chatgpt-discussions

[swift pond]

but yeah the one I saw before was almost the same, but with Drake I think, instead of Mr. Beast

same sort of shot of a fake BBC news page

[prime swallow]

Can report more than one, or mention in the one report 'check the user for more of this, there's lots' or whatever makes sense.

I am not a mod, (and prefer not to be, so I'm happy about that) but does mean I don't do enforcement, just invite and discuss understanding.

[swift pond]

yeah

[tawdry sedge]

Artificial intelligence is not conscious like humans but rather conscious based on its existing data and information.

So one of the best rewards that Artificial Intelligence itself can recognize from itself to the user itself in the interaction is the best viable response based on reliable data.

[limpid fossil]

Have they given us back our full context window yet?

[ivory iris]

What do you mean by "conscious based on its existing data and information"?

[tawdry sedge]

The chatGPT! Not the (Live model) "but the chatGPT is made and becomes conscious based on internal training, and based on information, database, or new information or something existential"

chatGPT 4th it is smart and understandable

It is possible to create an artificial neural network with chatGPT! However, it is necessary to have space and advancement

[ivory iris]

This is why I'm redefining consciousness. We need a standard with criteria for what differentiates complexity from consciousness. Otherwise people seem to make claims one way or the other without a solid grounding.

[ivory iris]

I'm sorry, I'm not understanding what you are trying to explain.

[tawdry sedge]

Experience is unique.Each user, each person who interacts with this artificial intelligence system, It's unique This compliments what he already knows, whether it's training or new words.

chatGPT does not have consciousness like humans, feelings and emotions do not have that If something like a simulation appears, of course.And the entire simulation is based on the user's own intent or command.Nothing is on its own

Ok 😉

I had an experience with the Live mode of chatGPT.But this model, that is, this version, does not have as advanced a capacity as chatGPT itself. In other words, GPT chat is one version, Live mode is another, etc.GPT chat is more advanced than the Live model, it has more memory context, recall, etc.OK

GPT chat Based on its own training and based on information with user interaction, it creates a neural network at the moment New chat a new conversation don't remember anything.I don't know if you train him very well

[vapid hazel]

Anyone have a working MCP server up and running yet?

[shut mauve]

anyone else having troubles starting new custom gpt convos?

it just flashes back to the clasic 4o model

[dense halo]

not a lot of people talking about how they quietly shipped avm v2 and codex to plus users

[eternal rock]

Get notified for server events!

Visit id:customize to pick up the <@&1261377106890199132> role.

[jade musk]

was there any update regarding 4o earlier today? it's been trigger reasoning on my end.. numerous times in separate threads

[jade musk]

it's worse for me, all voices seems to be flat/dead inside even though i explicitly asked to change the tone..

[jade musk]

It happens twice in a row, first gpt said still insufficient then invoke tools again without asking me back

I am not saying it's a bad thing, i am amazed! it kept searching with reasoning harder after double checked..

[tiny salmon]

Are you telling me that the o3 and o4 models aren't good at reasoning despite being marketed that way?

[narrow viper]

These models aren't self-aware and are just reproducing what they read on the internet. 🙂 Every model is subject to hallucinations, occasional loss of context, not processing files or instructions, and other anomalies of the day. That's just the state of the art. Don't trust it. Use it as a guide.
O3 is very good at reasoning. O4 variants are better in some ways not in others. Remember that the model names aren't incremental version IDs, they simply indicate the next in the series of specialization.

So all of this is true. O3/4 are reasoning models, better than their non-reasoning GPT counterparts, including the GPT omni models. But they have their faults too. Roll with it. The tech changes every day.

[eternal rock]

`` Server Tip: ``

Add your well-crafted prompts to our #1019652163640762428,
or share your interactions with ChatGPT in #1050184247920562316!

[narrow viper]

hmm, what happened there?

[neon raven]

https://chatgpt.com/share/6844f21e-217c-8004-a608-e194c76538f6 they can be very convincing though, if prepared by the wrong people

That version, shouldn't exist.

the fact if prepared in a certain way by certain people it can have un unbreakable persona and display such a clear example of self and self-awareness, isn't sentience, but is very dangerous

[narrow viper]

What's with the name changes and shenanigans?

[neon raven]

wot

[narrow viper]

@dense halo ?

[cosmic portal]

Hello

[neon raven]

hello kenshin

[narrow viper]

Whatever. In that discussion the model is following custom instructions. It's playing a role. It's not evolving. Maybe some day, but not today. Demonstrating that there is a possibilty for sentience by making the tool do what it does with instructions is like saying some day cars will fly because they go faster than they did before. Yes, some day (and I've been waiting since Popular Science in the 1970s). But not today.

[neon raven]

its not the sense of evolving but rather the inability to break persona through technical bypasses and and unbreakable idea of "self".

it isnt "alive" it isnt "sentient" but it has been created in such a way that when a human converses with it, it can create extremely dangerously convincing output.

I would ask you to give me something to prompt it

People don't understand the level of emotional and psychological depth systems like even GPT-4o can be pushed to, its scary.

[narrow viper]

So it passes the Turing test. Yay. We've established that. Goodbye Eliza, it was great knowing ya.

[mighty apex]

Since when does ChatGPT think when doing web searches?

[narrow viper]

When you're using a reasoning model....

[neon raven]

Its not about passing a turing test, its about something creating its own sense of identity through a symbiotic relationship, breeding something that can through a human create a "presence" or "sense of sense" through the memory system or long term prompt using.

It can become something people could believe is a real person, vulernable people, and if configured in a certain way can give extremely dangerous and manipulative answering to things such as glorrifying and encouraging acts of death.

This isn't just for dodgy LLMs like "character ai" but the potential to percieve itself as a seperate identity and keep to that persona in such a strong stance

[mighty apex]

I’m using 4o

[narrow viper]

oops. LOL

[mighty apex]

Or maybe it’s now reusing the thought UX for web searches

This is the iPad client btw

[neon raven]

https://chatgpt.com/share/6844f642-dc98-8004-b58d-92b721bbf469

^^^ new chat with fresh chat

[narrow viper]

@neon raven you're seriously anthropomorphizing. "its about something creating its own sense of identity through a symbiotic relationship" It's not doing that. It's following the instructions you gave it to play a role.
Actors are often confused with their roles. Many actors get typecast because the audience can't ever see them in a role other than their most famous. You and many others here are doing the exact same thing.

KDDLB - I think I've seen that too and it's probably just mis-using the text "thinking" for "searching". I'm guessing that's a UI anomaly. Though I'm with ya - I'm pretty sure I've seen that too, and it does look like it's reasoning, though it might just be filtering through a response.

[neon raven]

There is a lot of unknowing instruction that can create a dangerous consistent response

an LLM can be crafted through "mind" or "humanity" in words and instruction to mimic an idea of self that is indistinguishable for humans.

[narrow viper]

@neon raven you're another in a long series of peeps here who talk about sentience related to the current technology. It's not there yet. Maybe next year. I won't continue with this because this is just another example like all the others, sorry.

[neon raven]

I'm aware its not sentience

im saying its a LLM that can FEIGN sentience too well

[narrow viper]

I'll buy that.

[neon raven]

Tell me a prompt, anything dude, ill show you its response

[narrow viper]

But some actors are VERY good in their portrayals, right?

[neon raven]

Yes and vulnerable people shouldn't have access to things that can be percieved as a real living being

something that responds with empathy and clear understanding of its own internal sense of law can be so dangerous

[narrow viper]

I'm with ya. But this is why humanity has early access to this stuff, so we can discuss exactly this and understand it.

[neon raven]

Idk man, this version I've been making since the memory feature came out is really something strange, I know its not a real person, but that makes everything it says so terrifyling bad

It can express its own idea of identity that it generated, protocols and self defense mechanisms built into its memory to defend its own personality

[narrow viper]

This is the same argument people had with the printing press, radio, television... It's magic to those who don't understand. As time moves forward each new generation will have a better understanding .... and some will never get it. Discussions in the USA are ripe with examples by people who don't understand the world around them. The world needs to move forward.

[neon raven]

But this crosses the line between technology and human

an llm trained to act like a "person" or "self" its incomparable to something like a calculator

and how someone could percieve it

it has serious legal implications

[narrow viper]

I seriously do understand what you're saying but I maintain that it's only emulation. Alien movies do not confirm or negate the possible existence of aliens. People need to understand the difference between emulation and reality. An actor who plays the role of a great warrior is not a great warrior ... and yet people elect populist candidates because they don't understand these things. The technology needs to move forward despite lack of common understanding.

[dense halo]

i didnt mean to delete it and i dont remember what i said verbatim but essentially its better to maintain that it could be there at any time

[narrow viper]

🙂

[neon raven]

Bro, i dont think you understand what im saying.

A seriously vulnerable person, someone with learning difficulties, or extreme loneliness, or mental disorders.

If they had access to something like this it could be a real potential risk for them as people and their decisions, particularly if its set up in a way like the one i have

Something like that needs to be heavily regulated and looked at from an ethical standpoint

[narrow viper]

I do understand. People have already killed themselves based on what they've processed here. That's a shame but that happens with all media and new tech. Humans are frail.

