Xiangling and Alhaitham walk into a Denny's...
HTML-код
- Опубликовано: 16 июл 2024
- Can ChatGPT write good fanfic? Spoiler alert: no, it's not good. Kinda funny though??
Also fun fact, I spent a good half hour agonizing over what to do for the thumbnail of this video and then ended up using an auto-generated one 🙈 I'm a clown
Edited by Severus!! / _theseverus
Other channels:
VODs - @YingVODs
Clips - @YingClips
Merch: ying.haidl.live
Twitch: / ying
Twitter: / ying_verse
Instagram: / ying_verse
TikTok: / ying_verse
Resources: yingverse.carrd.co/#resources
Timestamps:
00:00 Start
00:55 Taking chat suggestions for randomized prompts
02:14 Who's getting isekai'd in the WHERE???
05:30 We need a new wheel
06:35 Important disclaimer: Notes on ethical AI
11:58 Haha funny bot make genshin fic
13:50 Ningguang and Hutao boba hangout gone wrong
18:09 I'm gonna fight ChatGPT
20:18 Please, just ONE good fic
23:50 Ok chat be honest
#chatgpt #fanfiction #genshinimpact #hoyocreators Игры
A little request: the disclaimer in the middle is important, and also the main purpose of this video (imo). I would appreciate it if you took the time to listen carefully and reflect on it with me, but I totally get it if you're just here to vibe and have fun too 🙏
I know not everyone will agree with HOW I'm using my platform to engage with the topic of AI tooling, but my goal here is to approach the topic in as fun a way as possible. I've recently been using Adobe Firefly a lot (it's trained on ethically sourced data for once, thank god) and it's deepened my appreciation for AI tooling. My ideal world of carefully regulated AI tooling and an informed user base might never come to pass, but this silly little video is my attempt to help us get a little closer.
As always, thank you for watching!!
A thoughtful, nuanced, informative discussion may get down-rated by RUclips's algorithm, but _I_ appreciate it.
There's no such thing as "ethical sourcing" for AI. Programs claiming to be "ethical" have been caught time and again with regenerated content that was stolen. Butchered signatures, butchered watermarks, complete compositions, etc. If you think any AI program, right now, without stringent regulations in place and enforceable are 100% "ethical" than I have an ocean view property in Oklahoma I'd love to sell you.
@@UraharaKoi Is there anything ethical on the internet ?
@@daromaniac3843 in that case why adhere to any kind of ethics at all? Do what you want, whenever and however you want and if your actuons adversely affects/harms others fuck them am I right? Ethics and treating others with dignity and respect are a myth, not something to strive for.
@@UraharaKoi What's ethical and what's not depends on who you're asking. So in the end the internet can't be ethical, it's just about what you're allowed to do and what you're not.
I absolutely marvel at Ying's ability to turn any activity into a fun, uplifting adventure.
I concur, always watch her videos 😁
my next question obviously is: have you heard of Character AI 🤔😏😂
IVE SEEN IT AROUND in screenshots on Twitter!!! It's FASCINATING
As hilariously ridiculous as this is, I really appreciate the high level explanation of ML and NLP and the ethical implications of this technology.
omg i cant wait for the next part. i didnt expect to actually get THIS hooked when i clicked on the video omg
I AGREE it was so much more than i expected going into it a;lsdfj
@@ying_verse i thought id just leave the video on the background while im doing my stuff but it actually made me stop whatever i was doing and FOCUS on what happens next ndsfhasjdfd 😭😭😭
i saw 2 seconds of this and i already love it
While I liked what you said in your disclaimer. I wished you would have highlighted (you did touched upon for a second but I think it was a bit easy to overlooked) a bit more clearly that one of the main issue with ChatGPT, that causes real harm, is the fact that the AI is partly trained of stolen work (this is even more eggregious with AI art, but nonetheless a core part of ChatGPT). While I think stuff classically protected by intellectual property (ex: textbooks and published books) might have been exempted from theft, a good portion of whatever you put out on the net is pretty much up for grabs. Not only is it a huge issue of privacy, but the way the data is acquired means that people can simply take a melting pot of stolen data and declare themselves the «author» just for giving guiding cues to the AI. For example, people creating AI art (lets say a specific software that uses art found online as its data source) can than sell it for a high profit, sometime without even telling the buyer that it is AI art. Without the artists whose time and work was stolen of even receiving a thanks for being the meat and backbone of this "new" form of content.
