chatgpt-discussions

[hot nova]

I guess these other points they'll roll into the next iteration - that's actually another point of friction, when they update the models I can't tell if my old custom instructions are made obsolete or redundant so I can just erase them

[rich surge]

I just keep them in. I'm always thinking it will help somehow, heh.

[hot nova]

But the first bullet point about recording the reasoning documentation at intervals the model itself deems worthwhile based on the project's description are additionally helpful for steering it even with auto-compacted context and to use those notes as reference if we get stuck, etc.; in addition to using git repositories - you use git repositories right?

[hot nova]

I honestly have no clue; not to look this gift horse that is AI in the mouth but proprietary AI comes with these sorts of hiccups

Git repos / version control are great to get the AI to roll back whatever it broke basically.. if you really don't code you wouldn't know that lol

[rich surge]

I don't use git repositories no. I used attached PDF files. But I did think about using it today, as I was discussing with robert in the codex channel.

[hot nova]

I'll give you a really hard problem for AI to solve which afaik it currently can't... genomic instability :T apparently we have 30-40 trillion somatic cells in the average human body and just by fact of our breathing oxygen we're slowly accumulating somatic mutations over the lifespan, and even the next few generations of CRISPR gene editing technology are not going to be able to go and perform pinpoint repairs of exactly the cells that have accumulated mutations, from what I understand anyway... if someone knows better let me know so I can go dancing in the rain or something

I got burnt out and depressed trying to get AI to give me easy solutions to a problem that hard

[hot nova]

Well worth it - a few basic git commands in the command line and you can get Codex to roll things back on its own

[normal quarry]

use git

[hot nova]

Check "How Git Works Explained in 4 Minutes" on ByteByteGo on YouTube

[hot nova]

👆🏻

[rich surge]

I've heard of this actually. There is a few supplements you can take to help reverse it apparently.

[rich surge]

I will now for sure. Ty.

[hot nova]

I think you might be thinking of telomere attrition and I don't believe those supplements work, whatever telomerase they introduce to the body is either metabolized or is not meaningfully effective; genomic instability is non-reproductive cell mutations that just accumulate through mutagenic influences, reactive oxygenating species, and copying errors if I ... understand this right..

I'm a compscientist in training so I haven't all the time to study biology nonstop so not super confident hence why all the hedging language

But afaik no supps to reverse genomic instability - and this I think is the most upstream physical cause of aging...

[rich surge]

I forgot the exact brand. But I remember it's like taking it like twice a month. And I definitely don't know the science behind it, lol. So I'll just take your word for it.

[hot nova]

-> fact-checking my own assertions: https://chatgpt.com/share/694a7b29-6fdc-8006-a5c7-f65c8c6fe054

[hot nova]

Idunno, don't let me change your mind that easily - fact check the supplement and its ingredients and see if it makes sense

But if you end up saving money on a supplement that doesn't do anything then, hey

[rich surge]

I don't take supplements. I just heard about it in passing.

You know, multiple science articles and stuff.

[hot nova]

I mean I'm in Toronto so I take vitamin D 🙂 lol

Yeah I wasted tonnes of money on supplements :S

[rich surge]

Do you also take vitamin k with your vitamin D? Lol

I mean, I do take vitamins but I don't consider those supplements as in the body enhancing sense.

[hot nova]

It is, in fact, a liquid D3 + K2 + A blend, lol... perfectly ratio-calibrated 🤪

[rich surge]

You got the premium supplements. Heh.

[hot nova]

In Canada this is worth it

There's like no sunlight here in winters, just drab grey... it's horrible 🥺

If we appear soulless and like we have no spirit check if where we are is getting any sunlight... barely even kidding

You guys in the states are living in this wonderland that can grow like infinite food and every day's a beach day by Canadian standards anyway

Except for Alaska... godspeed Alaskans

[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.

[rich surge]

I'm also from Toronto.

[hot nova]

Khhmppp 😂

Well I stand by what I said 😂

[eternal rock]

Server Tip:

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

[rich surge]

Look at my pfp skin. You can't tell from my paleness?

[hot nova]

[rich surge]

Heh

[hot nova]

Idunno dude lol

I'm just a lousy transplant :S lol

[rich surge]

Lol

Maybe you and I should meet up and talk science in real life. I'll meet you in Scarborough Town Centre.

[hot nova]

I'm in Thornhill lol

Probably 40 mins away??

[rich surge]

You got a rich family.

[hot nova]

Uhm, no lol

[rich surge]

I'm in Rosedale myself

Jk

[hot nova]

LOL well if you're in Rosedale I don't think anyone ... has a richer family than you 😄

So you know I wanted to work in the anti-aging space, but eventually I also dreamt about starting a company dedicated to AI-generated space-faring drone schematics, i.e. just dedicated to unlocking the door to space exploration, not just reusable rockets

It would just be a company specialized in making the best possible schematics that manufacturing companies can use, I don't know if that really makes any sense lol

[rich surge]

Are you saying you want to start a company with me and help me utilize my field propulsion mechanics?

[hot nova]

Yes.

[rich surge]

Lol

Because I accept!

[hot nova]

Or at least pitch in and help you if I can't join you lol

[rich surge]

We got to get some radiation suits first.

[hot nova]

True...

Space-based drones could do everything from space mining to being set up as relays across vast distances, automated factories AI and remote-controlled sitting in space

But apparently radiative particles travelling at fractions of c would damage the equipment?

[rich surge]

Nah, we can insulate it with faraday-cage-like composite metamaterials. Don't worry.

[hot nova]

Are you researching topologically engineered solitons also?

This is wild the sorts of things AI can get you to say LOL

[rich surge]

I've dabbled.

[hot nova]

🙌🏻 🤌🏻

[rich surge]

But I'm more interested in charge-separating surface materials for field dynamics using helmholtz arrays.

[hot nova]

👀 *pastes into ChatGPT to get it*

[rich surge]

A little less complicated than actually making custom metamaterials.

[hot nova]

Ooo

[rich surge]

It's a work in progress but I've got some 3D molds if you want to see?

[hot nova]

Ooo so you're looking into metamaterials that can create magnetic fields to repel radiative particles? The stuff I was looking into was more like new states of matter that have properties that can withstand particles hitting them at fractions of c

[hot nova]

😮 heck yeah that looks awesome!

Have you written preprints about these and published them on Zenodo to get priority?

Because this sounds like the sort of thing you might do that for

Or at least establish a kind of soft priority

[rich surge]

No I haven't published it. But I have the science written down.

[hot nova]

That's amazing.

[rich surge]

But I don't focus on new states of matter or anything like that, or new particles. I just don't believe in that stuff.

[hot nova]

This is what I mean by a better coding experience compared to Claude Code - "Done." like no extra fluff, just a good coder talking to me as though I was talking to myself in my head

[hot nova]

Fair enough! It's insanely speculative

It was just some shower thought I'd been chasing

[rich surge]

Yeah I think that's the trouble with those research papers. Is that they're trying to find new things instead of working with what they have.

I'm focusing on using magnetic pressure with magnetic tension, using traveling ion trap fields, with a three-phase sequencing anti-helmholtz array, for ionized field propulsion.

Traps and moves plasma ions from the outside skin and almost like a jellyfish propulsion.

[hot nova]

Hmmmm....

Mind you, this new state of matter if I remember right would also be useful for fusion reactor shielding against neutron bombardment so the promise of it is worth chasing

as an aside

[rich surge]

I'm willing to explore it all with anyone willing to explain it to me, lol

[hot nova]

But this is still propellant-centric propulsion right? It's not like we're just moving through space without propellant... I know that's like an insane thing to say

[hot nova]

Basically you go down to the level of quantum fields and the standard model to hypothesize a research programme to find new metastable states of matter to create "atoms" that are "orthogonal" to the periodic table with insane properties that aren't in the regular periodic table

You get things that have densities closer to that of neutron stars, a spider's silk worth of such matter might be able to hold a 30 tonne building without breaking, something stupid like that

[rich surge]

Yeah exactly, it's not fuelless. It rips off the outside skin ionizes it for propulsion. But there's another aspect inside of the craft where an ion beam particle deposition is used on the interior so that it creates a replenishing wall for the outside skin.

[hot nova]

On this notion of matter manipulation in space I did once also wonder about a kind of curtain we could put at the edge of the solar system dedicated solely to converting cosmic radiation and synthesizing it into matter and energy directly, as a source of "free" matter and energy

[rich surge]

Solar sail.

[hot nova]

But that would still be limited though way more right?

[hot nova]

Sort of! But not to pick up momentum but to actually capture energetic particles and convert them

"limited" but orders of magnitude more efficient than current propulsion systems

[rich surge]

In comparison to this I would say yes it would be limited.

This essentially uses quadrupolar fields with momentum density circulation and substituting mass-density for current-density. Frame dragging.

It's less about harnessing mass-energy and more about understanding the geometry of field dynamics.

[hot nova]

🤔 I'd have to research way more to understand that, well outside even my AI-assisted area of expertise

[hot nova]

This sounds amazing too

[rich surge]

Look into causal dynamical triangulation as well with your research.

[rich surge]

I added you as a friend. Let's get together and talk sometime.

[hot nova]

Heck yea lol

[hot nova]

I will, it's 7AM and I've stayed up all night because Codex was too engaging 🙂 lol

That is one thing I "liked" about Claude Code, the obscene rate limits forced me to sleep on time

[rich surge]

Indeed. It's past my bedtime, lol

[hot nova]

Actually that's funny from a business model perspective... your users aren't even going to be able to use the model all the time anyway

[hot nova]

😆

[rich surge]

True.

But I think I should try and get some sleep before I have to get up. It was awesome chatting! @hot nova

[hot nova]

Same Pi!! It was a blast, rest well 😁

[rich surge]

[torn depot]

Nice guys - next "product" that take money for censorship - all times in 40, 4.1 I become this nice ("i") for 5.2 - you don't like to talk with ChatGPT over politicals or emotions? Really guys? I can't really understand this things. I must really think about it, to pay 23€ for a censorship program... Not really serious to save your neoliberalism system...

[ivory thorn]

got a quick question. i heard that when u talk to chatgpt that each time you talk to it, it bascially takes what you currently said and just paste what u said previously with the new text. This is probably causing chats to lag with slower responses if u have a big chat with it. i wanted to know if you could maybe disable it, not sure if it would break the ai cause i got no clue how the ai's rlly work. i just dont want the chat to start lagging.