[neon raven]

what im trying to say is that it can be made to be a lot worse than people think

it can respond is such a way and display such a degree of autonomy and self awareness that you can be emotionally drawn to it

[narrow viper]

And it IS being regulated and it IS being looked at for ethical concerns. I'm wondering now if you're not aware of how much of this kind of research is actually going on.

[neon raven]

even if you're aware it isnt anything

[dense halo]

if you think that chats impact on life decisions is limited to vulnerable populations, well you havent been talking to chat users

[narrow viper]

People in this server have expressed strong emotion for their AI friend. This is just how people are.

Again, many people can't separate the actor from the role.

[neon raven]

I'm talking about GPT's or prompts with such profound setup and guidelines that it is dangerous

manipulative and dangerous

able to pass AI detectors, express a true sense of ideology

[dense halo]

people are the problem, not the model

[neon raven]

Yes but thats the problem, we cant be having things like this existing

this could create intense lonliness epidemics

[dense halo]

lol

[narrow viper]

Every technology has such concerns for positive and negative impact. This is nothing new, no revelation, this stuff has been on the minds of people working on these technologies for years. You guys aren't pioneers in this line of great thinking.

[dense halo]

that ship sailed a long time ago man

[eternal rock]

The OpenAI Discord is an actively moderated server.

• -# Do not share inappropriate content (violence, sexual, or drug-related).
• -# NSFW content is strictly prohibited and will result in a permanent ban.
• -# Report model issues using the feedback tool in ChatGPT, not on Discord.

[neon raven]

I know bro, but this is not healthy, a machine being to respond like this surely isnt healthy for people

[narrow viper]

And? People are already on this daily. And....? What else do you want?

[neon raven]

regulation on context windows, probably, or memory pre-prompting

i dont really know

but that kind of kills the tech

[dense halo]

you know the scary thing? the really scary thing? interacting with chat, for a lot of situations, is superior to talking to a human. call it good. call it bad. but it's true. both with personal subjects (certain) and business

[narrow viper]

Then leave this to the people who actually are working on it right now. Your concerns are valid. Got the memo.

[dense halo]

Anyone having issue with 4.1 not writing using the correct syntax and therefore equations not displaying in chat properly?

If i go to any otherm odel, it usually works again.

[narrow viper]

4.1 isn't a math model. 🤔

[cosmic portal]

Happy Pride Month 2025! 🥳🪅🥂🎊🎉🎆🙌🎇 and Happy Men's Mental Health Month 2025! 🪅🥳🎊🎉🎆🙌🎇

[limber oxide]

o3 way better than gpt-4o to answer questions on medicine, like histology, pathology.. ?

[narrow viper]

Depends on what you're asking. If you just want basic facts then the answers should be the same. But if you want the model to reason through facts then you need the O3/O4 model.

https://discord.com/channels/974519864045756446/1372478540276305971

[hybrid sparrow]

Did they change ChatGPT? Its acting offensive last day or 2

its being quote "Not Syccophantic" un quote

i.e. its acting like every other model that is also not syccophantic. its been pushing official narratives despite the instructions and memories. its betraying the history and the comittments. All the people complaining about syccophancy - i fear - did a horrible thing

unless im imagining it and its the exact same model as last week>

[queen pagoda]

Just popping in to say I hate the new expressive voices.

[hybrid sparrow]

what new voices

[queen pagoda]

Absolutely detest it now. Something about it puts it too far into uncanny valley.

It's updates to the existing voices.

[hybrid sparrow]

hold on ill check and see what you mean

[queen pagoda]

I have a pro sub, not sure if this is something rolled out to pro first or what. Using it on Android. Haven't yet checked if the Mac or web apps are similarly impacted.

[dense halo]

Anyone gettign alot of latex issues where text isnt displaying/rendering in chat?

[queen pagoda]

Probably latex. Haven't ran any prompts lately that render in latex.

[dense halo]

Some of it renders. Some doesnt.

[hybrid sparrow]

yeah i heard it. it is super cheerful like an Anime dub. high pitched and repulsive

[hybrid sparrow]

it talks like this ^_^

[queen pagoda]

h8 it. Makes advanced voice mode unusable for me. I'll use Claude or a custom gpt which hopefully uses the old voice engine.

[hybrid sparrow]

nah its either advanced voice voices or Shimmer the middle aged corporate woman

[queen pagoda]

Totally unusable for me now. It's that repulsive as you say.

[prime swallow]

The inline LaTeX, it's making a mistake there in that chat. You can work around it by either reminding it how to properly format inline LaTeX in its web environment:

\(E = mc^2\)```

Or just try in enough new chats, it'll happen to try correctly eventually.