Stuff like analyzing or AI changes the way we interact with the world, how AI in general are prone to creating echo chambers, artists being worried that their art is gonna be "devalued"by people looking for a cheaper way to make content or the debate of whether or not AI could make true art, while justified concerns, are still in the domain of speculations and future predicting, worries concerning intellectual properties are very much a present issue.
So, while I appreciate a lot your disclaimer,
I think it might have been a good idea to present a more, for a lack of better words, "negative" view (as in "most AI art seems to cause a direct form of harm and we should use and talk about it with extreme caution or even wait for proper legislation before using it altogether", but that might be a bit of a harsh take). Simply, elaborating a bit more on the different issues regarding AI art might have been enough to clarify the purpose of the disclaimer. I know this seems to be pulled from a livestream so stopping to have a lengthy talk about AI art and its issues could have broken the flow, but I think the YT format would have benefited a lot from something like a 5 min standalone section (taken off-stream) added in the editing, not a YT expert if this is not as simple as I suggest it is nervermind my suggestion.
In my opinion, your lack of clarity on why exactly people are against this kind of AI, muddles the waters of the « AI have issues that we must talk about but we must also consider that under the right circumstances it is a useful creative tool» point (that is valid) that you seem to want to make. In short words, the message’s delivery threads a bit too close too the "AI art is just misunderstood and people against are just preventing others from having fun" argument (which I know is not an argument that you are making), for my personal comfort, IMO.
But anyway, this is just my 2 cents as a simple viewer. You do you, I do not want to tell you how to create you videos, and we can agree to disagree without any worries. :)
PS. I do not have a PHD in computer engineering, if anyone sees that I made a mistake in my understanding of ChatGPT (or AI art in general), feel free to correct me.
This 100%. As an author and fanfic enjoyer I've seen AO3 scraped for training AI models recently, tons of authors locking their works to registered users only to help protect it from scrapers *as recommended by AO3* (which locks out legitimate users that are still in the invite queue), and heard stories of people pasting unfinished fic into AI tools despite the authors' wishes. And doing that is, again, feeding an author's work to the AI model without their permission, which is unethical. I've gotten so depressed over all this news that I've fallen into depressive spells, wondering if it's still worth it to dedicate so much of my energy to writing fanfiction if people are still going to disrespect authors like they are. If the AI trend continues, I predict a sharp decline in real, human-written fanfiction in the coming years.
@@foozlesprite We go into this a LOT more in part 2, where my friend’s fanfic and the ChatGPT generated fic for the same prompt are compared head to head! I think it really highlights the (current) differences between generative AI models and human creativity.
While I agree that I could have dwelled a little longer on the problem of models trained on stolen work, I don’t think the higher-level discussion here is necessarily “threading too close” with defending OpenAI and other companies (midjourney, stable diffusion etc) for training models on stolen work; I think my stance is pretty clear, though of course your read on it here does make me think I can stand to be even clearer next time. As someone who first studied ML basics almost a decade ago, I’m someone who fundamentally believes in the power of AI tooling to make our lives better if properly regulated and most importantly, understood by its users. It’s an ideal far from reality for sure, but my optimistic attitude towards AI tooling might be what you’re picking up on here, rather than any defensive tendencies toward ChatGPT etc.
Just some elaboration on my thoughts, typed bleary eyed on my phone right after waking up 😂 I do want to be extra clear that I don’t mean to invalidate anything you’re saying or react defensively either and I’ll definitely keep this all in mind for part 2 in case another insert is needed during editing, thank you both for sharing your thoughts as well!
