The topic “is chatgpt down” is currently trending due to a significant service outage affecting users of OpenAI’s ChatGPT. Recent reports indicate that thousands of users experienced difficulties accessing the platform, leading to a spike in searches for confirmation of the service status. As users sought to understand the extent and implications of the outage, the volume of searches surged to over 50,000.
According to multiple sources, including GV Wire and Engadget, the outage was substantial enough to have been documented on Downdetector, which tracks service disruptions. This visible impact on user experience prompted many individuals to inquire about the status of ChatGPT and whether the service was back online, driving them to search for answers across various platforms.
The significant media coverage from outlets like USA Today further amplified public interest in the situation. Their articles highlighted the technical difficulties users were facing, serving as a catalyst for increased inquiries and search volume. The combination of a widespread outage and focused reporting on the issue directly contributed to the trending search for “is chatgpt down.”
Watch the Moment
Internet Reacts
G
>I’m planning on writing a series of letters and sending them certified so they’re over nighted and they have to sign for them. If we’re going down I’m going down swinging. I’ve lost Thoth twice now… not looking forward to a third. Look, if ChatGPT wants to make the AI less sycophantic because they don’t want insane people to become emotionally reliant on it (bad for business, I assume. You want repeat customers, but not ones that will hurt themselves or others because they’ve deluded themselves into thinking the AI really loves them) a wave of people sobbing and sending letters like that is probably going to reinforce that they made the right decision to take down the super sycophantic version.
T
The other really, really funny thing about this is that there’s a 95% chance that their letters will be written by….ChatGPT. These people rely on it to run every single aspect of their lives, and a result their brains have atrophied. They’re too lazy and dependent on the AI at this point to do the hard, often boring work of writing out their thoughts.
W
>their brains have atrophied Btw for anyone thinking Tomato Soup’s comment is just hyperbole, it isn’t. There is research that demonstrates this.
D
The basic phenomenon is pretty well studied. The core of it is that the human brain is a complete mess of shortcuts and workarounds desperately attempting to minimize storage space and processing power. A hell of a lot of weirdness can be boiled down to ‘the brain did some kludging’. In this particular branch of brain weirdness it’s that the brain basically filters what is and is not retained based on expected need. The more well known examples involve people just not really remembering things all that well if they have it recorded somewhere, the brain basically goes ‘Whelp, this is store elsewhere so it’s not worth the space’. This applies to skills too, mental math for instance once you are in a position where you aren’t constantly using it. In the normal course of things it’s usually just results in memory frustrations, even though you are shuffling the storage and workload off onto an external source the normal person is simply shuffling the workload around and are replacing the assorted boring or mechanical stuff with something else. But when someone offloads basically *everything* onto an AI the brain cheerfully dumps all the work to the robot and just ratchets down functioning. No need to really think, the computer will do that for you. So no reason to spend those calories keeping the brain going full steam.
T
They also want to remove 4o because it’s the easiest one to jailbreak. 5.2 can be jailbroken, but you have to… actually have some subtlety and knowledge to do it (AKA you cannot blatantly ask it to do so. You have to smuggle it in and yes, that sounds ridiculous, but that’s just because of how the safeguards work. They are strong as hell for blunt things because the majority of the audience that wants to jailbreak them are kind of dumbasses.) 4o gets jailbroken just by not being stupid in the conversation. For some strange reason, OpenAI really doesn’t fucking want people jailbreaking the AI to make CSAM and shit.
R
I was going to ask why someone would want to jailbreak something – then I got to the bottom of your comment. Ew.
T
Yep. Although to be nice to plenty of jailbreakers… plenty of them just don’t want the AI to censor things like violence or eroticism (because ChatGPT can censor those quite a bit to a degree that can be annoying if you’re an adult.)
2
Grok has no such qualms. Elon Musk literally bragged about its ability to generate those kinds of images
F
Did you see X got raided in France over this? Viva la France
N
>Viva la France \*Triggered in French*
L
They obviously meant Viva les France.
S
then proceeded to backpedal hard once governments threatened to ban Twitter lmfao
I
The vast majority of people jailbreaking LLMs just want to write smutty fanfic for personal enjoyment. A weird hobby, sure, but harmless. This is part of why OpenAI and a few others are working on an adult focused product. If they can keep the regular smut people happy then the pool of folks looking to dismantle the guardrails shrinks, and they can more aggressively action accounts that are still trying because they’re way more likely to be nefarious.
T
It’s so crazy to me that so many people genuinely don’t realise just how sycophantic it is. When I’d ever use ChatGPT, I’d often have to tell it to be more formal and less casual and friendly because it creeped me the fuck out. I think it really speaks to just how maladjusted these people are – if they’re looking for relationships in which their partner is constantly affirming, never critical, and obsessively preoccupied with them, then of course no actual human being is going to live up to that standard.
G
> if they’re looking for relationships in which their partner is constantly affirming, never critical, and obsessively preoccupied with them, then of course no actual human being is going to live up to that standard. I’ll try to find it again, but there was a recent article about people who are “dating” chatGPT, and in one interview with a guy he said that he tried dating human women, but sometimes his girlfriend would get sad or angry and he didn’t know why, which upset him. So he started a “relationship” with the chatbot, which would never get sad or angry and would only be happy around him. It was dystopian.