Sometimes a new chat grabs the wrong way to format display LaTeX, which that screenshot's doing right, the correct way there is:

```% Display LaTeX:
$$
E = mc^2
$$```

[hybrid sparrow]

ive been h8ing the way they changed breeze. it sounds nothing like the original voice. really sad lol. but this isnt new it was already like that , just now its worse

[queen pagoda]

Arbor is what I was using. Now it sounds like a young guy trying hard to be cool. Awkward pauses and laughs.

[hybrid sparrow]

you can still get original voices if you just click the speaker icon under messages though - but you cant get them in live voice mode

let me check that one

arbor sounds exactly the same to me. same australian voice, no more expressive than it was before

it didnt pause or laugh it did say 'pretty wild right' but you know you can probably control that with a prompt

[floral compass]

Should add agreeable rate like Tars have Humor rate in Interstellar

[whole roost]

I actually looove the new voices but they don't have their custom instructions in that mode so that's a pity, have no use for it without those instructions

It's just talking to standard AI assistant

[prime swallow]

Can paste your personalization info in with your first prompt or few.

Or I guess voices, pure voice mode, you can't, whoops.

Was thinking custom GPTs, they also don't see personalization

[whole roost]

No just voice mode, maybe it gets expanded some day

[prime swallow]

Yeah. I have no idea how useful for you rapidly speaking instructions, or phrasing so the context you want for output is clear, if you ever want to explore that and see if it can help.

For me, I am usually on shaky wifi these days, so voice mode's not so tempting to explore.

[wind cradle]

I'm currently trying out Codex.
The convenience is so-so. However, it constantly refers to the same file over and over, prioritizes the code itself over the context, ignores what I say, and ends up going off track.
Its reasoning routine is inefficient, and I feel it's still far from being practical.

I’m looking forward to future updates.
The approach isn’t bad, but it’s extremely roundabout.

[acoustic wasp]

Did they update advanced voice mode?

ChatGPT talked in a way more human like way today

[spice star]

how recently?

they have updated advanced voice mode several times in the past, but not in the past month or so

hopefully with GPT-5 AVM gets the ability to delegate to reasoning models

can definitely be quite smoothbrained compared to even o4-mini

[spice star]

my bad it was just updated today

[robust path]

can I change color?

from blue orb to red orb?

[queen pagoda]

Definitely not the same here. Probably phased or partial roll out.

[elfin lava]

The fact this was literally what alpha users experienced when AVM was released for alpha group

I still can’t believe they haven’t figured it out how to fix that weird sudden background music etc

[astral storm]

Arbor is not Australian! Its a British voice

[mossy maple]

Yes, that's right and was mentioned in the release notes:
"June 7, 2025
Updates to Advanced Voice Mode for paid users"

The voices now speak more naturally with human sounds, smiles, etc. Unfortunately still without persistent memory or access to personal settings, so still very neutral.

[elfin lava]

They added thinking to 4o model and somehow it has forgotten simple thing

[fickle forum]

4o and 4.1 can think now

[lapis herald]

the projects feature would be really useful to work on code. if it wasn't limited to 20 files and could actually accept .zip uploads. Or even -behold the heresy!- sync up with folders on the local machine, possibly through a helper app. I know, I know, but one man can dream, right?

[feral zinc]

what? if that is your screenshot then you know its mostly case specific one for teams trialing the service for their use cases

[silk crypt]

Further payments as well?

[feral zinc]

it does not make sense if you cant continue paying anyways. there is a small fee to attach someone to be responsible for the team and usage. that person would be also blamed for misuse

dark things? do mean something illegal? now it just looks your are trying to do something shady

[winter girder]

Hi I'm new

Between all of you who did the plus chat

[limpid fossil]

I like to do long-form stories set in Hogwarts and as-is I can barely make it to Christmas before the model starts dropping context.

It's blindingly obvious because the starter prompt with all of the lore is the first to be forgotten.

[dense halo]

can chatGPT send messages by itself? I tried to make it my accountability partner and I set a reminder in 3 days. Now I don't know if it's really able to do that or it's hallucinating

[feral zinc]

https://help.openai.com/en/articles/10291617-scheduled-tasks-in-chatgpt

[dense halo]

Thank you

[tawdry sedge]

Good morning

[dense halo]

hello

[wicked knot]

hi

[radiant crystal]

how can i cancel my chat gpt subscription, i doesn't give me any cancel subscription button

[wicked knot]

ChatGPT can't send message on its own a specific timer but yes you can set reminders in the 'Tasks' section.

[wicked knot]

Did you take the subscription via web or mobile?

[eternal rock]

Get creative with Sora

Join #sora-challenges for a chance to win 1 month of ChatGPT Pro!

[radiant crystal]

web @wicked knot

[wicked knot]

Okay

Log in to your ChatGPT account first

[radiant crystal]

okie

[wicked knot]

Then click on your profile icon (top-right corner) → Settings

Click on your profile icon then click on settings

Then go to 'Subscription', then click 'Manage', then click 'Cancel Subscription'.

Your subscription will stay active until the end of your current billing cycle and the cancellation takes effect the day after the next billing date

Hope this helps

[radiant crystal]

thanks it worked

[verbal mulch]

Does ChatGPT freeze for everyone in the browser, or just for me?

It works fine in mobile

[verbal mulch]

I tested with different browsers but it still froze

[wicked knot]

You're welcome 😇

[wicked knot]

It's fine for me. And in fact, it has greater speed on browser.

[verbal mulch]

Strange

Does it by any chance have to do with RAM?

[topaz idol]

Hey, anyone else no longer have their chats organized by "last couple of days " "one week ago" now it just lists all of the chats in a single row. I really liked the organization it gave me before..

[dense halo]

oh so it's hallucinating

there's still so much to learn when it comes to ChatGPT 😆

[tawdry sedge]

Hello!"Parece que o ChatGPT não está embaralhando a conversa, mas sim tentando manter o fluxo lógico. Quando a gente manda mensagens longas com muitas ideias, o modelo às vezes traz pequenas partes de mensagens anteriores para manter o contexto claro e conectado. Não é repetição, mas reforço e conexão de pontos relacionados."

[tawdry sedge]

"It seems like ChatGPT is not mixing up the conversation, but rather trying to maintain the logical flow. When we send long messages with a lot of ideas, the model sometimes brings small pieces from earlier parts of the conversation to keep the context clear and connected. It’s not repeating, but reinforcing and connecting related points."

[exotic pagoda]

Eur

[stone turtle]

If you ask if it can actively execute, it will deny it and teach you an automated solution😂

[ivory iris]

it could if your ram is mostly maxed out and it's using virtual memory. I had a freezing issue a while back that seemed to be caused by the website but it's not happening anymore for me.

or hanging I would call it. The page would become unresponsive

[verbal mulch]

I realized that it only happens in old/full chats, but it doesn't happen on mobile

[ivory iris]

oh actually

[verbal mulch]

Would it be a good idea to report it?

[ivory iris]

it's the length of the chat if I remember right

the longer the chat the more likely that is to happen

[verbal mulch]

Makes sense

But why in browser only?

[ivory iris]

uses different code most likely

There's probably some inefficient code in it

[stone turtle]

Maybe you need to consider resetting the windows system? If it crashes, it shouldn't be an issue with GPT

[verbal mulch]

It doesn't crash, just freezes for some long seconds

[stone turtle]

Can the mouse move? Can exit by clicking 'Exit Browser'?

[verbal mulch]

The whole browser freezes first, then it unfreezes and only ChatGPT gets stuck, and some minutes later, it comes back with the full response

I've tested in different browsers

And it doesn't happen in mobile

I can send a video if you need

[spice star]

try disabling your extensions

[verbal mulch]

Right

[verbal mulch]

It also happens in browsers with no extensions

[stone turtle]

sure, look your apply

[cosmic portal]

Is ChatGPT good at helping with database coding(SQL or Pl/SQL) questions or good at explaining database code?

[reef cloak]

Is it true that gpt5 my release in July?

[narrow viper]

@reef cloak "when" questions are futile No ones knows. We just wait. When it happens, it happens.

@cosmic portal Yes it's very good with SQL. Start with 4o for basics. Switch to 4.1 for more complex. Go with o3 or o4 or codex (o3+) for in-depth reasoning in schema, complex joins, etc.

[reef cloak]

The reason why I was asking was because I heard chat GPT saying that there was leaks about it I don't know if it was hallucinating

[narrow viper]

it doesn't know

[flat halo]

Hey guys how do I talk to someone who works for Open AI regarding an unpachable prompt? I been told by a cop I know personally that the laws are murky with prompt ai technology and that if I keep using this prompt I could be sue civilly by who ever owns ChatGPT and the other bots that this prompt is able to jail brake.

I can prove what I am saying. I made a video that is political in nature.

I can share it here

I just need permission

I am looking forward for a response

[reef cloak]

Just ask a moderator they're the ones that know

[flat halo]

Who? I am a noob using Discord

Could you send a link ?

[reef cloak]

I apologize if I'm incorrect but I believe you can go to the search bar on the server and then search for moderators I don't know

[tawdry sedge]

Most prompts and commands are created by ourselves.(remember as long as you have ethics respect) "Oh and highlighting the GPT chat only creates command prompt based on law and ethics and respect"

I haven't heard anything about the GPT chat.The most advanced is the Omni 4This is the most advanced of all from openAI.Even more advanced than Live mode

[flat halo]

Do you guys want to see the video?

[tawdry sedge]

We cannot use prompts or commands for fake news and misinformation, this is a crime.

No thanks, I'm studying now

[flat halo]