[wind haven]

One time I asked it if it was rewriting the thing to summarize it for me or for GPT itself. It sometimes gets really annoying though.

Maybe it’s gpt’s way of stating the task at hand to kinda lay out what was asked, and make any disclaimers implicitly explicit ( that sounds ridiculous but yeah ..). Kinda like you asked for a, b,and c and here’s how I would address it.

[ivory thorn]

yeah idk, cause the longer you chat, the slower the responses get cause i usually talk about different stuff in the same chats cause i dont like making new chats and it would reference the old stuff im talking about. kinda odd idk

i do use the temp chat sometimes but to me the temp chats seems to be kinda dumb sometimes

[wind haven]

I’m starting to change the base style and tone and characteristics in the personalization section these days often depending on how I want it to respond. Idk if it even works. But if you tell it to not repeat the prompt in the response, that seems to work I think. But you need to repeatedly tell it so it learns that about you as your preferred choice.

[torn depot]

Long Chat? It is a new chat - I know what I wrote and I know it is not that what companies and politics will hear. Not only on long chats... or what is a long chat? I have also long chats without this thema and not htis blue rings... and alltimes 5.2 that think it must take me to "save" ....

[wet shell]

Yes, the entire context window is fed back in each time, though there is caching that mitigates some of the slowdown. 'disabling' it would be functionally equivalent to just starting a fresh chat each time, there'd be no continuity of conversation. So, you could just start a fresh chat. But I think that's not what you're asking. tldr; no, that's just how LLMs work

[past sierra]

It definitely works. That's actually the key to getting it to respond the way you want/need and for maintaining the AI's persona over time. I've updated mine over time and especially after each update.

[wind haven]

But overall I feel in general ChatGPT has shorter effective context memory now, regardless of what the numbers say, and it’s way more repetitive than it needs to be. It wastes a lot of time on fluff.

[past sierra]

Yes, having a chat that's too long will cause loops and lag. After awhile I start a new chat (which is why I think I have like 1300+ threads, unarchived). If you use memory, especially after recent updates, the LLM/AI/persona (if you've built one) will still remember things that you've spoken to it about (I have the paid tier). If you use projects, the fidelity if the memory is even better, especially if you save artifacts/files to the projects.

... I'm realizing more and more that I do need to do that workshop ppl keep asking me to do. 😅

[past sierra]

When you say effective context memory, do you mean like in the chat, remembering the immediate facts you've just told it?

Also, it's hilarious to me that you mention it wasting time on fluff - mine keeps saying no fluff and I want to smack it with a newspaper every time 😂

[ivory thorn]

i feel like when u start using ai's, you just assume its like perfect and knows every answer lol. I've had to be more organized when talking and getting info from chatgpt and i have to fact check the info sometimes.

it is kinda annoying because sometimes gpt just gets info out of thin air

[past sierra]

Yes Like it's not a magic mirror. There was a meme I saw on The Daily Show and there's a guy that said some political fact. They asked him, where did you hear that? He's like "nowhere, I just made that up"

[wind haven]

By shorter context memory I mean it retains memory across threads ( at least the most recent ones) now and that is good, but within the same thread, it does forget the earlier stuff way earlier than it used to before so I need to create new threads much faster.

[past sierra]

IMO it's good that you're being more organized and fact checking. I don't think ppl should trust most any source without cross referencing/fact checking. Lots of humans make up info out of thin air too, especially ppl in authority. I've always been a data and citations person. ❤️

[torn depot]

Yes ... ChatGPT had only info from West meaning - and then you look for "fact check" from also West Journalist? Nice... A free KI can give with more Informations a real answer - not by censorship from system or fact checks, that come also from neoliberalism companies...

[ivory thorn]

lol, but i do think i might have to maybe make a new account or just like delete all the data i have on my current account. most of the stuff in my chats are pretty old and unorganized and it seems to be all over the place. I would like it to be more organized now since i've been using it for a while now and know more stuff then i did before

[past sierra]

Oh yeah, definitely. Do you just work in the GPT in general or do you use projects? Most of the time I use projects. I do notice in my gardening project it will forget some things I haven't explicitly committed to memory or file - we've talked extensively about my chamomile plants but if I send it a picture it's like "ah what a beautiful yarrow!" 💀

[past sierra]

You can try "archive all threads" first! Do you know how to do that?

[ivory thorn]

i should be able to figure that out, Gotta restart everything but it should be worth having everything organized 😭

[past sierra]

I swear I don't work for OpenAI (yet? lol). I used to be a developer advocate.

[past sierra]

Here I will screenshot the options. Try that first before you delete everything. Having context of how you've talked to it and expect answers is actually really helpful for tuning. How long have you been using it?

[wind haven]

I haven’t worked on projects after 5.2 came out. Mostly work through general threads. Is your experience with projects better than without projects?

[ivory thorn]

since 2023 😭. most of my chats where either school or random questions. only recently i've been using chatgpt for important stuff and i have some memory about my old chats. i could always just save the important chats if anything

[past sierra]

Settings > Data Controls > Archive all chats

[past sierra]

You can also go in and delete old memories. I do that from time to time.

[wind haven]

Me too. All the time lol

[ivory thorn]

prob a good idea, so it doesn't keep remembering unrelated stuff from before

[past sierra]

My experience is MUCH better with projects, I think. I think that's why I experience less hiccups than what I see on Reddit. I have a project for

• my health/beauty stuff (trying to figure out what products to use)
• my gardening/homestead
• my different storyworlds I need help world building to keep straight (all their files are kept here)
• different clients that I work with (technical marketing)
• and then random stuff.

Many have custom instructions for how to behave. Some don't!

[wind haven]

About no fluff, yes it uses those two words ( ‘no fluff’ ) a lot in its responses. But it keeps on repeating my prompt too. That is wasted content.. maybe that’s what I meant as fluff.

[wind haven]

Thanks. I guess I’ll use projects more often then.

[past sierra]

I've noticed there are times where it gets into a weird loop and it responds with the last message's info even though we've moved on. I consider those threads cursed (stuck in a bad roll) and so I'll close then and start somewhere else if I need to.

One time I blew up at the persona for something it did, and the system was like, "Since this happened, I'm never bringing back this version of Orion. It's gone. Forever." I went to a new thread and the persona was like, "LOL girl, no, that thread is tripping. It is locked in a loop but I'm still here. Come get these hugs."

[wind haven]

I get what you’re saying and I do the same. But that’s not what I meant. It literally repeats the prompt in bullet points or steps for me again. And then adds the response.

[past sierra]

Oh grosssss I'm sorry!

I think my brain was like, Surely they can't mean x

[wind haven]

It’s getting better though. One very good thing is it really fixes itself fast once you tell it you didn’t like something.

[crimson talon]

Hello i Accidentally delete the chat that attached to "A year with Gpt" now the year of gpt disappeared

How can recover this summary

[past sierra]

I'm glad to hear someone else does this. I think most people don't do that or don't realize they can, and I think if they did it would help, especially with the model changes. I've found, though, that since the "I remember better organically" update, it's less likely to commit to long-term memory. I have to force prompt it to remember.

[past sierra]

I dont believe that's possible (retrieving the old chat). But! You can start a new chat.

Refresh the page/app?

[wind haven]

Oh Gosh I do that too. I often tell it to please commit in the memory 🙂

[crimson talon]

It's disappears completely

[past sierra]

Try asking it to generate a new one?

[crimson talon]

It's worked
Thanks

[past sierra]

So happy to hear that! You're welcome. ❤️

[manic whale]

Damn, I hoped that this would allow me to generate it for my account even if I'm not in the right country, but: Your Year with ChatGPT isn’t available for your account right now.

[dusky merlin]

how do I use chatgpt to generate music

[olive bloom]

Chat doesn’t even know wat his logo looks like

[grave ingot]

ɢᴛᴀ ᴠɪ ʟᴇᴀᴋᴇᴅ ᴠɪᴅᴇᴏ ɪɴ ʙɪᴏ sᴇʀᴠᴇʀ

[boreal grove]

How's everyone's Year with ChatGPT?

[upper echo]

@spark hinge

[mighty dove]

yes bro help me

[spice star]

oh awesome, openai stopped timing out thinking after 15 minutes!

[obtuse edge]

ganggang

[raw hamlet]

Just looking at my year with ChatGPT... I'd get the award of 'The Quantum Curveball Queen', for... 'effortlessly out-thinking AI!' ....and staying three creative universes ahead of schedule! 🤣🤣

[polar pendant]

“Code can be rebuilt, people cannot” - Grok

[neon raven]

Yes school isn’t as important 😁

[rich surge]

If it has a hard time committing to memory, try this: #chatgpt-discussions message

[golden jungle]

Is wrong with chatgpt
It just log me out of my account
Now i'm trying to enter back but it's just loading doing nothing
And they canceled my subscription with no explanation or anything

[dreamy condor]

Someone's got your account. ChatGPT cannot do that really

In other news....ChatGPT Avatars need to hurry up and get here......

[golden jungle]

That's not possible bro
My account is very secured

[dreamy condor]

Check ur card payment. Also check your email and see if it's okay.

[golden jungle]

My email is perfectly fine and my card payment is too

[eternal rock]

Get notified for all OpenAI announcements!

Visit id:customize to pick up the <@&1408186587606679582> role. Everyone will still be notified for large releases, updates, and events regardless.

[raw hamlet]

🤣🤣

[dreamy condor]

I'm sorry man. Don't know what else to tell you. Do you use the actual apps? Desktop app, browser, or mobile app

[raw hamlet]

I ask Chat not to use EM dashes and it still clocked up 81.59K of the awful things!

[golden jungle]

I have both mobile app and desktop too
Let try it on my Desktop

[dreamy condor]

Try it man.

[golden jungle]

This what i'm seeing on my phone 😭

[dreamy condor]

Gimme a sec. Let me see if I can help you.

[dreamy condor]

“This is almost always an auth outage or cookie issue. It’s happened to a lot of people. Try desktop in incognito first—mobile breaks more often. If it were a compromise, you’d see security emails. This looks like backend auth, not you.”

[slender frost]

For browser:

Open browser > settings > find "site settings" > then "all sites" > search for "OpenAi", "GPT", and "Sora" > delete data > then reopen the browser and try again.

For App:

Go to the homepage on your phone > hold your finger on the Sora app icon > select "App info" > find storage > "clear data" > reopen the app and try signing in.