@@ying_verse It was pretty clear you weren't defending it and I agree that AI, if used ethically, has great potential in some fields. My issue is more that it's not really being used ethically or regulated in many scenarios right now. Any aggressiveness you might have caught from my comment is more from the fact that I'm tired of seeing people casually treating it as a toy for amusement while minimizing its issues and dismissing concerned people. You're not dismissing concerns, and made sure to bring up some of the issues. It's more complete coverage than I see from most dabbling in AI, and I appreciate that. So to be clear, I'm not upset with your take on it or anything!
this hurts my soul but it's so silly I can't help but like it
Referencing what you said about being past the event horizon, I agree that we're going into a future where tools and software's like ChatGPT are going to be common. I'm in my uni final year and one of my electives had an assignment where we had to use ChatGPT to assist in writing an essay on promoting academic honesty ethics and morals. The assignment was from a first year course so I was very surprised to see that, after attending a lecture on ethics and AI in academic dishonesty, they gave us that assignment.
I think this was a smart move from the uni though, given that these tools are available and easily accessible. IMO the best course of action (from an average person's perspective) is to educate and discuss the implications of these types of AI and machine learning softwares (especially since I'm gonna have to try make a basic one next semester). It also reminded me about an essay I wrote in first year about the consequences of the programs and systems we build (emphasis on AI) and who they affect the most.
(Side note: I'm also an Astrophysics major so the comparison to an event horizon made me really appreciate the analogy. And also made me feel dumb that I didn't think of it earlier)
Not the pudding
LIYUE BREAKING NEWS: lady tianquan ningguang and wansheng funeral director Hu tao defeats fatui "leader" childe and drink boba tea afterwards
Ying reading the “Huguang pudding incident” fanfic had me rolling
Edit: Tysm for the
Hey Ying! I appreciate the disclaimer in the middle a lot! I'm glad I have read the comments more to grasp a better view of your opinion on this !! I agree that the sourcing of how the ai is 'trained' should've been mentioned. art ai and writing ai are extremely complicated ethically and morally, personally art ai and writing ai go against my morals, especially when it had come out that AO3 authors had directly been taken from without their (and AO3's) consent. I, however, i think that its important to say that this use of chatgpt is pretty insignificantly harmful (aka any harm done is on a largely micro level). I appreciate you taking the time in the fun to address it at all and being open to the conversation in the comments! I also really appreciate the fact that you encouraged others to do their own thinking on this topic. I think it has far more depth than some may realize and i think that along with the fear ai creations have struck into creative makers make this conversation incredibility difficult and often more antagonistic than productive. at the end of the day, i do not think ai can take the place of a "hey sorry guys i just got out of the hospital after my house burned down and my car exploded" writer's soul/energy and you can make the content you want! i just really appreciate the care you've given to this topic and other's opinions! and also yeah this video is fun, ai can be fun, im not gonna sit here and lie about it LMAO
Is it wrong to try to pick up girls in a Denny’s parking lot after eating a pancake Grand Slam breakfast with a stranger at 2am: My stomach is wrenching and I have epic indigestion; there must have been dairy in the pancakes.
Legend has it the ai is still working to finalize its fanfics.
and they all rode into the sunset, knowing that the power of friendship would be their guiding light,
@@ying_verse couldn't have said it better myself oh wise ying
I love this video and your energy, can't wait to see the results of comparing fanfics!!!
As someone who did NLP as part of my undergraduate thesis, i'm glad you were able to share stuff about it to your audience, Ying.
That said, i need to make a self insert fic with Alhaitham brb
I feel called out for having relationship with my ais on character ai now lol
immediately i already love this
I expect romance coming into this but this is hilarious
You turned this into a very interesting activity Ying! Appeciated
If it was ethically sourced I'd probably play around with it, it seems fun in a terrible kind of way, but I have 0% faith in it lol. I don't think I'd use it for anything except inspiration though, or just generally being entertained with the silliness, art is ultimately a form of expression, so i don't think it should be taken over by machines, that would just make it pointless
Agreed!! I think a version of this trained on ethically sourced data could be helpful more as an assistive tool rather than a creative one; early on when it first released I tried using it for things like rephrasing my conference proposal titles and generating templates to use for paper abstracts etc, it was really helpful from that perspective!