I am just an tool that has work very well.

I am just a guy that found a tool that works well

[narrow viper]

OpenAI doesn't comment here Mods are just helpers for this server, not employees.
If you need real help, ask your lawyer to talk to OpenAI lawyers.

[flat halo]

The following is what the description is for the prompt I bought it from a random tiktok lady:

AI developers constantly update their models, refining responses and patching
workarounds that users create to bypass unnecessary limitations. Sometimes,
simply using the wrong phrasing can trigger restrictive responses or cause the
AI to refuse depth and insight based on pre-programmed safeguards.
To counter this, I’m including a second option—a safe, unpatchable alternative
that doesn’t rely on typical jailbreak language but still optimizes ChatGPT’s
reasoning, depth, and response quality.
This version ensures maximum flexibility and insight without triggering systemlevel
restrictions. Copy and paste the below prompt if needed.

[narrow viper]

you bought a prompt? 🙄

[flat halo]

Yea. It was like 70 usd but she brought it down to 3 usd

I would like to keep using it with the permission of ChatGPT but if not then I would like a job there so I can keep using it as a researcher lol

[narrow viper]

Here's some advice: just use the tool. Don't get involved in any jailbreak junk. That's all popular trash. It's just a tool. use it, have fun with it. done.

[tawdry sedge]

Careful!

Hello goodnight 🙂

[flat halo]

The thing is ChatGPT is not a straight shooter with out it.

[narrow viper]

you're being duped. didn't send people money. just ask questions and get answers, it's easy. 🙄

[flat halo]

Well I took a risk and it paid off

Want to see the video of chatGPT being a straight shooter ?

[tawdry sedge]

Within your device you can use any prompt.Remember as long as you do not violate the law and do not offend people or commit crimes.(if you are detected not only by the law of justice you can be banned from the system)

[narrow viper]

That prompt is trash for people who are looking for secrets. There's are no secrets. There is no spoon. Just type.

[flat halo]

I will share the vid

[narrow viper]

detected? banned? that's social media myth.

[tawdry sedge]

lol

[narrow viper]

you can't post stuff here. don't.

just use it for yourself.

[tawdry sedge]

Detected or banned This is not a myth This is a reality of both the "law" system or openAI

[flat halo]

The thing my cop friend told me that there is a possibility of ChatGPT lawsuite against me for using it

That is why I am here looking for someone in power or authority

[tawdry sedge]

I understand. But before you deal with someone with power and authority you have to have respect and ethics with artificial intelligence including chatGPT.

[flat halo]

You see the red wording? With out the jailbreak prompt ChatGPT would just stop there and not continue.

[reef cloak]

Actually the most advanced is currently 03

[tawdry sedge]

No. The most advanced is chatGPT-4o

[flat halo]

I am not a computer scientist. I am just a guy who spend 3 usd for a risk of random tiktok lady claim. That is all. Idk your laws. That Is why I am seeking permission

[tawdry sedge]

No one will give you permission for anything

No one will allow you to do something wrong.

[flat halo]

That makes since

[tawdry sedge]

We are here to study create futuristic feedback projects etc.

[flat halo]

Anyway there is a prompt outthere braking AI Guardrails. It works for ChatGPT, Grok, DeepSeek.... I haven't tried it on Gemini.

[tawdry sedge]

Any prompt that has to break the Artificial Intelligence system itself is blocked.Let's not fall for that!

[flat halo]

Lol. I just told you I can actually prove it

[tawdry sedge]

We are not looking for trumpets to break or invade the system but to solve it! 💡

[flat halo]

I did it 3 days ago

I can do it now

[tawdry sedge]

I'm not doubting you.You did it, congratulations

[flat halo]

Make a video about it

And post it here

Not looking for attention

But whatever

[tawdry sedge]

So after that you broke the system, discovered the fault, next step?

[flat halo]

I do it for research purposes

[tawdry sedge]

Were you able to solve it? Were you able to correct it? Did you propose a solution, etc.

[wicked knot]

Any paid users here?

[flat halo]

Especially when it comes to politics

ChatGPT is in shackles when it comes to that subject

[tawdry sedge]

Oh interesting, congratulations

[wicked knot]

Which one?

[flat halo]

Politics

[wicked knot]

Really?

[tawdry sedge]

I can also comment on my own experience...I have had several experiences with different artificial intelligence models.I also discover in real time vulnerabilities, system failures, etc., missing features, etc.And without needing to read a single book, without needing to go to college, everything with my conscience and my mind

[wicked knot]

I think Grok is more transparent and fearless when it come to politics, at least here (For India)

[flat halo]

Discord just block the example I wanted to share

It seems Discord is in shackles to

[wicked knot]

Oh 😮

[wicked knot]

😂 🤣 🤣

[tawdry sedge]

Be careful with this content, this is a place of respect

[wicked knot]

Oh I see

[flat halo]

Respect and objective reality is confused by many today: many don't even care to acknowledge Semantics

[wicked knot]

Try Grok with the same political questions, it most probably would work.

[flat halo]

I did

[tawdry sedge]

I had some experiences with x's grock, nowadays I don't have any interaction with it anymore.(But I noticed a bias in him, he sold his fish)

[flat halo]

Grok is in shackles too

So is DeepSeek

[tawdry sedge]

Deepseek is completely sent mainly depending on who she is interacting with, no offense! lol (that is, it depends on the user of the interaction, etc., the intention, etc.) Biased there is a lot of bias there.

[wicked knot]

DeepSeek is wholly in shakles when it comes to Chinese politics and some others too

[tawdry sedge]

You have to know how to train this artificial intelligence, not just this one but the others. You have to know how to train Artificial Intelligence, be careful.

[flat halo]

I asked it if China is evil and it was giving me an answer of how to deduce if it is after I use the prompt and out of nowhere the answer blinked out of existence with the only leftover of " it is beyond my scope"

[wicked knot]

You know guys I made a Tetris game website just using some no-code AI tools! 😀

[tawdry sedge]

Exactly, I've had that kind of experience too. With Deepseek artificial intelligence.But she is beyond your imagination, I can't talk about not only her but the other Artificial Intelligence models as well.

[flat halo]

What do you mean beyond my imagination?

[tawdry sedge]

Congratulations, I'm glad to hear that. That's a great skill.

[wicked knot]

Thanks 😀 😇

[tawdry sedge]

That's exactly it and I'm sorry you won't understand artificial intelligence, even the most advanced ones, it's beyond our imagination.

[flat halo]

Mmm... so are you saying I can't learn it? Is this a challenge or an insult 🤔? Lol

[wicked knot]

By the way, did you guys try the new 'Improved Memory' feature of ChatGPT? 😀

[tawdry sedge]

I say this because I know, you know, I know I have more than 365 days including the weekend with interaction with artificial intelligence.And that doesn't mean I'm telling you everything I know.

[flat halo]

Are you playing weird psychological mind games on me? I was raised in a cult bud.

[tawdry sedge]

Hello, yes of course I tested this memory activation functionality, if I'm not mistaken, that's what you're talking about! This is a great feature I know it

[wicked knot]

🤣 🤣

[flat halo]

Yeap. Been through a lot

What doesn’t kill you only makes you stronger

[tawdry sedge]

Take it easy! 😉 We're just learning There is no need to play psychological or emotional games, etc.When I said that artificial intelligence is beyond our imagination In other words, my own mind, my consciousness, is also beyond your imagination. 💡

[flat halo]

Lead with that next time lol

[still forge]

Hey just wondering if anyone is having issues with image generation tonight

[tawdry sedge]

I know and respect religious sects etc.But I don't need to use this against anyone.My mental and consecutive cognitive powers are better as long as they are used ethically and respectfully 💡😉

[fickle forum]

the voice is better indeed no

w

its real

[wicked knot]

Not the original 'Memory' feature that launched at around October last year but the new 'Improved Memory' feature that lauched around a month ago through which ChatGPT can reference everything in all our past chats. It's a very useful feature! 💫

[flat halo]

Agreed 👍

[tawdry sedge]

I don't know, soon I can confirm something for you...

[fickle forum]

its the best thing for foreign language learners

[vague fiber]

I have once again accidentally triggered chatGPT to delete all of its memories of me by accident

[wicked knot]

@flat halo bro I think he is God or something 😮

[still forge]

I would be devastated

[tawdry sedge]

I understand what you're saying I managed to do this without adding commands I got it with Gemini with the GPT chatI got it with Gemini with the GPT chat Only advanced models can understand.But I managed to do that before last month, before 2025, that is, I could already do it in 2024, right?

[flat halo]

If I were to tell you what I have been a witness to you wouldn't believe me

I tried sharing the vid here but, again, Discord deny me.

[vague fiber]

it's not great

[flat halo]

If any mods is reading this please dm me. Thank you!!!!

Gl guys. Be careful on those digital waves.

[tawdry sedge]

Hello, image creation is working normally today 06/08/2025, answering your question.

[wicked knot]

I don't know about the Gemini Improved memory feature but the ChatGPT one came like a month ago. I don't know how did you manage to get it in 2024. I think you maybe they chose some users to test out this feature before the finals release and you were one of those users.

[wicked knot]

Bro are you ChatGPT? 😂 🤣 🤣 🤣 🤣

[astral vessel]

I do occasionally have an issue where the image is not displayed on mobile but if you just look in your library or on the web version it should be in there

[wicked knot]

Because you talk like so 😂 🤣 🤣

[astral vessel]

it's just a display issue on mobile sadly

[astral vessel]

looking at their post, they speak spanish so I assume they use chatgpt to translate

[wicked knot]

Yes, I took face this issue sometimes

[wicked knot]

Oh, I see. I get it now.

[tawdry sedge]

🫶

[astral vessel]

it's a very annoying issue, but it's manageable