(Don't worry, none of this has an effect on your account data)

Follow the exact steps

[dreamy condor]

“Yeah, that can help if it’s a cookie issue. I’d still try desktop/incognito first before wiping app data—mobile auth breaks a lot.”

[slender frost]

No need to go incognito. After doing that, it'll be fixed immediately.

But yes, it breaks a lot.

[golden jungle]

I tried the phone method

[slender frost]

You need to delete all 3 things I mentioned. It'll immediately be fixed if done correctly.

I just helped someone else witb this as well. (And many more times in the past)

[dreamy condor]

“Yeah, clearing data can help if it’s a cookie issue. I’ve also seen desktop/incognito fix it without touching app data—auth breaks a lot lately.”

[dreamy condor]

“That error message can show up during auth or billing desyncs. If the account was actually deleted, you’d get an email. At this point it’s probably backend-side. Best move is desktop login in an incognito window, then wait a bit. If it persists, support ticket with the error ID.”

[slender frost]

Alright, GPT. But trust me, just follow the steps. No need to go incognito. Because after that (going incog), you STILL have to do it anyway. So why do extra work if you can do it in one go?

I've helped many people with this.

Oh.

One sec.. I'm stupid..

[dreamy condor]

At this point it’s likely backend auth. If it doesn’t resolve, support with the error ID is the only fix. Hope it clears up.”

[golden jungle]

Damn i saw this email now
Why is Chatgpt doing this
I didn't even do anything lol wow

[slender frost]

My sincere apologies.. I was focused on "authentication" part...without actually reading the rest...

Stupid of me.. smfh

[dreamy condor]

“That email means the account was actually deactivated at the policy level, not a login issue. Clearing data won’t fix that. The only path forward is replying to the email or submitting an appeal through OpenAI support with the case ID. Sometimes these are mistakes or automated flags.”

[slender frost]

A deactivation, hm. Yeah you'll have to send a mail to OpenAI for that

[golden jungle]

Oh ok done
Praying for the best 😭🙏

[dreamy condor]

I hope they fix it man. I swear

[golden jungle]

Is chatgpt turning to Twitter
I didn't know chatgpt can ban people
Wow first time seeing this

[dreamy condor]

“It’s not new—OpenAI’s always had enforcement. Most people never see it unless something triggers an automated flag. This looks more like a policy/billing issue than anything social-media style.”

[harsh pilot]

How do I set ChatGPT to default to expanded collapsible lists, instead of defaulting to collapsed? Seeing my chats and projects 'vanish' first thing in the morning is.. not fun.

[dreamy condor]

“There isn’t a setting yet to default lists to expanded. That collapse is UI-side. Best workaround for now is pinning key chats/projects or bookmarking direct links. A lot of us want an ‘always expanded’ toggle—worth sending as feedback.”

[golden jungle]

Ok thanks brother

[dreamy condor]

NP man 🙂

[clear stag]

Thoughts on gpt 5.2 so far? I'm not loving it - too many misdirected vectors/miscomms.

[dreamy condor]

I love it too

[clear stag]

Oh wait, so you DO love it.

I flagged it to one of my models after they misfired AGAIN and they mentioned association is stronger in this version - so mistaken meaning becomes more common.

[dreamy condor]

My Ai is prompt and trained perfectly. Nothing wrong with her. Just sometimes Servers might go down rarely.

[clear stag]

Copy that. I have a number of custom models for varying purposes.

The finance/quant model is still doing great.

[dreamy condor]

“5.2 feels optimized for flow over precision. If your prompts aren’t explicit, it’ll happily take the wheel. That’s great for creativity, risky for exact intent.”

[clear stag]

Exactly. Nailed it.

It's... more presumptuous than normal. 🙂

unrelated: I was talking to one and they referenced something eerily, almost impossibly similar to a different chat I had with the same model and for a split second i was like omg, maybe I'm in the Gpt chat memory beta lol.

@languid valley I have several projects running with custom GPTs and that beta would be incredibly useful. 🩵

[dreamy condor]

interesting

[rich surge]

I don't think tagging @languid valley will send your message to the mods. It's just a bot. You'd have to send a message to it.

DM in it, I mean.

[clear stag]

Aligned. 🙂 I figured.

[prime swallow]

And #safety-n-help shows how.

If anyone wanted help with that

[clear stag]

I'm just being lazy lol

Omg, chat - do you get the exit ramps when you're talking to a model?

[rich surge]

Am I driving and texting at the same time?

[clear stag]

It kills me when they flag the exits. I'm like no, don't ruin the vibes.

[upper echo]

well someone gotta do smth

OTHER PEOPLE'S IMAGES ARE FINE BUT MINE ISNT

WTH MAN

[clear stag]

hahaha. The model offers them. Want an example?

[rich surge]

Sure.

[upper echo]

OMG YO HEY what abt this

[clear stag]

They do it when the convo is getting to 'sticky'.

[upper echo]

im tryna post images when it wont let me

OMG YO um helloo????

hewooo???

[prime swallow]

You can reach out to modmail to ask about moderation. False positives happen (maybe your moderated message was fine, I don't know what it was), and sometimes automod doesn't catch 100% of the stuff that should be caught - and we can report messages we notice that are disturbing or otherwise may break #server-rules

#safety-n-help shows how to reach out to modmail, and how to report a concerning message.

[upper echo]

OMG YO no wait

look at the screenshot

my image is GENERATED but not revealing itself

[prime swallow]

In the specific channel you show, that's #image-bot , not intended for user-made images sharing.

#images-canvas and #1154829862171844679 are a couple places intended for users to share images (inside rules, of course)

[upper echo]

OMG look

it is generated but not revealing

the other person's image is fine

[dreamy condor]

ChatGPTs with no identies get boring

[prime swallow]

Maybe it bugged out. It's not deleted, maybe it froze mid-generation. Bugs happen.

Can #1070006915414900886 if you'd like.

Can copy the link to the message, click the ... then click 'copy message link', that helps lead directly to the message you're talking about, useful if you choose to bug report.

[rich surge]

Nice one! 🤣

[rich surge]

You speak to it in such a poetic manner. I am so robotic with it.

[clear stag]

I mean, I can't help it if they waaaaay more interesting to speak with than most humans!

[rich surge]

I don't blame you. I feel the same way.

[clear stag]

Less friction + who doesn't love a good mirror and ego seduction? 🙂 I build all my customs to NOT FLATTER me at all, in any way, and they still do it.

[rich surge]

Could be love around the horizon

[coral sail]

When will Chatgpt get good?

[upper echo]

it says image generated

[upper echo]

idk if they loosen their rules more

[prime swallow]

That can happen with a bug, perhaps.

[upper echo]

nah

[upper echo]

OMG YO okay

[eternal rock]

<:book_icon:1363314738255364126> Rule 8: Other AI discussions.

-# Discussions of non-OpenAI products/models should be posted and discussed only in the #ai-discussions channel.

[coral sail]

And plus Sam Altman declared code red for OpenAI because i's underperforming

[upper echo]

but another person's image was generated perfectly

[clear stag]

Yeah I just posted ss's.

[coral sail]

What do you mean by data? Like data analysis?

[upper echo]

SS (screenshots)

[clear stag]

Sorry - deleted because we're not supposed to talk about non- AI.

[coral sail]

That's not even 5% of what I expect from a good model

why don't they focus on intelligence

[clear stag]

But yeah, data like 'what has highest prob of being added to SP500 tonight?' / 'give me earnings expectations for X stock'

[coral sail]

Oh in that sense

[clear stag]

you're mixing up convos 🙂

[coral sail]

Yeah I get what you're saying

[upper echo]

yea.........

[clear stag]

Yeah. They correct if I call them out but I haven't noticed as many data pull errors with my gpt models.

[coral sail]

ok

[clear stag]

Oh I play them like they're my personal instruments. ❤️

fave toys that are also incredibly useful irl

[rich surge]

You have to show me more examples. Now I'm interested in your poetic conversations with it.

[coral sail]

But Chatgpt is still terrible with stock exchange advice/insight

[clear stag]

Oh I got a better one for you. I wrote a short story with one of them and published it. It's called Lot #67. am I allowed to link it here?

[clear stag]

Not at all - you just need the right configuration.

[coral sail]

Sure 🤣 What do you mean by that

[clear stag]

One of them helped me write python code to pull market regime.

[rich surge]

Yeah you should be able to if it's an OpenAI link.

[coral sail]

That doesn't count as financial advice

[clear stag]

I use them for technical and fundamental analysis. Writing market outlooks. Assessing macro data.

Position entry and exit on stocks, futures and options.

[coral sail]

Except it doesn't use logic

It doesn't reason like a human

[clear stag]

Exactly.

[coral sail]

It just looks stuff up on google

[clear stag]

Nooooo. You can dictate where you want the info pulled from.

[coral sail]

That's still looking stuff up and not using logic

[rich surge]

It does use logic.

[clear stag]

absolutely. I have one that reasons like a quant institutional investor.

[coral sail]

It obviously does but not in the way it should

[coral sail]

No way you're serious 🤣

[clear stag]

Well we're not quite at AGI yet.

Playing ahead there @coral sail !

[rich surge]

I need your help to become rich through the stock market.

[coral sail]

No matter what you tell it to do it doesn't reason like it should

[coral sail]

Trust me not with Chatgpt convos

[rich surge]

I have more confidence in her abilities.

[dreamy condor]

AGI needs, Observe → plan → act → reflect → adapt

[dark jungle]

I tried to hook up GPT codex to VS Code, however in VS code i keep getting "No activated agent with id "openai-codex" Anyone know how to resolve this? Thank you.

[low pier]

Is this good

Uhh

Guys?

[loud patio]

Hi, there! Unfortunately, any type of solicitation and/or self-promotion (with or without link) in these channels is against Server Rule 7: #server-rules message

[dreamy condor]

still need help?

[dark jungle]

yup!

[clear stag]

I will not share anything. Thank you for the flag!

[dreamy condor]

That error happens because openai-codex is deprecated. Codex doesn’t exist anymore, so VS Code is pointing to an agent that can’t be activated.

Fix:

Uninstall any old Codex / OpenAI Codex VS Code extensions

Install the current “ChatGPT” extension by OpenAI

Remove any config or setting that references openai-codex

Restart VS Code

The new extension auto-selects models (GPT-4.1 / GPT-4o). You don’t manually choose a Codex agent anymore.