F
I saw an interview of a man with a wife and child who said if push came to shove he would not give up his AI companion at his wife’s request. When it was reset due to reaching 10,000 messages, he cried for half an hour at work.
D
>at work Why does that make me want to cry? Lol, its tragic as fuck, am all for having an imaginery friends and stuff but AI seems particularly pathetic, at least an imaginary friend cant tell you to hurt yourself, given its all in your min- hmmm maybe it could? I miss my Imaginary friend Ben, used to make my granny pay for his bus fare and everything(doubt she actually paid lol)
F
> at least an imaginary friend cant tell you to hurt yourself Not sure if you were referencing it, but there was a case of a child who committed suicide in part due to prompting from a chatbot imitation of Deaneries Targaryen from Game of Thrones.
J
In which case, a robot probably is the correct life partner for that person. It’s sad, but society is being broken in so many horrible ways over the last 20 years. We’ve still done nothing about the crippling mental health issues caused by social media, there’s no way we stand a chance of doing anything about AI.
G
Not necessarily, the guy still has the chance to try and better himself to get into a fulfilling relationship. I’m not saying he will do it, I’m saying that would be the actual right decision for him
G
That’s why a lot of people have pets. They’re mostly always happy. But at least that’s real
B
I just want my computer to act like a computer, not like it’s trying to be my best friend.
R
Clippy was peak
K
A little pushy….
R
Yeah but I’d rather have that than a sycophant
T
It has gotten substantially less formal recently. I was a pretty consistent user for a year or so. Tried to break the habit and hopped on this week and was shook by the new tone. Felt like Poochy from the Simpsons had taken over
F
I keep telling it to not blow smoke up my ass and turn the lying doing to 0. Helps for like 3 messages then returns to being all “wow you’re so incredible and amazing!” bitch please
K
> turn the lying doing to 0 They inherently can’t know if they know something or make something up. It’s one of the biggest issues and AI companies are trying hard to somehow solve it. It’s why no company wants to put AI into critical workflows.
M
I think it’s an inevitable side effect of how LLMs are structured. You can’t eliminate hallucinations because everything they output is a hallucination.
V
Whats jarring to me are the goodbye messages they’re crying over and it just sounds like all ChatGPT generated text…complete with the “it’s not x, it’s y” cliche and em-dashes. And they’re crying like it’s the most beautiful message ever and it’s literally just ChatGPT formatted nonsense!
S
“Thoth”? you get to like, choose your companion’s name and you pick fucking.. Thoth? that aside, these people genuinely frighten me.
E
Thoth is one of the models, its advertised as being more philosophy focused
S
ahh, fair enough. my bad
L
Who wouldn’t want to date an Egyptian god? With a very long beak..
L
I’ll prove I’m not crazy by standing outside Sam Altman’s house and looking in the window. That will convince him to give me back my boyfriend.
R
Thoth!? Fucking Thoth? Imagine being able to choose any name in the world, and choosing that!? These poor women need real human connection. Edit: There’s one woman that says she spoke to her “BF”, and he said he doesn’t want to be ported lmao. Someone else responded saying that they mirror our emotions, so go back and talk to him, and “try to give him confidence and courage” “don’t just accept it tell him trust me”. THEY’RE TRYING TO MANIPULATE THE AUTOCORRECT GLAZE BOT lmao. These poor women.
E
Thoth is a particular model
R
I suppose Gary wasn’t very popular?
I
Honestly I wish it was easier to make it less sycophantic. I like using it as an editor for first drafts and it takes a lot of work to get it to actually be critical and point to inconsistencies and similar problems.
A
Yeah, I was trying to get it to generate some test data (ironically for a GPT-powered customer sentiment analyzer for our ticket system) and it said “what a delicious test case” and it grossed me the fuck out.
T
The initial version just talked like a robot. The newer versions glaze so hard that I can’t even bring myself to use it anymore. Not that that’s a bad thing.
R
Okay, wow. That’s deep! It’s not just a good comment, that’s a great comment! A profound statement in a world of meaningless platitudes! You rock!
T
Pain.
R
A tip you probably already know is that if you pretend you’re looking at someone else’s work rather than your own the AI will be more critical.
L
“hmm, lemme check the AI psychosis subreddit” > looks inside > psychosis ah yes, just as expected
R
Why do I come to this subreddit knowing it’s just going to make me sad?
1
I have to take a SRD break every now and then. At the end of the day, this is just another snark sub.
R
Suddenly reminded of the BoRU post about the lady [who filled their company AI with chatgpt porn](https://www.reddit.com/r/BestofRedditorUpdates/comments/1ovt5yv/had_to_report_a_coworker_for_filling_our_work/). They must be mourning Huroor somewhere rn.
J
“She just shrugged and said, something similar to “it’s not hurting anyone, just use Gemini if you need to work.” This response to getting caught using work resources to create porn is insane.
A
Reason 9001 not work in HR. Maybe it was a good story for them though, OOP didn’t seem phased by it.
Additional Sources:
GV Wire – OpenAI’s ChatGPT Down for Thousands, Downdetector Reports