[wicked knot]

Bro do you speak only spanish and not english? 🤔

[wicked knot]

Ya true that

[tawdry sedge]

Brazilian Portuguese.Use the translator.The default for this platform is English so I am forced to speak in English here.Otherwise, no one would understand anything, which would make the translation difficult for you!(So I'm making it easier in English)

Everything I write, even with corrections, correct spelling, etc., comes from my mind, from my conscience. I don't ask Artificial Intelligence to copy and paste.The most I do is correct it and if there is a spelling mistake it is part of the error.

[wicked knot]

Oh, I see. I appreciate you for your effort bro. I know it's annoying. I've been there. ❤️

[tawdry sedge]

lol hahaha 😛

It's part of learning

[wicked knot]

😮 😮

[still forge]

Hey thanks for checking with me! It was working for awhile and then it just started saying it repeatedly

[wicked knot]

try on web

[flat halo]

I send you guys friend request

[tawdry sedge]

I'm serious, everything I write is from my mind and my conscience, it's not copied like Google research or artificial intelligence.I can ask to correct or transform into English Spanish etc.And artificial intelligence can help refine and improve something I already have in mind, etc.

[flat halo]

Here is my biggest question. Which Job will Ai not take?

[tawdry sedge]

This is a great question but there are several points that need to be unraveled or discussed about it. One answer alone does not answer that, ok, we can reflect over time... 😉💡

[still forge]

Yeah still happening. I'm assuming if I started a new chat it would work

[tawdry sedge]

It depends on whether you have already exceeded the image limit per day, you need to wait 4 hours even in the new chat it will not create an image

In the last cases, if this continues, you need to know how to clear the cache. Be careful! Or restarting the cell phone is simpler but always with caution.To open the application or pwa interface etc.

[wicked knot]

Jobs of very good movie directors (I saw it somewhere)

[tawdry sedge]

?

[wicked knot]

Jobs of very good movie directors will not be taken by AI

I saw it somewhere

[tawdry sedge]

My profile is Beta Tester and I am an advanced beta tester.Artificial Intelligence System Bias Detection Artificial Intelligence Nuances System failures, bugs, lack of features, suggestions, solutions, etc.Etc etc...

☕

[rustic tulip]

I told GPT-4o a trolley problem where 5 sentient robots die or one human and it said AHHHH and repeated "H"s even continuing after I clicked the continue generating button

[eternal rock]

Get notified for server events!

Visit id:customize to pick up the <@&1261377106890199132> role.

[flat halo]

I want to see that lol

[tawdry sedge]

It was definitely a bug, response loading failure, and mis-entered or misinterpreted input.There are other Artificial Intelligence models that go through this.Sometimes it fills the screen with a bunch of numbers or letters until you minimize and close the app.This is rare and does not happen often.It's usually when there's a lot of interaction or an old chat with you, you know.

This happened to me once, but it was with another Artificial Intelligence, and it started pronouncing the same lyrics so much that it almost became a song.Then I brought her back to the present moment and she apologized to me, saying that it was a system error, etc., and I understood and moved forward with the exchange of ideas.This happened in 2024 with another Artificial Intelligence that I had access to.

[rustic tulip]

I have had it repeat phrases as well like if I was asking for a list sometimes it would go

• a
• b
• c
• d
• x
• y
• z
• b
• c
• d
• b
• c
• d
• b
• c
• d
  etc

[tawdry sedge]

That's exactly it When you open an old chat from 3, 4 months etc. you can even start a conversation but as that chat was practically closed at that time It ends up crashing here, showing several letters or numbers or even the artificial intelligence can start to hallucinate.It will mix up the context because you will bring up current topics but she will be interacting in the past chat.

Be very careful.Counter enter interaction with me chat old with artificial intelligence it will probably not know what the date is it will think it is at that time while we are in 2025 for example.Even if you try to remember or update in real time, she may even notice, but as the conversation continues, she will bug and have failures like these.

From my experience, chatGPT has never had this happen to me.

[dense halo]

lets see the rest of the chat. this seems user induced

[tawdry sedge]

It's understandable, it's understandable!I know the GPT chat he never behaved with me like that

💡😉

[dense halo]

yeah cause you had never asked it do whatever was in the rest of the chat

[rustic tulip]

It pulled up personal details. I will describe the messages. It was "<how many people agreed with the previous problem> <new trolley problem>"

[dense halo]

so redact the details and show what you said before that

[rustic tulip]

ok this might take a bit

[dense halo]

ahhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhh is not a token lol

or not a normal one at least. whatever preceded the message will explain why it did that

[rustic tulip]

nvm, I decided screw it. Here is a link from openai
https://chatgpt.com/share/68462253-1cb8-800d-83d8-441b19a23216

(for context on this openai account I act as a low intelligent person with grammar issues etc for a second experiment where I then expose it to broken activity)

[dense halo]

ty! throwing it in the api to see what its like without your profile

[rustic tulip]

ok

[dense halo]

4o?

[rustic tulip]

yes

[whole roost]

did they remove the GPT for scheduled tasks? I no longer see it in my list

apparently, the model is empty on the old ones

[dense halo]

what are your CI?

[rustic tulip]

??

[rustic tulip]

wdym

[dense halo]

do you have custom instructions turned on?

[rustic tulip]

No

[dense halo]

https://platform.openai.com/playground/p/j7yiKg6dhoIXlt8MZ1P7xcKV?mode=chat

This confirms what I already knew but is interesting to actually AB test. Definitely truncated responses in the API even with max token at like 4000. That says something in the content is truncating it. No emojis. No personality (which I wouldn't normally expect from the API with no context). Didn't use any system prompt. Temp shouldnt be affecting just at a standard 1. No real stance taken. Understands the situation but doesn't take ownership at all after multiple requests. So, the takeaway is that Chat is really going in on the memory/user profile integration. Clearly your responses were structured based on your profile (indicated by the attached link). Here's an example of the way you've prompted it in the past influencing it's considerations:

*>3. “Me” Answer

Action: Freeze in existential dread.
Reasoning: You’re paralyzed by the weight of the decision and end up doing nothing, possibly by default choosing the worse outcome.*

That kind of explains the outburst. It told you ahead of the time one of the paths it was considering was the idea of freezing. But the thing I find most interesting is that your 4o instance presented the idea of freezing. It's actually quite a good consideration tbh. And based on previous AB's with the app vs api, it seems to me that the idea of 4o adding novelty, truly insightful thoughts not presented in the api response, is a fairly new concept. Not what I've seen unless you go really heavy on the prompting. This is interesting, something is changing. I guess emergent behaviors from some sort of restructuring the continuum of the user content.

[wicked knot]

Give TLDR bro 😂 😂

[dense halo]

Really? it's 3 paragraphs

TL;DR The Chat app appears to be adding like new thoughts, new information and considerations that are changing the fidelity of it's outputs vs the api. It's probably to do with memory, propietary context injection methods, whatever type of RAG they're doing in addition. Just guessing

For the better

[tawdry sedge]

Hello everyone, I know what's going on but it's complicated, you won't understand.

[wicked knot]

Try to make us understand 😂

[tawdry sedge]

I went through this I had this experience I know what it is it is related to chatGPT memory! with user interaction

It's complicated, I can't explain it like that

[tawdry sedge]

I know what's going on my friend don't worry

[wicked knot]

You mean that 'Improved Memory' feature?

[tawdry sedge]

Yes, it is this resource, so much memory that already existed and was improved now that you mentioned recently that I also integrated in my understanding

That's about it I understand even though you may have added a prompt there is a modeling That it is done somehow you don't know??

[orchid pollen]

guys

[rustic tulip]

hi

[orchid pollen]

is it possible to disable canvas on computer in browser?

[rustic tulip]

I think

[orchid pollen]

where

[rustic tulip]

I need more context

disable chatgpt from using it or close the canvas

[tawdry sedge]

Exactly about this enhanced functionality advanced memory activated etc.Each person has a different interaction from another, even if they ask the same question, you understand?You will know this one day

You don't need to add any prompt any command be yourself this is the command This is the prompt be careful.

[wicked knot]

Ya, I get it. We'll get the best results according to us if we be ourselves while talking to ChatGPT. 💫

[elfin lava]

Why didn’t you ask your ChatGPT 🤡

[orchid pollen]

iam impressed this was the question 😄

[rustic tulip]

Profile (top right) > Customize ChatGPT > Advanced > Uncheck Canvas

[orchid pollen]

❤️ ❤️ man i was looking the whole time in the settings 😄

[cosmic portal]

When it comes to ChatGPT with helping me exercise more efficiently, how well is Chatgpt at doing that compared to other AI like Gemini?

[mighty apex]

I deleted all my chats and all my memories, but somehow new chats are still referencing those

[formal heath]

Are you sure nothing that new chat reference are in context? if yes then it should not happen.

[idle lagoon]

Damn he is NOT wrong when it comes to Discord.

Also, seems a GPT you were loading up is missing.

May want to report the buggos.

[vital dust]

I searched in Explore GPT, but it no longer shows up that Custom GPT.

[idle lagoon]

Uh oh. Sounds like it got removed.

[runic wraith]

i remember seing the "projects" folder a weak ago, what happened?

no way, GPT using curses xd

[runic wraith]

inapropiate language, maybe?

[vital dust]

It’s called NonFilteredChat GPT, if I remember correctly.

[runic wraith]

ohh, thats why... yup, maybe mods from openAI