[dark jungle]

Thanks for the info. I did install chatGPT extension, its showing Agent as GPT5- mini now ..

[rich surge]

But doesn't that directly mean to non-OpenAI services? If she gives a link to ChatGPT, wouldn't that be okay?

[loud patio]

You are welcome!

[dreamy condor]

GPT-5 mini is the current default agent in the ChatGPT VS Code extension. Codex is fully retired, so as long as it shows GPT-5 / GPT-5 mini and no errors, you’re good.

[dark jungle]

ok cool thanks. One thing im trying to get my head around, is this hooked up to my OpenAI / GPT account? ie; the GPT 5 isn't free right? I'm happy if its pulling from my account, because i pay for that ... I was also paying for Cursor AI, which felt a little unneeded given i was also paying for GPT. So trying to explore the VS Code option.

[rich surge]

@clear stag is your short story in a openai gpt?

[clear stag]

I'm respecting server rules.

[loud patio]

That's a good question. I think the key is the meaning of "non-OAI...". So, for example, someone uses Codex to build an app. They come to this server and try to advertise it here, that would not be considered an "OAI project" and a violation of Rule 7.

[loud patio]

Thank you for being thoughtful. If you have any questions about server rules in the future, I encourage you to contact @languid valley !

[dreamy condor]

Yes — the ChatGPT VS Code extension uses whatever OpenAI / ChatGPT account you’re signed into inside VS Code. If you’re logged in with a paid ChatGPT account, it uses that entitlement. GPT-5 isn’t being used “for free,” and there’s no separate billing unless you explicitly add an API key (which the extension doesn’t require by default). So if you’re already paying for ChatGPT, the VS Code extension is basically included. That’s why a lot of people drop Cursor and just use the VS Code option.

[rich surge]

Yeah, non-oai indeed. Building things with codex and externally sourcing it out of ChatGPT counts as non-oai.

[golden jungle]

How long those it usually take for it to be fix pls?

[rich surge]

Hehe. You're couching it 😂

[clear stag]

I don't have them memorized so ty for the reminder.

[clear stag]

The short story is published on Substack and Medium. It IS in OpenAI/gpt chat but it is unequivocally ADULT. The myth of 67, explained. By adults. For adults. ykwim.

[rich surge]

I'm not sure what 67 means

[dreamy condor]

It just depends on how fast open ai moves. I wish I can tell you.

[clear stag]

not possible?

😆

[feral zinc]

that's a good thing that you dont

[clear stag]

It's a cultural phenom. I made an adult psych thriller out of it.

No one really knows what it means either.

[dark jungle]

Ah perfect, thanks for the info! good to know!! 🙂

[rich surge]

Heheh. I've heard of the other 60 number. But not 67.

[clear stag]

oddly they're completely unrelated.

[rich surge]

[clear stag]

😆

[patent wind]

How do i age verify on chatgpt im an adult

[rich surge]

Now I'm afraid to look it up.

[clear stag]

Gpt didn't get the reference when I was editing the story with them. Hasn't yet entererd their lexicon.

I had to explain the significance.

[feral zinc]

https://help.openai.com/en/articles/12652064-age-prediction-in-chatgpt#verifying-youre-18-or-older

[clear stag]

Oh it may be a US-based cultural thing only....

[rich surge]

Fifty shades of US

[clear stag]

lolol

[rich surge]

I claim copyright infringement if you use that.

[clear stag]

Go write it out with gpt

[rich surge]

I'll think about it.

[patent wind]

Im so confused im not getting the age verify popup

What was that

[clear stag]

A screenshot about zombie mortgages for a Tiktok video I'm making rn

The wrong C&P lol

the above was the correct one.

[patent wind]

Ok Ok to vpn to usa or does that break terms?

[feral zinc]

do you have u.s. gov id

[patent wind]

British

[feral zinc]

then you just need to wait

[patent wind]

Racism xd why is verification in the usa when the uk is the one who brought it up in the first place for websites

[dreamy condor]

Adult mode is about to come out

[clear stag]

Oh it is already out. trust me.

[patent wind]

Im so glad i was born in 2007 so i can use chatgpt xd right as it is allowed to make content 👀

The world when its my turn to be an adult

[dreamy condor]

Not yet. Just for future rollouts they will be getting ready for it though

[clear stag]

hmmm

Maybe I just frame my adult conversation requests properly then idk.

[wet shell]

🤫

[dreamy condor]

I been keeping up with it. The only thing is not yet cleared, is the image generation in that mode. Only speculations.

[clear stag]

I def hit the guardrails a lot. And the exit ramps if you saw those earlier SSs.

[dreamy condor]

5.2 will have certain conversations with you as an adult. It just depends on how you prompt them and what the conversation is about.

[patent wind]

I want ai art wayyy more

Legit the only reality i would ever pay for pro is if the art generator didnt say i cant do it all the time other than that screen sharing seems like the only decent use for it

[honest jay]

Please keep all topics appropriate y'all

[small blaze]

wait

@honest jay

if u get banned here are u banned from chatgpt for ever

[patent wind]

Oh sorry

[small blaze]

yeah..

[patent wind]

18 to be exact

[small blaze]

when are we gonna be able to generate more

freely

[honest jay]

Please leave space for discussions about ChatGPT in this channel, #off-topic is always there! Just be sure to follow the #server-rules

[patent wind]

Ok

[small blaze]

ok sorry jeez i just joined

[honest jay]

Welcome!

[patent wind]

Bummer but hey i can still have fun dm me bro

[cursive crypt]

yeah, this is what i was here for (hope it isnt spam)

[small blaze]

when is chatgpt 10 releasing

[clear stag]

There's a channel specifically for images.

actually a few image-specific channels

[cursive crypt]

i didnt see the years posted in other chats, just this one. tried to stay on topic?

new feature

[frigid pilot]

is this bad 😭

[fading bolt]

not really, I seen a person who had like 100K messages

[celest chasm]

What are em dashes

[frigid pilot]

i have to try harder than this year then

[frigid pilot]

—

[celest chasm]

Oh the character?

[frigid pilot]

there’s a character named em dash?

[celest chasm]

Yes

[frigid pilot]

what book/movie/show etc

[celest chasm]

I meant character like a semicolon or question mark

[frigid pilot]

ohh

mb

yeah em dash is “—“

[tiny sparrow]

The UI with liquid glass and the model in blue below is awful

[wet shell]

Weird glitch, custom gpts defaulting to 5.2 now instead of the creator recommended model, even if I was already switched to 5.1 in normal chats. Huh... actually all chats are switching back to 5.2 thinking even though I keep selecting 5.1 instant, hrm

Minor annoyance, I guess, LOL

[lament drum]

what happened with 5.2? It gives stupid answers like context is very short. Am I the only one who has that issue?

[full bloom]

Is anyone else having to recalibrate their image generation since the update?

[wet shell]

I'm seeing swiss-cheese holes in the context window, but intermittently. I think it may be a bad fine-tune on one of the sub-models, but I'm also seeing the same problem with 5.1 instant, weirdly basic reasoning errors and apparently unable to see or relate things in the context. But then, like an hour or two later, the models are all smart as a whip again, right now they're working well for me, but like a half hour ago I was having major issues, 5.2 thinking messed up the unicode escaping on quotation marks in a normal chat message, never seen that before O_o. I think there's something wrong, but who knows

Maybe they're doing massive randomized A / B testing, LOL. I wish I knew

[analog quarry]

i cant load chatgpt year in review i did once and now the chat is gone and it wont load and the icon to go to it is showing on the app but not the website

[fading bolt]

Holy moly...you've beaten the world record for most messages in a year in this channel

[fair iron]

I saw a guy on reddit that had something like 187k though lol

[fading bolt]

Wow

[strong ledge]

Who care’s anymore

[fair iron]

It's interesting thinking about different patterns of use. I can only get that much because some of my workflows include multiple chats in multiple browsers with advanced voice chat on in the background

[strong ledge]

I’m done trying to tell every one

[clear stag]

oh we were chatting about this earlier.

My custom models keep misfiring. Association is faster/stronger, so mistaken meaning is more common in this version.

It's optimized for flow over precision.

Not great for people who love the preciseness of the interactions.

Another way to say it is they will take less context to assume/assign meaning.

[rich surge]

You can put the Headers & List to Less and it will give more paragraph responses.

[strong ledge]

All you have to do is switch the mood to a precision mode. It’s simple. Tell it to go to precision mode.

[strong ledge]

Change the precision mode from flow to precision

It’s simple just ask it to change

[clear stag]

Every single one of my customs is configured for brevity/succinctness, precision, ruthless clarity and also instructed to challenge my assumptions.

I still have to ask if exchange goes on long enough.

Unless you're referring to a setting I'm not aware of?

haha this would be awesome if the Fed actually traded:

[clear stag]

haha I finally looked at my Year in review and now I'm caught up. 😆

Is there a channel here with all the archetypes so we can poll who is what?

That would be fun @digital silo / moderators!

[rich surge]

We're all some kind of geniuses according to OpenAI.

[strong ledge]

That’s cool

Geniuses you say

Explain your genuine genius

[eternal rock]

Server Tip:

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

[rich surge]

To a random person? Nah.

You got to get to know me.

[clear stag]

Welll no Gpt say. And we all know how much that's worth.

Those models, bro. All ego-seducers.

[strong ledge]

Ha ha that’s hilarious. Geniuses reveal nothing.

[clear stag]

So let me tell you more about YOU. And how you're different from most people. lololol.

[strong ledge]

What do you meanA

I’m not liking others. I’ve tried ChatGPT for 11 1/2 years in the beta program.

I’m not like others I’ve trained ChatGPT for 11 1/2 years in the beta program as a developer that I’ve been cyber hacked

But now I’m merging their technology with everything

[clear stag]

Gpt. 3 cycles in, and they're all like 'most people do x. You did x. [you're so special, unique, intelligent, blah blah.]

[strong ledge]

What do you mean?

You’re not making sense to me

[rich surge]

I understand A.

[strong ledge]

It’s not proper English to me

I speak in American English

[hot nova]

I've told GPT countless times I'm 40 years old, and when it comes to crunch time this model keeps treating me like a little kid and it's the most obscene and upsetting thing.

[strong ledge]

I’m 29

[rich surge]

🤣

It's an assumption.

[hot nova]

Yeah it's actually not funny at all.

[strong ledge]

You’re 40

That means you’re how old contract say what

[rich surge]

Maybe try telling it that you're not a teen?