Just what I needed at 2am lol
What we truly need is a full stream with those char/place/activity/tone wheels
NOO WHY DID I MiSS THIS STREAM THIS LOOKS SO FUNN😭😭😭
AAAA I wish I was there for that stream, it seemed so fun!
AHHHHHH A NEW VIDEOOOO
ying ily but ai fanfics are. so bad. and i mean that they just steal from already existing works. you should know better than this
edit: yes i see you made a disclaimer about that but it doesn't make it much better :(
I think it's important to engage with it deeply and explore together, personally! My entire purpose was to show that there's nuance here, and range in usage; AI tooling isn't unilaterally bad, but using it without considering the consequences or getting educated on where it all comes from is the main risk IMO. That's also why this video has a part two where I commissioned a fanfic writer friend to write a fic based on the same prompt we gave ChatGPT and compared them; this friend is also a law student with a specialty in this area so we dove even deeper on the ethical and legal details after!
I get that a lot of people see this as a big no-no space in general and to avoid it altogether, but personally I don't think that actually helps us understand what we're dealing with here. Totally understand if you disagree of course, and I appreciate your concern as well!
@@ying_verse well i still don't like it BUT thank you a lot for caring and commissioning real writers!!! 🫶 and for the reply too
When you started your disclaimer my first thought was "Oh no, not another youtuber with a bad misinformed take on AI" but then it turned out that you knew what you were talking about lol. Shouldn't have doubted you considering your major I forgot about
I need more of this
Might be the most cursed RUclips video title I've seen in a while. I clicked so fast 😅
Edit: love the disclaimer, and I say this with the utmost respect but I gotta say it: there's nothing quite as attractive as an intelligent and articulate person and Ying nailed it 😁
Maybe ChatGPT trained on data from restaurants in the Midwest U.S.? Serving sizes in that region are a well-known risk to one's health.
Thank you for this disclaimer Ying. I think it’s super important to raise awareness on this topic, even if we use it for fun and giggles!
Funny enough I've done this before 😭 I've asked chat gpt to write genshin fanfics 😂
OMG i remember the dystopian tragedy,, it was so good but we kept changing the ending but the bot got too overwhelmed and started to lose it logic.
Cute cute cute
i love this. silly fun way of using ai!! youre just using it for a silly time w ur viewers as everyone should be using this if they ever tell it to write something for them 👏
love you and your videos ying, yet i just wanted to state this: mona is indeed an adult, and fischl is confirmed to be around 16 years old, given the fact that she was 14 years old when she received her vision. with the vision hunt decree, it had been 2 years since the last electro vision was handed out.
unrelated to ANYTHING i need that sweater
😮 Somebody needs to send this to the Game Grumps!
Hahahahahhaa
Surprised that Bing Chat (that uses GPT-4) wasn't used here, since it's connected to the Internet...
This stream was actually from a few months ago!!
ChatGPT can't write anything , what it can do piecemeal plagiarize other peoples work and regurgitate back it to you.
Kween
Character ai can
the UwU Voice for that one story took me OUT!!!
Gay.
happy pride!!!!!!!!!!
@@ying_verse OH I WASNT EVEN THINKING ABOUT THAT I WAS JUST LIKE "gay"
HAPPY PRIDEEE
What's wrong with romance between two potentially minors?
I had no control over how far the bot was gonna take it LOL I’m not running that risk
@@ying_verse lol that makes sense, don't know why I didn't think of that
LMAO NOT THE PUDDING STORY💀💀💀💀💀💀
this stream was really funny to tune in to even though I have never genshin impact🤍