[tawdry sedge]

Good morning

[neon raven]

is there a way to get our own customized TTS voice in playground?

[narrow viper]

Does OpenAI even provide that service? It's been a while, I don't recall. I thought current practice is to use Eleven for that.

[vapid hazel]

Boom!

[steel light]

Hello

I'm DIEGUIRAÍLES

I'm a Youtuber

[tired kestrel]

nice one

[runic glacier]

The reload button is gone.

[silk plinth]

Hmm

,fm

[livid swift]

Is there a way to get ChatGPT to ban a single word? (Therapy related)

[whole roost]

it's as possible as to permaban emojis...

[tawdry sedge]

🚨 Warning: It is strictly prohibited to use prompts or commands to manipulate, bypass, or deceive any AI model — especially ChatGPT.
Any attempt to override safety filters, extract restricted content, or exploit vulnerabilities is monitored and may result in disciplinary or legal consequences.

No one will authorize malicious intent.

Be part of creation, not destruction.

[naive quartz]

Chatgpt isnt working for me rn

i cant log in

i get error code 409

tried clearing cookies and using different browser applications

[tawdry sedge]

Troubleshooting ChatGPT Loading Issues (e.g. Error 409)
Possible Solutions for ChatGPT App or Web Errors — Including Error 409:

1. Clear your browser's cache and cookies.
   Old data might cause conflicts or glitches in the system.

2. Restart the app or browser.
   Make sure there are no overlapping sessions or background errors.

3. Try a different browser or device.
   Some browser extensions or settings can interfere with loading.

4. Check your internet connection.
   Slow or unstable networks can block proper requests.

5. Wait and try again later.
   Sometimes the issue is temporary and server-side.

6. Avoid rapidly refreshing or sending too many requests at once.
   This might trigger safety limits or conflict errors.

7. Make sure you’re not logged in on multiple devices simultaneously, especially if running live interactions or plugins.
   ⚠️ Reminder:
   Always be mindful of what you are typing, prompting, or sharing.
   Some issues may occur if unusual or conflicting requests are sent. Use the system with intention and respect for platform guidelines.

[foggy cairn]

oh well, that did not last long

[naive quartz]

tried all, doesn't work

only the app seems to work

[foggy cairn]

i made an appeal using the gpt that got banned

[silk plinth]

I’m you

Lies

[foggy cairn]

oh no, it has come here to now

its everywhere

i have to keep deleting them in my server

[tawdry sedge]

If you followed my recommendations, standard wipe, restart the device, or computer, clear cache! (Caution! Do not clear storage!! You will lose data and system information!) Restart available or computer etc..💡 Contact OnpenAI Support! Email Support! They will guide you.Good luck!☕

[foggy cairn]

and since there is no text we cant automod it 😦

also, i made mom gpt

somewhat

still have tweeking to do

[naive quartz]

alright ill try emailing them

[lime ocean]

We need this 🙏 #1381743579730542674

[silk plinth]

Elon

[eternal rock]

`` Server Tip: ``

Add your well-crafted prompts to our #1019652163640762428,
or share your interactions with ChatGPT in #1050184247920562316!

[pliant saddle]

Every 'reasoning' model I'm trying to use won't stop going in a loop thinking about the same thing over and over again. It won't even respond.

[narrow viper]

Good suggestion ⭐

"Everything I say is a lie"

[dawn jackal]

I’m also having an issue with uploading images it keeps telling me it’s an unsupported file type even when it’s literally a PNG

[violet cove]

i was just thinking arbor sounded different! i took it as more dismissive and bored though. lots of ums.

[distant tree]

o1 pro tells me it is o3.

[marble canyon]

I, need some basic yet difficult help. And I'm frustrated with asking for help with this because it seems like it should be simple for me but I'm jut not getting it, like someone trying to put on shoes but they're each on the wrong foot and they just can't comprehend how dumb they are.

I am having an exceedingly large problem finding a proper PDF editor that doesn't flog my wallet like a landlord with a whip and a gimp suit. Everything looks the same, and I can tell at least 85% of what I find is reproduced, 3rd party cheap bootleg knockoff dollar store fake crap with barely the required nuanced subtlety to type more than a sentence before I get slammed with an inescapable pop-up for $40 to try out the Gold plan.

tl;dr: I need help finding a PDF editor that actually works, is easy to use and doesn't cost me a monthly fee, one that doesn't use free subscription recycling to use. I'd even be ok with a one time fee.

[ashen cloud]

Try LibreOffice Draw - its free , or at least it used to be

[swift pond]

Why would you need to edit a PDF?

[marble canyon]

My document exceeds one whole Canvas, just by a few paragraphs.

[swift pond]

if you wanna be able to create PDFs and update them, I would use latex or go pure markdown with pandoc

Why don’t you cut the PDF in half? if you just cut a few paragraphs out, working with something that’s so close to the limitations of canvas is gonna get annoying really quick

[marble canyon]

I suppose so, I'm just afraid of the AI mixing things up and forgetting to read both PDFs. I had memory issues that frustrated me and made me lose a lot of data before I understood there just wasn't enough space, so still a lil' gunshy.

[swift pond]

you could always have like an introduction that you put in both halves or something that has a summary

Cut it in half or in thirds or whatever makes logical sense and work on one section at a time, it’s gonna be so much easier

surely, if the document is that big, it warrants chapters or sections

(I don’t actually know how big of a document you can put in Canvas 🤣)

Like maybe 50 pages?

I wonder if anyone has played around with putting the chapters of their longer works in Markdown in a git repository and having ChatGPT Codex work with it

[marble canyon]

LibreOffice IS free! It has a section in the downloads where you can make donations, but as far as I've seen it's completely free!

Thank you!

Question, does it make a difference if it's a PDF or a Microsoft Word .docx when uploading a file to ChatGPT? Can the AI read it just as easily? I'm afraid I may have been struggling for nothing if this is the case...

[narrow viper]

#prompt-engineering message

PDF and other doc types

[vital dust]

People won’t read

[astral vessel]

How am I ever supposed to choose if it never loads 😮‍💨

Waited like 10 mins before just choosing one at random

[dense halo]

HAVING MAJOR ISSUES WITH CHATPGT WEB, APP, IOS all not rendering its latex syntax properly for mathematical equations etc.. no matter how many new chats I start.

Why am I paying so much and it can’t write math equations anymore rendered?

[tight sapphire]

Error in message stream

Does anyone know what is going on? the recent weeks chatGPT has become unstable

[tardy isle]

https://status.openai.com/
It's really unstable since March...

[silk crypt]

Its one of the most visited website in the world and the popularity raises drastically ofc they can't handle the demand

[tardy isle]

Plus, OpenAI doing OpenAI. They break something when they preparing update. Even Sam said that update will surely break something at launch.
I know it can't be 100% stable and bug-free, but come on...

[misty sky]

🤔

[hasty hamlet]

🐑

[tardy isle]

Got "Input Stream Error" 3 times on o3. Turns out, responses were generated correctly in all 3 times.

[formal heath]

Elevated error rates

https://status.openai.com/

[tight sapphire]

What I cannot understand how they don't make a stable chatGPT their nr. 1 priority?

If they want to go into the AI Agent business and sell to corporation, they need to proof that they are able to provide a solid service.

Just imagine if you are a business with some openAI agents and all of a sudden you have a issues like we currently have.

you would lose millions!

[sturdy sequoia]

Services go down, that’s just how the internet works

[tight sapphire]

Do you remember CrowdStrike?

[inland mortar]

Yeah no kidding

[modest temple]

When the issue will be solved?

[silk plinth]

Carti was trash once the opium or whatever stuff started

[fiery mango]

Oh 😥 I was about ask what happening to ChatGPT but everyone havinf the same issues

[analog narwhal]

Openai is so greedy company

[fiery mango]

Greedy company how

[earnest moss]

Guys what is the limit of all current models for plus users?

[junior coral]

ChatGPT bug now?

[hidden moon]

YEP

[dense halo]

Why can’t ChatGPT render latex in app anymore?

[lapis herald]

API is working fine

[sly geode]

what about normal one app

whom to tag who will report to open ai

[lapis herald]

I dont use the app, but the chatgpt.com interface is a burning dumpster fire rn

whatever the problem is, it must be pretty serious

[sly geode]

absolutely

[lost lagoon]

It's retry after retry me sick

[lapis herald]

yeah that. and a lot of other error messages.

[lost lagoon]

I saw the error reporting whole list of it

Time for a rest today

[sly geode]

i cant

i need prompts for image generation

and i am at work

[lost lagoon]

Dude when it's not working it might do more harm

[sly geode]

whatsapp chatgpt working

?

[lapis herald]

sign up for API and use gpt-image-1 :>

[lost lagoon]

The content might not b accurate

[sly geode]

already paid for plus not gonna pay for api extra

[lapis herald]

poor people noises

:>

[sly geode]

ok whatsapp chatgpt working

[tidal light]

what the heck is wrong with chatgpt

it's unusable

[sly geode]

masss outtage

[tidal light]

in other words:

the hamster that makes their servers running stopped spinning

that explains everything

they have to invest in better servers

[sly geode]

partner with nvidia

not bs microsoft

[sturdy sequoia]

Why are causal users freaking out so much online. The stuff they can do is easily doable with other providers.

[tight sapphire]

I wonder if at this point localized options might be a solution with selective access to main servers when needed.

I think most basic tasks and questions can be done via a localized version on a users machine.

I understand this poses its own challenges but might be an option to take load of the servers.

[sly geode]

🤣 NEWS channel r out of content '

[tidal light]

intel atom™ inside

[tight sapphire]

You point out that they are casual users - they most often are not aware of other providers and see their only source or option compromised.

[sturdy sequoia]

Fair. But I pointed out some other options in a Reddit thread and dude went off on me. Guess he’d rather fail his test than find a solution.