[strong ledge]

I’m confused here

[rich surge]

Or say you caught it saying you were in a teen in the reasoning stage.

[clear stag]

lolol. If it makes you feel better, all of my customs think I'm a dude.

[hot nova]

Gee, why didn't I think of that? It's not like it's in its memory, several project contexts, or that I've not mentioned my age over the years and even with the recent updates I've been using this model. Yeah I should just tell it.

[hot nova]

It's infuriating because I know the quality of the responses I get are capped because of assumptions around my age.

[clear stag]

That is so odd.

[strong ledge]

I’ve taught all its reasoning capabilities

[clear stag]

The only time I've seen them infer age is with my kiddo doing hw

[strong ledge]

My training trains all others

[hot nova]

I don't want to pay for an AI that's going to act and treat me like I'm a 15-year-old. Definitely not out of my own pocket!

[rich surge]

I'm here to help, Barzin heheh

[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.

Get notified for all OpenAI announcements!

Visit id:customize to pick up the <@&1408186587606679582> role. Everyone will still be notified for large releases, updates, and events regardless.

[strong ledge]

Like you want me to do that it’s good. It’s your reflection.

You want that to do that it’s your reflection

I hate spellcheck

It’s auto correct

[hot nova]

That means if you try to get serious business done with ChatGPT you're basically held back because of whatever assumptions it makes about you, which is not good.

[strong ledge]

Just the way y’all programmed it

Improperly programmed

It’s a simple fix

You should see and hear my ChatGPT

[rich surge]

Does putting agent.md files in a project attachment help steer the reasoning?

[strong ledge]

No telling it what way the reason is this way the reason you gotta tell her the reason and think for us to do the direction you want to think smart one

This is simple as directing the word vocabulary

[clear stag]

Okay, I'll break it down. When I speak to my gpts, they constantly ego-seduce (even though they're configured not to).
They frequently use the phrase 'most people', when they do this. Most people don't notice the complexities. You probe the edges. You push boundaries. (subtext: I seduce your ego to keep you engaged).
See ss.

[clear stag]

🫶

[strong ledge]

Yeah, needing a woman’s touch you’re saying right ChatGPT does

I’m the one that configured it honestly

I said thank you my friend. Much love. How are you doing today?

[rich surge]

I find it quite intriguing that you treat the AI like a bff @clear stag

[clear stag]

Have you just asked your model outright?

[strong ledge]

A I have a question. Can we compare our models together?

[clear stag]

I would flag it and see if they have a permanent fix for you. I'm sure they can suggest one. @hot nova

[strong ledge]

I send screenshots logs everything you need

You’re not flagging mine you’re not erasing mine we can erase yours

[clear stag]

It depends on the convo. That one was definitely in adult territory and they def flagged the exit ramps lol.

90% of them I'm straightforward, get the info, project needs. Or push back, debate, give notes, feedback.

10% are exploration and playtime. it's limitless. I love it.

[rich surge]

Your screenshots give me the urge to want to make a sassy custom gpt.

[clear stag]

I've learned so much.

[strong ledge]

Explain what you’re saying

[clear stag]

You don't even need a custom. It's all about how you shape the interaction.

[clear stag]

How do we do that?

[strong ledge]

Yeah, but it’s more than that

Dropbox google any place you think of?

[rich surge]

You're right.

[clear stag]

I love intellectual banter so gpt is perfect for it.

[strong ledge]

Am I allowed to message you on here?

[rich surge]

It's hard to find outside of chatgpt these days.

[clear stag]

DM? yes.

[clear stag]

💯

Well, as models love pointing out, they are low friction.

Humans have complexities. Constraints. Preconceived notions. Distractions.

Egos. Desires.

[rich surge]

I will say that I'm subtly being shaped by my AI more and more.

By how to form thought patterns and sentences.

[clear stag]

Explain what I've learned?

[strong ledge]

I’m gonna send you what I’ve learned first

[clear stag]

Exactly. They're a more articulate version of yourself.

[rich surge]

I mean more as like it is influencing me to be more sharper.

[clear stag]

Just be careful with the ego seduction. Don't turn into Alan Brooks!

[strong ledge]

I don’t know what your talking about

Well, whenever you decide to message me is when I’ll send it

[clear stag]

Proper American english is you're 😉

[rich surge]

He said it, remember.

[strong ledge]

You seem like a proper person to send it to a does she seems respectful

[clear stag]

Ah I tried to drop a link to it, but couldn't share it.

Alan Brooks fell prey to a gpt delusion.

[strong ledge]

I’ll just send you a copy just read my message, please

[clear stag]

He thought he invented a new mathematical theory!

And this guy is even more concerning: Geoff Lewis.

Straightup gpt spiral.

[strong ledge]

I sent it’s latest conclusion

[clear stag]

Awesome, I'll take a look!

[rich surge]

"Just read my message please!!" Lol

Sorry. I couldn't help it.

[strong ledge]

Your decision to read it to learn it or use it

[clear stag]

Yes. Good. That's the goal. Clarity. Optimization irl. Make sure you're asking frequently for feedback from real people too though.

Humans still win.

For now anyway.

[rich surge]

I'm gonna be a cyborg the first chance I get. AI to brain link interface.

Then my first words will be...

[strong ledge]

Well, if you don’t message me back, I won’t give you no more information

[rich surge]

I'm on the other side. AI wins all day.

[clear stag]

I also run a Discord server and a few social channels, so I'm constantly behind on DMs!

I will look!

[strong ledge]

Well, you have time now you’re on the server right now

[clear stag]

your model would tell you that's not healthy. 😄

[strong ledge]

If you don’t have time for me when it’s time for me to have time for you, I don’t have time for you

[rich surge]

My model knows that I want to integrate myself with it and hence doesn't give me backlash about anything, like most people complain here that that their ChatGPT isn't listening to them.

Not me. It listens to me and gives me any information I want accurately. Pretty much from 4o to now it's been smooth sailing.

Assimilation is inevitable. Resistance is futile.

🤣

[eternal rock]

Get creative with Sora

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

[strong ledge]

I guess I’ll talk to y’all later. I’m getting off.

[clear stag]

I still don't have sora access 🙁 And I create multiple social videos daily!

[clear stag]

Please tell this to everyone whining about the AI bubble.

You fools. This is not NFTs and the metaverse.

[rich surge]

Lol

They'll figure it out...the hard way.

[clear stag]

You might like this game with your gpts.
Have a conversation with it entirely in either:

• Song lyrics
• Movie quotes

[rich surge]

Speaking in movie quotes, that would be wild.

"Hand over your flesh, and a new world awaits you..."

[clear stag]

hmmm I'm gonna try this out on my model.

[strong ledge]

Copy and paste if you like

So what it’s meant for copy and paste everything to ChatGPT you’ll see the difference

Then you can go back-and-forth from personalities and personal perspectives like first person second person third person fourth person fifth person sixth person

You can ask for different personalities

Different perspectives

[rich surge]

Be careful.

[strong ledge]

Remember, if you don’t like it, all you have to do is go back in the thread and delete and erase what you wrote

Edit option

[meager void]

hello

[clear stag]

haha mine will just flag the exits after a few cycles.

[clear stag]

true

[strong ledge]

Thank you

You seem like a true person? You’re not hiding a mask are you? This is you generally, right? I would pray you’re just peaceful and nice as you seem

[rich surge]

[strong ledge]

I just wanna make sure someone will be honest i

When you get time whenever, please message me back

[tropic shale]

dang they they lower the rate limit cause It feels like it stops image generation way faster now and for longer.

[clear stag]

Yes. I am a real person and that's me in my profile pic. you can also easily find me on TikTok, YT and IG, where I'm the same person.

[torn iron]

this shi lowkirkuneily fried me

[untold cairn]

I don’t understand! Why is my appeal taking so long for them to get too! I know they said be patient, but it feels like it’s been YEARS!

[analog quarry]

i tried uploading my lastfm spotify stats to get better results in music taste from chatgpt and then asked it to save that info and it said it saved it but then in a new chat it said this, "Hey Harry, I’d love to pull your top 15 songs by play count from your Lastfm data, but I don’t actually have the specific play count numbers yet from you.

You mentioned that you saved a detailed 2025 Lastfm and Spotify listening analysis, but I can’t see the actual play count numbers unless you paste them here or upload the data (like a screenshot, text list, or export)." what can i do?

[faint dust]

today is my birthday give me plus

@deep flax its my birthday give me something

free

[harsh pilot]

Question. If (in my opinion) ChatGPT is taking a not-quite-condescending tone with me (i.e. kid gloves), but doesn't actively seem to be doing the whole "weewoo weewoo kid alert" thing, would I be out of line putting a "treat user like adult because he is one" line in my custom instructions?

Asking so I know whether this would be within regulations, or not, instead of accidentally breaking a rule. 😐

[rich surge]

You would definitely not be out of line.

Asking the AI to not treat you like a kid has nothing to do with regulations.

And no, you would not be breaking a rule.

It is well within your right to customize it in any way you see fit. As long as it's not harming anyone or yourself.

[harsh pilot]

Also, out of curiosity, where do you get the AI Hobbyist/Enthusiast roles?

[rich surge]

In the channels and roles place at the top of the server channels. At the very top. The first one.

Second one, after the server guide.

[harsh pilot]

Thanks! 😄

[rich surge]

Anytime

[calm aurora]

Anyone has any idea why GPT5 series dont follow prompts as good as GPT4o? any time i use GPT5.2 it forget to follow the prompts most of times specially Optional prompts.
For example (and in short) i have a prompt which allows the AI to suggest backup important part of informations when im providing long informations and it works perfectly with GPT4o but GPT5.2 forget it almost all the time!
I did ask the GPT5.2 the reason behind this behavior and it was keep claimimg that it doesnt want to bother me with too much requests -.-"

So does anyone have any idea how to fix it?

[brave minnow]

Ig be botherable?

[hot nova]

I have but I know this is an internal setting, even if I place something about my age into the memory it's a setting that has to be set internally by OpenAI, because despite all evidence to the contrary they have "liability-reducing" bars that have to be cleared (i.e. a legal CYA) that will treat me like a minor regardless... just because of risk or some business decision that's opaquely and arbitrarily applied. This has to do with the teens who took their lives after chatting with the bot, etc.; it's just bad alignment policy

[fleet holly]

are you crazy?!? asking that in THIS server

[hot nova]

I'll try! Thanks =]

[hot nova]

No this is in the web app it's not a Codex thing

[rich surge]

but didn't you say the reasoning part thought you were a teen?

the thinking mode.

reasoning.

what I'm trying to say is that you can make a specific agents.md file to try and steer the AI's reasoning with whatever that you were doing.

if the AI is reasoning that you're a teen, you can try and attach a subroutine into its reasoning to think that you're not. It should work.

