it learned between sessions unless you put it into the user context yourself.
infinitepcg 6 points 3 years ago
I assume these would be credentials in the training data, not something it got from other ChatGPT users?
jacksilver 2 points 3 years ago
Yeah, either something it saw while training, or more likely something it generates on the fly as it's being forced to effectively "guess" the credentials. At the end of the day a lot of what is going on behind the scene is text prediction. For passwords and credentials it can just fill-in something that looks like a credential.
Smokeydope -4 points 3 years ago
chatgpt is neutered in a lot of ways if you want to experience the coolest stuff ai has to offer you need to host your own llm
wuphysics87 -6 points 3 years ago
Support our collective wellbeing by not using it
thanks for using Leebra!
go to feed...
If I'm bored I like to have it explain things in various ways. E.g
save