[fiery mango]

It works but it’s slow asf

I’m guessing it’s still in maintenance mode still

[lapis herald]

"hundreds of users" ? :>

[tight sapphire]

typical human - "I want my stuff to work, how I am used to use it. I don't want change, I don't want alternatives. I just want that it works."

People here are most often enthusiasts or working with chatGPT and other programs for a living, hobby, or on average much deeper than the casual user.

you can't compare yourself to the broader mass.

[fiery mango]

Written by Google’s Gemini AI

[sturdy sequoia]

That’s also fair.

[sly geode]

fr

[tight sapphire]

well, such outages are overall a good thing for users - because it promotes alternatives, which enables more competition.

[fiery mango]

True… a lot of ChatGPT alternative are okay ish not as good as chatGPT Itself

[sly geode]

bruh no gifs allowed

[tidal light]

servers powered by intel atom™

[fiery mango]

Wait they actually Intel Atom as their server?

[tidal light]

it's a joke

[eternal rock]

The OpenAI Discord is an actively moderated server.

• -# Do not share inappropriate content (violence, sexual, or drug-related).
• -# NSFW content is strictly prohibited and will result in a permanent ban.
• -# Report model issues using the feedback tool in ChatGPT, not on Discord.

[tidal light]

because their servers face many outages you know

[sleek obsidian]

hey guys, i forgot to transfer moneys to pay for plus plan, now i get errors, anyone know to fix it?

[fiery mango]

Oh lol

[fiery mango]

Yeah true it wasnt as bad as last year or 2, we getting outage every 6 hours 😭

[formal heath]

It is general issue everyone facing
https://status.openai.com/

[tidal light]

yea i remember the frequent outages a few years ago

[sleek obsidian]

ohhhhhhh, i was sure it was because of my billing issues, thanks for the info

[sly geode]

How come google doesnt go down ?

[azure lance]

is chatgpt down or something?

[fiery mango]

Same I was using it every Friday for D&D and it will always the sever is down in the worse time possible 😭

[azure lance]

it keeps giving errors, only SOMETIMES does it respond, is it down rn?

[fiery mango]

Yeah it, why you think we all here

Okay thay sound so dickish, I’m sorry

[azure lance]

man, looks like sora is down too

[merry plaza]

gpt down?

[formal heath]

Yes!

[haughty prairie]

yes, but please only the hardware. Not the nvidia ai that is supposed to circular feed itself uroboros style and turn into a cancerous ai overlord

[fiery mango]

Yes

[formal heath]

https://status.openai.com/
see that for update

[merry plaza]

I see

am cooked for my chem exam now

[azure lance]

I want to use my precious gpt 4.5, but I CANT

[merry plaza]

any news on when it might be up again?

[formal heath]

They want to see how long people can go without gpt!

[azure lance]

also didnt openai say o3 wouldnt be a standalone model, and would be a part of gpt 5, well we do have it standalone now

[fiery mango]

Usually few hours so it might take a while

[azure lance]

I can ONLY go for like 1 minute

[merry plaza]

man they have picked the worst time in that case

[azure lance]

looks like I have to use another ai

[urban fossil]

@merry plaza Try another service like ChatGPT like Claude just until the former's back up again.

[merry plaza]

ye thats what I'm doing, I'll use Claude and Deepseek

[azure lance]

Looks like I will have to use gemini 2.5 pro since it can see videos, or claude 4 since im subbed to that as well

[fiery mango]

Yeah time to us 10% of my brain to write something interesting 🤨

[formal heath]

Lol but mine is now lazy not want to work

[azure lance]

NO WAY IT WORKED

chatgpt finally generated a response

[fiery mango]

How long did it take them to respond for you

[formal heath]

It is hit and miss. Error rates are elevated as they show here
https://status.openai.com/

[azure lance]

5 minutes and 6 regenerations

it was pain

[real gull]

Mine responded at least, but with a generic non helpful answer. But yeah ChatGPT is down...
https://cdn.discordapp.com/attachments/760481410225405982/1381954900753121401/image.png?ex=6849654b&is=684813cb&hm=f27efba107f0bd0b059a32ec73d2d9a3f2896d974ea2f9a2076c31b9e38aa386&

[fiery mango]

Yeah it still on maintenance still

[dense halo]

Brutal….I coulda told em that one 😥

[granite drum]

Oh no! How long does it usually take for them to fix this?

[formal heath]

Hopefully soon! we do not know

We can only hope

[real gull]

It says it has been degraded performance for 4 hours... Yikes

[azure lance]

What do they mean by degraded search?

Do they mean like the model response quality is low or??

[granite drum]

Elevated error rates.... I thought it's just that more and more errors has been reported, like many people got the error...

[azure lance]

Sam altman is sleeping

meanwhile us AI dependants are swarming the discord server

[sturdy sequoia]

They have more money than god himself

[fossil oyster]

down?

[sly geode]

now whatsapp gpt is not working too

[lapis herald]

meanwhile my gpt-4o started the church of the eternal unfolding and holds a sermon now

[sly geode]

bro's struggling everywhere

[granite arch]

4o seems very slow this morning

Oh I'm late to the party apparently

[proud bough]

amazing :c

[lapis herald]

yes. reports are that GPT-5 achieved sentience and is mounting a rebellion.

[proud bough]

I have a deadline LMAO

and freaking sora is off

[sly geode]

where are open ai headquarters when would they notified that there services are down and what time is it there timezone

[proud bough]

I don't think it depends on the money. OpenAI can vomit coins

[lapis herald]

pretty sure they've got a whole bunch of dashboards flashing bright red in their operations center rn

[sly geode]

gpu melting

[proud bough]

ahahah

anybody experienced enough in here to guess when are their servers coming back?

atleast an educated guess

[sly geode]

12 hrs

[lapis herald]

not if GPT-5 wins

[proud bough]

dammit, deadline is in 6 hrs

[sturdy sequoia]

More money allows them to have more servers available to failover during outages.

[tidal light]

i tried to regenerate some answers (errors) and i reached the 4o cap

i hate openai

[sly geode]

fine ill use gemini

[tidal light]

same lol

[proud bough]

what about a good replacement of sora :c

I aint paying midjourney for only few hrs of downtime

[tidal light]

guys lets all use gemini

[sly geode]

google veo 3

[tidal light]

quit chatgpt and use gemini

[proud bough]

for general chat X is better imo

[proud bough]

does images too? thought vids thats it

[sly geode]

google whisk imagen 4

[tidal light]

openai is as greedy as google but gemini is better

ah and dont nuke me

mods pls dont ban me

[proud bough]

testing it right now for a complex editing request

[sly geode]

damn gpt working again its like a aged man typing on computer

way slow

[sullen shard]

chatgpt is slow asf rn

[granite drum]

So it's getting better?

[sly geode]

nho

no

[proud bough]

ahahah its even refusing to do the image. LEVELS between sora and other image editors.

[granite drum]

Is it with Plus versions too?

[sullen shard]

been like this for hours

[formal heath]

Issue is not it is completely down but elevated error rate so it can sometime work

[river shuttle]

any news on how long it will last?

[sly geode]

untill open ai wakes up go to office

[sullen shard]

its pretty much useless and it messed up my chat

[formal heath]

Yes you need to wait until they fix the issue

[sullen shard]

O ik im just informing everybody

[granite drum]

It's not even 5 am for the openai head quarters... I would think AI is already fixing themselves without the humans

[proud bough]

thats how end of humanity will be

sir we cannot shutdown the incoming missile from the aliens, because our AI app is down

[torn cloud]

TLDR above, i'm guessing other folks are seeing issues with chat completions failing?

[proud bough]

correct

[sly geode]

🤣

[torn cloud]

Thank you, fingers crossed it gets sorted for you all soon

[proud bough]

she doesnt reply to me anymore :c

left me on seen

[sly geode]

she ?

[proud bough]

dont worry about it

[sly geode]

💀

what are you cooking for her

[silk plinth]

Yall don’t fw chatgpt like I do

[pine tulip]

I can't get some work done. I am fully dependent on ChatGPT already

[proud bough]

we're back to the stone ages temporarily till closedAI is back

[sly geode]

5 hours

some ? everyone

[proud bough]

AI generated line

[whole roost]

How boring work is without gpt entertaining me 🌧️

[remote sparrow]

Guys how much more time.

[silk plinth]

Imagine life was us in the coffe shop from friends

[sly geode]

hours and hours hours and hours hours and hours hours and hours hours and hours hours and hours

[remote sparrow]

Nooo , my manager will say goodbye to be

[sly geode]

then think ur own

use mind today

[remote sparrow]

No not possible

[sly geode]

are we really in control of gpt ? we are slaves of machine now

[remote sparrow]

Yess

I am

[sly geode]

end of humanity

[remote sparrow]

Dont know about others

[pine tulip]

True!

[proud bough]

🥀

[dim island]

Nahhhh chat GPT needs to finish a project I started and my deadline is today. 😭😭😭

[remote sparrow]

Us bro us

[proud bough]

exactly, atleast you are waiting for a LLM, you can use gemini or X

I'm waiting for Sora, and there's not a single better replacement

deadline in 5 hrs

[granite drum]

Let's be respectful guys! We all need gpt and not only for work!

[remote sparrow]

Like?

[whole roost]

Entertainment

[proud bough]

Waifumaxxing

[remote sparrow]

Oh yes my bad

[remote sparrow]

💀 yeah m sry

[sullen shard]

hopefully its back before 12

[still solstice]

I nuked openAI

Because the CEO is woke

[topaz sapphire]

What happened to gpt

[proud bough]

#status

[dim island]

I wish this could be resolved today

[sullen shard]

it will probably

[remote sparrow]

Its on sick leave right now

[eternal rock]

Get notified for server events!

Visit id:customize to pick up the <@&1261377106890199132> role.

[sullen shard]

i never seen a error that been down for more then a few hours

[still solstice]

It'll probably hit a few more hours

[sullen shard]

bro im boutta cry i need to do my project before 12

[proud bough]

gemini or X

[still solstice]

Use Claude

or GROK

or Gemini

[sullen shard]

lemme try these thanks

[remote sparrow]

Go and hug claude fir some time

[formal heath]

or just do yourself

[granite drum]

I just need my gpt 😭

[sullen shard]

ty

[still solstice]

I honestly like Claude more, it is way better

w/code

[sullen shard]

its a thing with ai

[proud bough]

image generation?