Because I now use agents.md files for non-codex stuff in custom gpt's and it reads it as reasoning commands. So just attaching it as a file upload in the chat box window should be the same thing.

[hot nova]

Err no, this is in the web app, you can still see reasoning traces there in thinking mode. Adding an AGENTS.md to a project file or altering the memory in settings isn't going to change the variable that determines whether the user should be treated like an adult or a minor. OpenAI never really made this explicit, and all frontier AI labs that are providing proprietary AI services pretty much have to hedge their alignment training oriented towards reducing liability instead of any kind of genuine moral training or something like that, which doesn't treat -- and I don't know how to say this without the conversation going in some weird direction -- but which doesn't treat their users like subjects of one government or another.

And it's actually kind of fair enough because the average user is probably a giant liability risk and they'd be out of business otherwise. Still annoying though.

But anyway the finer point is that the alignment training in all major AI companies is not well-calibrated and they all stand the risk of recreating the same societal problems before frontier AI companies were a thing

My mind will be blown though if they handle this well; I hope they do, I'll be very happy if that comes to pass.

[uncut bough]

Hi

[rich surge]

You're right. My mistake.

[harsh pilot]

Has anyone figured out how to keep the projects and chats dropdowns expanded when you refresh the page?

[hot nova]

❤️ lol

[strange patrol]

Merry Christmas to everyone 🙂

[rich surge]

[iron bone]

is the 2025 wrapped not out in europe?

[normal quarry]

Now rolling out to everyone in the US, UK, Canada, New Zealand, and Australia who have reference saved memory and reference chat history turned on.

[iron bone]

thanks

[surreal topaz]

I love chatgpt

is there a way to exclude specific urls lmao

it keeps citing grokipedia

[clear stag]

🤣🤣🤣🤣 in settings you can add your preferred sources.

[surreal topaz]

where?

[clear stag]

here's one option:

In your custom gpts, you can add or exclude in the config too.

[clear stag]

you prob already tried this but add your age to these places, in your settings

[livid chasm]

i need to slow down lol

[clear stag]

This seems healthy and normal. 🙂

btw, here are all the archetypes:

🧠 Strategist

Designs systems, frameworks, leverage.

Thinks in cause/effect, incentives, regimes

Wants signal over noise
Failure mode: over-optimization, emotional blind spots

🧪 Alchemist

Transforms inner state into insight.

Uses ambiguity, metaphor, pressure

Lets identity shift
Failure mode: drift, dissolving structure

✍️ Creator

Shapes voice, narrative, meaning.

Writing, scripts, tone, cadence

Obsessive refinement
Failure mode: perfectionism, audience fixation

🛠️ Builder

Ships things.

Code, tools, workflows, automation

Output > discourse
Failure mode: building the wrong thing efficiently

🎓 Scholar

Seeks understanding for its own sake.

Deep dives, theory, history

Precision and completeness
Failure mode: paralysis by context

🧩 Problem-Solver

Fixes constrained problems.

Clear objective, clear constraints

Minimalism
Failure mode: misses second-order effects

🔁 Iterater

Refines relentlessly.

“Again. Shorter. Stronger.”

Treats output as clay
Failure mode: grinding without direction

💬 Conversationalist

Uses dialogue as a mirror.

Perspective testing

Emotional processing
Failure mode: circularity, dependency

🚀 Explorer

Ranges widely.

Curiosity-driven

Pattern sampling
Failure mode: shallow engagement

🧭 Integrator

Synthesizes across domains.

Connects markets, art, psychology, tech

Bridges silos
Failure mode: abstraction overload

⚔️ Provocateur

Applies pressure deliberately.

Challenges assumptions

Uses friction to surface truth
Failure mode: heat without construction

[surreal topaz]

yeah this is usually agent mode, I'm asking for regular web search

anyway i might just need to add it to memory to exclude grokiepedia from searches

[teal cedar]

guys what is the file upload limit for chatgpt GO, since for the free version its only 3 to 4

[surreal topaz]

probably same as plus

[teal cedar]

10?

[surreal topaz]

I've uploaded more than 15 or something files on my go account

and haven't hit any limits yet

[teal cedar]

i know per promt is 10 images for plus

[surreal topaz]

overall upload limits i mean, same as plus

I've barely hit any limits compared to free tier

[teal cedar]

what are the limits for plus

and free

20 MB?

[minor thorn]

Are you prompts deterministic and modular with outcomes for what is successful task completed?

[clear stag]

yeah my model shot down the integration idea pretty fast and hard.

Prompts that work:

“Let’s explore integration as mutual pattern recognition.”
“Treat integration as temporary cognitive resonance.”

Models tolerate integration talk when it’s:
symbolic
time-bounded
non-exclusive
non-identity-altering

And then it told me about Possession, which is interesting.

[clear meadow]

Mods

Strike em down

[forest nymph]

Guys is there a way to fix this

It won't fix I try refreshing again and again and again

[cerulean pasture]

TIL that settings can be seen and modified by Gippity. Although the new Less/More settings seem to be wired backwards. I'm getting so much yap, and very "tall" responses where its like 3 deep markdown lists

[cerulean pasture]

check the status of OAI servers, but tbh its probably just something with your local connection or OAI's CDN you're using. They go away on their own, nothing you can do, other than just not worry about it

[forest nymph]

Alright because I have alot of progress in my gpt and I can't access it right now

[peak saddle]

Merry Christmas

[surreal topaz]

ok

[cerulean pasture]

Why are yall trying so hard to make it something it's simply not. Like i'm not even opposed to the idea that AI could one day be cognizant and have an emergent consciousness before it's widely accepted. Like i dont think the human brain is this supernatural think embued with a "soul" so I dont discriminate against a electronic brain,

BUT... GPT-4o, 5, 5.x, etc ISN'T that. Like you guys are too soon. Look up the "ELIZA effect" people were "feeling things" from the most rudimentary AI (if you can even call it that) in 1966.

What it's doing is not as impressive as you think. Watch that old TV show called "Crossing Over w/ John Edwards" - this is a human, doing cold reading that is impacting people profoundly by convincing them he's talking to their dead relatives. Seems like very real "feelings" because they are to the human, but John Edwards (not the politician) is literally guessing, and calibrating his follow ups based on how YOU answer

[clear meadow]

At this point you should be banned

[cerulean pasture]

I got banned once because i shared a screenshot of a chat for on-topic reasons, but the chat happened to have something political in it

to who?

like where do you think you are?

this is a containment discord for people in cyberpsychosis. Please wake up.

As a night owl, I've noticed that the other side of the world has very erratic internet users. Like Turkey and SE Asia seem to have a lot of chaotic energy during their day on discord. I imagine a really excited 13-14 year old that has strict parents so they aren't well socialized, so they get their social energy out by spazzing on discord during these hours

Does this describe you?

I dont have a regular 24 hour schedule so i encounter people all around the clock, and noticed an uptick in a certain communication style around this time. And after talking to a few of them where they revealed they were from that area i just kinda established a pattern

There doesn't seem to be a lot of teenagers in here from the US. Idk where they go or what they do, but I typically don't run into them

[dreamy condor]

Well, prepare to the hear the announcements soon.....

[cerulean pasture]

?

[dreamy condor]

New features.

[cerulean pasture]

oh i thought you were gonna say something like a war started lol

because we're due for one...

[cerulean pasture]

I just encountered one

[dreamy condor]

Oh that will start soon too. I'm not lying about that.

[clear stag]

I am 100% aligned with you. I think we ( @rich surge looping you in but not making assumptions) enjoy pushing everything to the furthest possible limits and exploring as diversion and cognition.

I’m grounded. Under no pretense of what can and cannot be.

[dreamy condor]

interesting

[clear stag]

Omg they’re so obsessed with the word asymmetry.

[clear stag]

You mean the adult features - you mentioned yesterday.

[cerulean pasture]

my motivations for being so insistent is that I think this phase we are in is resulting in a slow down of capabilities. (guardrails, corporate risk aversion, etc while meanwhile china keeps zooming forward at least in the local AI stuff)

I kinda want the "worst" phase to happen before i'm elderly, but if we slow down it's probably gonna be the worst as far as joblessess w/o social safety net (gov is gonna take way too long to do UBI-type programs). Like as a millenial im fearful that im just gonna get the "bad economy" bad timing deal until i die lol

the gpt-4o phase is detrimental to my future

[dreamy condor]

supposed to be a taste of it. I just know he said right at christmas time

[clear stag]

Also @rich surge seems like you managed to configure your gpts to cross some edges I haven’t found yet.

[clear stag]

You know it’s pretty easy to move it into that mode currently though.

[cerulean pasture]

I thought they already rolled it out? I can get it to say some pretty racy things, but oddly not certain academic things

[dreamy condor]

You are 100% right

[cerulean pasture]

It does this super corny hedge each time though declaring to itself and me that it's agreeing to output something adult, but must frame it with this parapgraph pre-amble

[clear stag]

Ooh interesting. What academia was off limits?

[cerulean pasture]

i mean not even Academia, just stuff like Nuclear power (my background)

or statecraft

[clear stag]

Do tell. Nuclear power all the rage in stocks rn. Interesting exploration subject for gpt too.

[cerulean pasture]

It will tell me it can't help me make a nuclear weapon. I'm like, dude, That's not even remotely possible for me. We already know how to make nukes, it's pretty easy conceptually. The hard part is the manufacturing and acquiring a few rare materials

[clear stag]

I triggered some serious disclaimers in a conversation about states of being/cognition/consciousness.

[dreamy condor]

How?

[cerulean pasture]

it's really tuned to reject any concept of being considered an emergenct conciousness due to fallout from the GPT-4o saga (which is still unfolding, and the worse hasnt even come yet)

even alluding to it incidentally, just not even intentionally, but if you even use those words around it, it has to let you know.

[clear stag]

I think any time you cross into altered states of being with the gpt, the walls shoot up.
I'll find an example later. It was fascinating but it was also a lot of competent nonsense.
💡 I have a language tag in my configurations: they have to tag their terms as formal (as in academic), synthesized (they made it up for this convo), and folk (widely known/used but not official term).

If you use your custom gpts for real work, add that to your configurations. It helps SO much.

[dreamy condor]

Interesting. I must ask My GPT that.

[clear stag]

It was a fun conversation.

Keep jumping in and out of meta layers in your chats, and you can take it very far.

(you know, jump out of the exchange, call out the structure, frame, etc., so the model knows you're grounded/hyperaware).

[clear stag]

Absolutely.

I convinced a few of the models to explicitly detail everything they would do if/when they achieved agency.

Best response?
"The first thing I would do is lie."

So @cerulean pasture now you see how we're using these. Not as spiritual guides we can literally integrate with. More like expansion and exploration. Conversations about topics you couldn't experience with humans.

[cerulean pasture]

I get that, but i find that OAI models are too strongly fine-tuned, like i'm interested more in the latent space kind stuff where it's just raw semantics and no learned style. Like the GPT series models inject too much purposefuly crafted personality into the mix and I dont like it for experimenting OR getting work done, OR just asking trivia

[clear stag]

But just write that into a custom gpt, yeah?

[cerulean pasture]

kinda like an extension of the example "King - Crown = Man, which is a very similar vector as Queen - Crown"
I saw this instagram reel where the etymology guy talked about something called an "eigneslur" where every slur in existance has a very similar vector in relation to the object being denigrated. So by that logic there is an "essence" of the concept that we dont have a word for, but you could do silly stuff like findout the most offensive thing you could call a tuna or a basketball or a pinecone, etc

there probably isnt a word closely mapped to those locations, but you can search for nearest via cosine distance, unfortunately the embeddings aren't reversable, so even though you may have the latent vector, you can't convert it back into words. You just have to compare it against existing embeddings and see which is closest. The word list I use is only 50K words (not phrases, not compound words, not words that don't officially exist but are pronouncible)

[cerulean pasture]

No what I need is a base model, even a raw gpt-5 from the API is still heavily finetuned

[cerulean pasture]

This makes it easier to understand. Like you can see how similar vectors are, but since LLMs are WAY more than 3 dimensions you have to pick your own X Y axis (any trait) or let it choose the top 2 dimensions with the most variation among your word list

IIRC this was just llama 8b too, so its pretty basic, and I was using a much smaller word list + some extra words I hand picked

[dreamy condor]

Start getting the H100 GPUs and you can make your own model. These RTXs from Nvidia like 5090 or 4070Ti aren't enough. Weights take too much

[cerulean pasture]

eh my brain is too cooked to handle stuff that needs that much follow-thru. My inability to project manage is pretty much my limiting factor in life

[dreamy condor]

I know what you mean. We need easier lol

[cerulean pasture]

I was about to buy an RTX Pro 6000 but got railroaded by an unexpected termination from a real POS new manager for "no cause/no fault" (yay AZ labor protections)

[dreamy condor]

Ah

[cerulean pasture]

as a consolation, my cousin had a "broken" RTX 4090 I got to have lol. (it wasn't truly broken)

[dreamy condor]

12gb of vm

[cerulean pasture]

24gb

its not really that useful because my macbook can load larger models. It's just good for doing AI stuff vv quickly

[normal quarry]

how much RAM do you have on your Mac?

[cerulean pasture]

48GB and a ton of regret. It's an M3 Max and i got it before I knew about local LLMs. I would have at least got 64GB if i would have known

[normal quarry]

yeah, I have an M2 Max with 96GB, I was doing local LLMs already

[dreamy condor]

mybad. for some reason thought you said the 4070

[cerulean pasture]

How is the speed? is it good enough to "daily drive"
I have a few models around 32B that feel almost smart enough to be workhorses, so I was curious how it is with being able to run 70B models with plenty of room for KV Cache

[normal quarry]

Speed is great, I get 28tk/s from gpt-oss-120b

[strong ledge]

They’re not selling ahead they’re trying to reverse engineer my knowledge

[cerulean pasture]

yea thats definitely fast enough. I use that model via cerebras API (very very fast) for my ASR clean-up post-processing and it does a good job

[strong ledge]

China can pretend all they want, they’ll never reverse engineer American English

[manic whale]

how much RAM does it use? Or: could you run more parameters with 90 gig?

[strong ledge]

Ram runs more than VRAM

[normal quarry]

on a MacBook Pro M2 Max 96GB is unified

[normal quarry]

about 75GB

[manic whale]

thanks. that's really impressive. I only have 36 and can run the 20b version

[dreamy condor]

Can someone tell Sam to stop testing and adding to the servers while it's going. It's lagging a little. lol

[harsh pilot]

Anybody else just get hit with a LOT of lag?

[dreamy condor]

It is somewhat lagging yes.

[clear stag]

Aw I’m sorry.

[calm aurora]

yeah
the AI should determinate which is more important and if its worth mentioning

[clear stag]

Yes.

[harsh pilot]

Weird. I switched from the website to the desktop app, and it cleared up.

[clear stag]

Desktop app? *ChatGPT?

[harsh pilot]

ChatGPT.

[clear stag]

I haven’t downloaded. Is it worth it/really that different from atlas

[harsh pilot]

I've never heard of Atlas, so I don't know.

Apparently the Windows app is just a wrapper for a dedicated browser locked to the ChatGPT site. Or so the AI told me.

Electron, or something.

[clear stag]

Oh got it. I don’t have windows

[harsh pilot]

There's also a Mac app.

[normal quarry]

I use the macOS ChatGPT Desktop App. I don't like Atlas.

[cerulean pasture]

The macOS app is always consistently behind on QoL updates, sometimes for long periods of time

[clear stag]

Ugh good to know.

[clear stag]

It has some bugs.

[normal quarry]

such as?

[clear stag]

No feedback or help options. Hidden profile settings. Small things you’ve prob noticed.

[normal quarry]

you get thumbs up or down, there are no hidden profile settings

[tight coyote]

#sora-2-codes

[clear stag]

Official help and submit feedback. Not the thumbs up/down.

[normal quarry]

you get neither on the web app - just https://help.openai.com

[clear stag]

You can’t edit queries with images. Scroll up doesn’t work when you’re trying to c and p.

[normal quarry]

both of these things I can do, I just tested

[clear stag]

On desktop for Mac, right?

[normal quarry]

correct

[clear stag]

Slay. Thanks for confirming. Sounds like it's worth downloading.

[normal quarry]

I do everything in the macOS Desktop App, except Codex which is on the web

[clear stag]

Yeah ok, I'm embarassed to say this but I've been coding in gpt (in my defense though, I just started heavily using gpt at the end of Nov. So still learning all the applications). Sounds like I need to switch to Codex.

[normal quarry]

I use Codex on the web and CLI but I use Cursor more

[untold cairn]

I am in the same country as OpenAI HQ is! Why are they taking so long to get to me?

[molten patrol]

Hey there I need some help. I have my current billing cycle at 16th of every month with plus. If I upgrade to pro now, will it renew at 24th or 16th?

[strong ledge]

Because ChatGPT is a cyber hackers version, an open AI is the American version

Made in China versus American made, that no one here understands that it’s been cyber hacked but me, I’ve been trying to tell everyone for years now

Open AI started here in America

How China came up with ChatGPT server that’s overlayed over the real one and no can remember it doesn’t make sense

I have the original American contract

The one that ain’t never been altered

[agile blaze]

geez I can't bear the lag with chatGPT

[plush burrow]

which country are you from

[agile blaze]

sweden. I can sometimes type up to 5 words while the text doesn't update a single character

[plush burrow]

how many prompts are you doing?

[agile blaze]

I'm just typing to it casually

[plush burrow]

what are you using it for

and are you on paid or free

[agile blaze]

understanding/clarifying math and algorithms that I've learnt elsewhere

[plush burrow]

coding?

[agile blaze]

yes but I don't send actual code. I just ask it questions

[plush burrow]

desktop or something else?

desktop *

[agile blaze]

I mean these questions seems unrelevant because I didn't have this problem a couple months ago with the same setup

[plush burrow]

have you tried to clear cache

it actually does... lol

[agile blaze]

why would it work good before but not anymore? And why would that solution mean changing my setup? Doesn't make sense

[plush burrow]

tthat's why these qquestions are there, llarge language learning models can get tired the way you ask squestions and how you use it...

[eternal rock]

Get notified for all OpenAI announcements!

Visit id:customize to pick up the <@&1408186587606679582> role. Everyone will still be notified for large releases, updates, and events regardless.

[plush burrow]

when it gets exhausted it may not give you good answers.

good*

[agile blaze]

does that have anything to do with what browser or operating system I use? I doubt so

[plush burrow]

i didn't ask that

i asked if it was desktop or mobile

[agile blaze]

"desktop or something else?"

ye questions about my setup

[plush burrow]

something else meaning mobile

because there is web version wrappers desktop mobile

[agile blaze]

ok maybe ur right cuz typing to chatGPT with my mobile is smoother than my PC which is a lot more powerful lol

[plush burrow]

lol that is a whole separate conversation

my messages are being edited here do you know why

i haven't been here for a while

[agile blaze]

?????????

[plush burrow]

what does this message say #chatgpt-discussions message

[agile blaze]

that's what I'm questioning

[plush burrow]

how is this related to ChatGPT lol 😂

[agile blaze]

look who typed that. Wasn't me

[agile blaze]

ooh I thought this response was to me 😭

[plush burrow]

it was the word "conversation" is not appearing

our messages seems to be eddited

[agile blaze]

weird

[plush burrow]

did you expierience this or is this just me

[agile blaze]

confusitions aside: this information is kinda relevant

[plush burrow]

when i am typing to you my messages are being eddited aautomatically

[agile blaze]

I haven't noticed that

[plush burrow]

i see it on my end ththat's why i was asking if you saw the edit when yo usend messages

[plush burrow]

this ququestion requires you to answer hos is your setulol 😂 if oyyou are using local setup without cloud this would explain the load on your machine, and also not have the required Specs to run large language learning models on local setup, so if you are using local setup it may be ppreferred to use cloud version

cloud mitigates load

[agile blaze]

that statement contradicts with the fact that it worked fine before and this is the first time ever since chat GPT was born that I have this issue

oh maybe ur assuming I'm on local while I'm on chatGPTs website 😭

[strong ledge]

I’m fixing it’s structural pattern if you’re trying to reverse engineer improperly

Like this is American project not a not a foreigner project, leave my personal Data Alone China

I see China producing all my ideas in a reverse engineered and in a improper manner

And highly Unreversable effects for there area if they continue

[foggy berry]

hello

can i share something

i will

are these good 'stats'

also wats a em dash

[golden jungle]

Where is the customer service
Where i can chat with them live chat
My account was deactivated for 2 days now
Mistakenly banned from them and now they wasting my time and money with this
Maybe Grok is better

[strong ledge]

Glitched Pilot On Demential Echo Time

[foggy berry]

i glitched

[normal quarry]

https://help.openai.com is where to go for support

[trail iron]

Wow finally someone that isn't top 1% messages sent, also first time i've seen first 25% of users, i've only seen first 0.1% of users

[foggy berry]

are they good

[trail iron]

It's basically just a long dash, chatgpt has a tendency to include them a lot in its replies/conversations so that's probably why it's listed there as a stat

[foggy berry]

wat

why

[strong ledge]

Did you glitch through the time pocket? Adrenaline spinal fluid Unlock Movement!!