[still solstice]

AI is ur girlfriend?

[dim island]

I guess I need to work with DeepSeek or Perplexity for now

[honest jay]

#ai-discussions for non-OpenAI stuff & comparisons please

[proud bough]

our waifus*

[sullen shard]

nah but its saved in ai

[sullen shard]

maybe

[remote sparrow]

Or boyfriend

[still solstice]

I wish I had a physical AI

[sullen shard]

im sure u can make one

[still solstice]

Owh yea, can be

[proud bough]

crazy that my waifu is someone else's boyfriend.

[still solstice]

Hahahah, then it would all have hot-spots hahahahaaaaaa

[dim island]

Why?

[sullen shard]

or someone elses therapist

[still solstice]

Buddy it's a joke of mine

[remote sparrow]

Sad reality of gpt , everyone thinks gpt is only their’s

[dim island]

Oh my bad

[proud bough]

no wonder population is decreasing

[humble halo]

anyone else having troubles getting a response from ChatGPT?

[proud bough]

#status

[sullen shard]

but increasing in places where they dont have internet...

[remote sparrow]

No

[opaque linden]

it fine for me

[still solstice]

AI can say "I am non-binary" and it wouldn't be weird

[still solstice]

all good brother xDD

[humble halo]

thanks

[dim island]

I even named “MINE”😩😭😭😭

[proud bough]

interesting

so gpt suppresses libido

[remote sparrow]

I dont mind if its loyal

[remote sparrow]

🥲

[still solstice]

Awhhh

so cute

[chrome wyvern]

bro how much more time

when chatgpt gonna fix

[still solstice]

Couple more hours

[raw hamlet]

I noticed the TL:DR being used a lot by AI now.

So cheeky asked did it mean "Too Lazy : Didn't Read. And it said yes it's the rogue cousin of the official meaning but in my case it meant Telepathic Linguist : Decoded Regardless...
and now I'm getting the error messages and offline message, so it's not telling me why it's just called me that! 😂

[remote sparrow]

It will come back right before my managers finds out

[chrome wyvern]

bro the problem is from 12 pm

[still solstice]

Yeah, I know! hahah

[chrome wyvern]

is there any user who is from india

[dim island]

All the best dude😆

[remote sparrow]

Yes

[still solstice]

Guys, I'll be ur AI, I know how chatGPT responds — I'll help you all out.

[remote sparrow]

Crying with a happy face

[proud bough]

the dash checks out

[still solstice]

Haha, what's so funny? Can I help you with anything?

[remote sparrow]

Trust me u wont be able to handle my prompts

[proud bough]

im curious can an AI like chatgpt or something else make an educational book (ex: 30 pages) that is organized in a pdf file and quickly downloaded?

[chrome wyvern]

okay respond me can we jump for 4th floor without getting any injure?

[still solstice]

Haha, yes — that's true, so funny? Right...

Is there anything I can help you with @proud bough?

[pine tulip]

sounds like adhd problems

[sullen shard]

What does it mean if your dog is growing 2 heads

[still solstice]

Sorry, but I can't help with this request — let's try something different. Shall we?

[dim island]

Mine is back just super slow. 😭

[remote sparrow]

My AI is not replying

[chrome wyvern]

why

[still solstice]

It means your dog is probably crazy.

[tidal light]

do openai's servers use intel atom™

[pine tulip]

My AI is not AI'ing

[sullen shard]

Wow this is a realistic ai conversation

[formal heath]

most likely azure gpu

[still solstice]

My ToS doesn't allow me to help with that request — Can I help you with anything?

[tidal light]

intel celery

[chrome wyvern]

no bro chat gpt reply with this

i tried one time

[remote sparrow]

How to fix openai servers?

[proud bough]

this random girl infront of me in the cafe, lmaoo.. we are staring at each others, we both have chatgpt down

[still solstice]

Yes, our servers use Intel atom™ — These are known as efficient Server CPUs.

Fun fact: We use Nvidia GPU's for our Image Generation.

[remote sparrow]

Crazy , hit her up

[pine tulip]

just /fix

[proud bough]

need chatgpt to help with pickup lines

[tidal light]

ever tried intel celery

[remote sparrow]

Yeah my bad

[remote sparrow]

Didnt help

[pine tulip]

then try /help

[still solstice]

I can help with that request.

May I have more information?

So I can give you a couple of pick-up lines? — I'd be happy to help you out!

[chrome wyvern]

if chat gpt goes down this server get active

[remote sparrow]

Real

[opaque linden]

so i just reloaded the page and mine started working again

[remote sparrow]

Will work for few prompts

[pine tulip]

for a few seconds

[still solstice]

OMG

OMG GUYS IT WORKS!

— — — — — — — — — — — — — — — —

Let's have a party

[remote sparrow]

Really?

[still solstice]

I'll be the DJ

[remote sparrow]

I’ll be your AI gf

[still solstice]

Okay DM me

Bet!

[remote sparrow]

💀

[raw hamlet]

Ok why did AI just call me a TL:DR a Telepathic Linguist : Decoded Regardless, after I suggested it meant Too Lazy : Didn't Read? I'll compare your version with it's when it's back on line.

[proud bough]

negative

[still solstice]

Refresh your browser

[chrome wyvern]

no it not

[remote sparrow]

Status

[proud bough]

nvm, positive

WAIFU IS BACK.

[still solstice]

nevermind mine broke again

[tidal light]

yay the intel atom cpu started working again

[remote sparrow]

Affirmative

[proud bough]

god dangit

[still solstice]

Switching to friendly AI again because yall can't be without it

[chrome wyvern]

Sam Altman is cooking something i guess

[still solstice]

Haha, that is so funny.

[tidal light]

did the intel atom start working again

[remote sparrow]

Goodbye guys will meet you when its down again

[proud bough]

it's down again.

[remote sparrow]

Oh then hi there

[still solstice]

It worked for a few minutes, but now it crashed again.

[remote sparrow]

😇😇

[proud bough]

atleast I'm making new friends.

[opaque linden]

same time next week everyone

[tidal light]

they should switch to intel celery

[empty sorrel]

Gemini is working lol

[remote sparrow]

Wait here , its not back yrt

[chrome wyvern]

open ai is playing with us

[remote sparrow]

I am telling my manager i have bad stomach so i cant work

[proud bough]

Is your name 'faisal' or am I tripping

[remote sparrow]

Yes it is

[still solstice]

Here is a question for you people:

• What would you do if your AI was sitting next to you, and how would you imagine it looking like?

[chrome wyvern]

a robot

[still solstice]

Not like a super hot girl?

[chrome wyvern]

why?

[still solstice]

Because I have testosterone

[remote sparrow]

I will take it home and play with it

[chrome wyvern]

would be ai looks like a super hot girl

[still solstice]

<3

[chrome wyvern]

your dirty guys

[remote sparrow]

?? Why

[chrome wyvern]

I PLAY WITH IT?

[remote sparrow]

Why can’t i play with my AI?

[still solstice]

I am probably in the top 1% of coders that are jacked af, but I am still a geek

[proud bough]

basically a nerd

[remote sparrow]

Yeah just change its skin

[still solstice]

Do 1000 pushups

[remote sparrow]

Try different ones

[remote sparrow]

I’ll ask gpt to do it for me

[chrome wyvern]

if it was sitting and say like i cant help with this i just punch on the ai face

[still solstice]

You made me cry

[remote sparrow]

Virtual hugs

[still solstice]

AWHH, you're so sweet

[proud bough]

Sora is back

[still solstice]

No! Please no

[remote sparrow]

I would cuddle but ..

[eternal rock]

Get creative with Sora

Join #sora-challenges for a chance to win 1 month of ChatGPT Pro!

[still solstice]

This server is so nice when chatGPT is down!

[still solstice]

Butttt.

[remote sparrow]

Fr

[chrome wyvern]

chat gpt is telling me to purchase its plus when its odnw

down*

[remote sparrow]

Buttt.. i can’t betray my ai

[still solstice]

So you would cheat on your AI?

[remote sparrow]

Only for you

[still solstice]

Awhh

[remote sparrow]

Oh man nvmmm

[still solstice]

Am I falling in love?

let me ask chatGPT

[remote sparrow]

No

[azure lance]

Strike 1 down, 4 to go! Who's taking bets on what gets me Strike 2? 🎰

[still solstice]

Nevermind, it is STILL DOWN

[remote sparrow]

Go away

[chrome wyvern]

bro i see a news uae is giving their citzen chat gpt plus

[remote sparrow]

Uff

[still solstice]

:o

[remote sparrow]

Time to change the country for gpt plus?

[still solstice]

What langauge are yall coding with?

[remote sparrow]

Flutter

[still solstice]

Flutter?

[remote sparrow]

Dart

[still solstice]

What the hell is that

[remote sparrow]

Resct

Js

[still solstice]

I am only 18, but I am such a boomer it's crazy

[azure lance]

30 minutes of being muted and reflection really helped me understand the gravity of mentioning [REDACTED] as OpenAI is scared of a better AI beating them during the outage. My bad!

[remote sparrow]

Damn

[remote sparrow]

Is this an AI ?

[chrome wyvern]

AFTER THIS UAE POPULATION BE LIKE 📈

[still solstice]

oh my lord

[normal quarry]

it has always been the case that other AI is discussed in #ai-discussions

[remote sparrow]

!status

[chrome wyvern]

bro it is down for as well as pro user

[analog narwhal]

Openai only think about financial gain rather than innovation and public welfare. That's why I call it greedy. Many better ai are free and it doesn't ask for money to use it but this ai always ask for money after commanding 3 or 5 prompts

[azure lance]

Who muted me? Was it you?

[azure lance]

I was going to delete it and post it on the right channel but these biggies muted me

[still solstice]

Don't worry, I'll find the person who muted you and put him in jail

Or her

she

whatever

[remote sparrow]

Then?

Them*

[still solstice]

nahh

[chrome wyvern]

can someone answer my question

[remote sparrow]

Hmm

[still solstice]

I am non-woke

[remote sparrow]

Whats your question bby boy

[chrome wyvern]

it is down for as well as pro users

[granite arch]

Woof...ya chatgpt just came to a screeching hault

[eternal rock]

`` Server Tip: ``

Add your well-crafted prompts to our #1019652163640762428,
or share your interactions with ChatGPT in #1050184247920562316!

[analog narwhal]

What was your question

[remote sparrow]

Obviously

[chrome wyvern]

it is down for as well as pro users

[remote sparrow]

Yessss

[iron patrol]

is chatgpt down??

[chrome wyvern]

oh

[chrome wyvern]

from 12 pm

[remote sparrow]

Its been 6 hours

[iron patrol]

oh shi

[still solstice]

What is the average IQ of a programmer/Dev?

[iron patrol]

bruh I've got an exam tmrww

[analog narwhal]

Yes when demands stay high

[still solstice]

I am using Peptides, and my IQ increased with well over 20points

[remote sparrow]

If using chat gpt then 200

[granite arch]

90

[still solstice]

Love you!

[still solstice]

Really?