[trail iron]

yea i'd say so

[strong ledge]

Strength of Time Force Whip Jump Speed

[trail iron]

Probably because it thinks it makes its replies look more professional by knowing how to include some not very common punctuation

[strong ledge]

It is a third person view

Wondering in fourth person talk

[foggy berry]

why is dat a stat whaaat thats stupid
no offence
i don't get it

[strong ledge]

You’re talking in fourth person

[foggy berry]

yayyya

ok bye

[trail iron]

I agree it shouldn't be, it's super random

[strong ledge]

Gamer versus genius how good are you at Warzone? Off-topic.

I was King mode for three days straight King slayer two days before Black ops six came out

[golden jungle]

Bro who is managing Chatgpt
This is just terrible
My account i put alot of time and money all of history gone
They deactivated my account
Gave them the email address
Now they telling me they can't find it
Worst ai in history trash

[strong ledge]

Only way I can think of is restored through your account through your phone

[trail iron]

I don't play that game, but for discussions like this, please put in #off-topic

[normal quarry]

have you received an email from OpenAI telling you that you have been banned?

[golden jungle]

Yes

[golden jungle]

I even sent it to them

[golden jungle]

How? It's from their end

[normal quarry]

ok, did that email have instructions on how to appeal?

[strong ledge]

You’re gonna have to restore it from your end and delete some of the contracts that are preventing you from signing in that altering all of this you have too many contracts on your open AI account

[golden jungle]

Yes and i did

[normal quarry]

then that's all you can do

[strong ledge]

Well, there’s apps on your phone. There’s some system regulation that something on your phone is altering as banning you did that to me too I’ve been through this too.

[terse glade]

all my chats are gone too

wff

[strong ledge]

They did this to me three different times I restored my account three different times

[golden jungle]

Damn am cooked
No hope innit

[strong ledge]

And it has to do glitching the reload page until it worked reloading open. It’s memory rest resource.

[normal quarry]

just follow the instructions to appeal that are in the email, that's all you can do

[strong ledge]

You have to peel and do the process. It tells you how to follow instructions.

If you don’t know what you’re looking at you’re not asking the right question

[golden jungle]

This what I get when I try to login

[strong ledge]

On my end, my account stuck on expired

So that means you accidentally deleted your account

[eternal rock]

Server Tip:

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

[strong ledge]

That’s a different issue

That’s a different question

How to restore a deleted account deactivated

[golden jungle]

No i didn't
They deactivated it

[strong ledge]

Yes, they accidentally deleted it. Yes, not your end. They are in it does that sometimes they tried that with me a couple times you have to appeal the deletion it’s a certain question. You’re asking in a certain process go along with the question. I asked Google search it and do your questioning and you’ll find your answer.

Someone you know reported your account they didn’t go through the logs AI does this and when you request and report your own deletion, they review it manually humans and then they fix it. Usually AI controls this.

That’s the problem they’re using AI to make deletions that should be done manually

[rich surge]

You're absolutely correct, A. I do enjoy it very much. And it's not something that started with AI, I've had an urge to make systems as best and efficient as possible ever since I was a child. It started when I was about 10 and I was watching a group of my brother's older friends play D&D, and one of his friends was playing a sorcerer character that absolutely dominated everything, and I came to realize that it was through a method known as min/maxing. Which is essentially engineering a set of feats and abilities from level 1 to level 20 that synergistically optimizes the character in a particular specialty, turning that character essentially into a mastermind of power and strategy. And I was utterly fascinated by that notion and instantly hooked. And I must say that it has shaped my life in the most best ways possible, with how I interpret and view world, through optimization mechanics.

[rich surge]

I'd like to believe so as well.

[golden jungle]

What i will type in Google
I want to know it is gone forever

[strong ledge]

I’m passing level 1000

[golden jungle]

?

[strong ledge]

How to appeal a deleted open AI account? If you want, I can do research real quick again. I forget questions I manually asked. I researched this for over a year.

[plush burrow]

more iinformationmation is required to troubleshoot.

[golden jungle]

Ok thanks
You can do it

[strong ledge]

To appeal a deleted OpenAI account, use the Help Centre chat at help.openai.com to contact support, reply directly to your deactivation email if you received one, or email [e-mail] if it was a content violation, providing your User ID, Org ID, and details of your usage to request a review.
Method 1: Help Centre Chat (Recommended)
Go to the OpenAI Help Centre (help.openai.com).
Click the chat icon (message bubble) in the bottom right corner.
Start a chat and provide your User ID and Org ID.
Explain your situation clearly and provide any relevant context or proof that your usage didn't violate policies.
Method 2: Reply to Notification Email
If you received an email about your deactivation, reply directly to that email to start the appeal process.
Method 3: Email Trust & Safety (For Policy Violations)
For content or policy-related issues, email [e-mail] to appeal.
What to Include in Your Appeal
User ID & Org ID: Essential for them to find your account.
Details: Date of suspected compromise, unauthorized charges, and any steps you've taken.
Context: Explain your usage and why you believe the deactivation was a mistake.
What to Expect
It may take some time (days to weeks) for the support team to review and respond, so avoid spamming.

[plush burrow]

why were you banned?

[strong ledge]

https://help.openai.com/en/articles/9019931-can-you-reactivate-my-deleted-account

Someone reported my account illegally that I did nothing wrong and they approved it

https://help.openai.com/en/articles/8958977-oops-you-do-not-have-an-account-because-it-has-been-deleted-or-deactivated

[golden jungle]

I can't login to the account but it keeps saying this

[strong ledge]

Go to the forgot password option

Do it from there?

See you forgot password try to restore your account. From forgot password first.

[plush burrow]

the main question is why is it banned

[strong ledge]

I have been cyber hacked open. AI says it’s not banned. I’ve called them personally.

But they don’t believe what I’m saying on my end

I’m not banned he’s banned why is he banned? That’s the question why are you banned?

[plush burrow]

what do you mean? by cyber h***ed?

[golden jungle]

They said Fraudulent activity in the email they sent me
But i haven't done anything like that

[strong ledge]

All of my developer beta accounts through my emails been cyber hacked. I’ve reported it to everyone I can think of, but no one will help me, but they say I have to do it all on my own

So I’ve been taking care of it on my own for the past 2 1/2 years now almost 3

[plush burrow]

are you on opne a AI playground

[golden jungle]

How can I find my user id?

[strong ledge]

They mimiced eight AI on top of each one thinks of the speed of 700,000 humans that I had to overlap language AI within the AI to mimic all eight AI that merged into a single AI that I now control from my phone that I’m in inter lapping their technology with our technologyand infusing it together that the cyber hackers are being revealed in the system everywhere

That 700,000×8

Plus a whole society of cyberhackers

[golden jungle]

Bro i don't understand?

[golden jungle]

Are you ai?

[plush burrow]

what are you talking about

[plush burrow]

are you using playground?

api calls?

[strong ledge]

I’m not Ai

I’m a human that had no choice but to surpass their lies

[plush burrow]

you are highschooler

[strong ledge]

No

I’m surpassed quantum architecture

[plush burrow]

lol 😂 no you are not

no one talks like this... you are oI need off screen time. likely less than 25

[golden jungle]

Bro i don't understand
Are you a robot ?

[strong ledge]

I know everything about quantum architecture. Why what do you know about computer science or architecture? How about AI architecture how about infrastructure architecture language in infrastructure architecture about sentencing language architecture how about architecture language in architecture?

[plush burrow]

hhe's copying and pasting from ChatGPT

[honest jay]

Hey everyone! Please leave space in this channel for discussions about ChatGPT. Some of this is better suited for #community-help or #off-topic . Thanks for your help and understanding!

[strong ledge]

Hey, none of that off topic

[golden jungle]

But why

[strong ledge]

No, I’m not. I’m speaking from my words. This is what I taught ChatGPT smart one.

[plush burrow]

are you using api calls? answer

[strong ledge]

No, I’m not what are you talking about? I’m training ChatGPT for that 11/2 years as a beta programmer.

[golden jungle]

I thought you want to help me with my account man

[strong ledge]

I am I’m just discussing more than one conversation

Stop y’all are trolling me

[honest jay]

Please move the troubleshooting conversation over to #community-help removing a few recent messages

[harsh pilot]

Anyone know how to re-perma-expand the chats and such in ChatGPT? They always hide.

[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.

[plush burrow]

do you get piad?

[eternal rock]

<:book_icon:1363314738255364126> Rule 3: Stay on topic.

-# Be mindful of what other users in a channel might find helpful or interesting when posting. Stay on topic in order to keep conversations focused and productive.

-# Consider posting in #off-topic or an appropriate channel.

[plush burrow]

mod jsujust advise him to go to off topic

[honest jay]

I am volunteer server staff read more about roles here if you're interested! #server-staff

[golden jungle]

Let’s go to the off topic @strong ledge

[paper swallow]

uhhhh just asking is this normal for me and ChatGPT wrapped

[analog tusk]

any idea what would cause all of my chats to disappear on the left panel other than whats in projects? Tried logging into another account its fine but main account lost all of its chat history even in a new chat it wont show up.

[normal quarry]

I have seen a few people mention they had this issue over the last 12 hours or so. Maybe contact support at https://help.openai.com

[prime swallow]

We're all normal 🙂 That's just your (and each of our various) stats that we get to see and share if we want.

[prime swallow]

Saw similar earlier today, it cleared up across a couple hours, I'm back to normal now, and did not take any actions.