chatgpt-discussions

[somber cloud]

how do you use it for ham radio?

[steady holly]

Last warning. Please stay on topic.

[eternal rock]

<:book_icon:1171408210398289941> `` 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.

[normal quarry]

I had it fix bugs in my Jove Jupiter radio receiver software and do calculations for some work I am doing at the Astronomical Society of Victoria on the radio astronomy 8 metre dish

[dense halo]

awesome

[prisma furnace]

really?

o1 pro did it?

[normal quarry]

I also use it for ham radio, this is a map of radio receivers that received my WSPR transmissions over a week. I had o1 Pro calculate mean distances per watt and frequency of reception by solar conditions

[dense halo]

cool

[normal quarry]

o1 feels like a great dream, I have to pinch myself to make sure I am really awake 🙂

[swift pond]

You’ve convinced me, I’m gonna start using o1 lol instead of 4o all the time

[prisma furnace]

yes even with me he solves complex codes maybe he is stupid in the first response but after the first response he starts doing what u want when he understands what u want exactly

[eternal rock]

Get notified for server events!

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

[normal quarry]

if i am as happy with Pro in 4 weeks than I am now then I will keep subscribed to Pro

feels like great value for money

[prisma furnace]

As I heard in the last article a comparison was made between o1 and people who have a doctorate degree and they said that o1 is equivalent to the intelligence of a person who has a doctorate degree.

[normal quarry]

yep, it is pretty wild when you think about it

[prisma furnace]

Wait two years, and u will see how AI will develop to the point that it will surpass the best humans and u will see new treatments and things entering our world that u have never seen before.

[normal quarry]

I can't wait

[vague pendant]

is there any limit for o1 Pro as a pro user like the o1 preview before? Or is it unlimited requests

[normal quarry]

o1, o1-mini and GPT-4o are unlimited for Pro Users

o1 Pro is limited but I don't know what the limit is

I have used it a lot, no limits affecting me

[maiden reef]

why is gpt not allowed to curse?

[dense halo]

I thought it could

[frozen birch]

Thank you, Robert, for confirming my point. I’m glad we finally agree on the 'limited unlimited' logic!"

gn8 😄

[dense halo]

I think if you put in the instructions, be casual, to curse is okay

it will from time to time

[maiden reef]

really

[cloud pewter]

but if openai detected suspicious usage on pro it’ll limit you from using o1

[dense halo]

I change my insturction a long ago, but I think so

[maiden reef]

i'm just using basic gpt tho

[dense halo]

oh, in that case 🤷

[orchid pollen]

200$ for pro?

what the

[maiden reef]

so you have to trick gpt to curse

[orchid pollen]

20$ -> 200$ was there no room for something inside of that range

lol

[maiden reef]

for what?

[prisma furnace]

[orchid pollen]

what the hell

i would buy a higher one for more o1 but wow not 200$

[maiden reef]

lol there is nothing called pro

[dense halo]

plus got o1 full release

[frozen birch]

lets say that they have.... "too delicate" system... 😉 anyway i must say once again bro - i never wanted fight with anyone haha, but now really, gonna sleep 😄

[prisma furnace]

u and another person can pay or u and 3-4 of ur friends and its use is unlimited

[orchid pollen]

extra for you

[maiden reef]

that is scam

or you are a troll

[orchid pollen]

no

go on the website lol

[prisma furnace]

It's new xDD

[orchid pollen]

bruh

[dense halo]

yeah its like the biggest AI news in the last few days

[orchid pollen]

go on the website dont ask the ai man

[prisma furnace]

[orchid pollen]

jesus...

[maiden reef]

that cant be real

[prisma furnace]

Asking AI instead to click on plans

[maiden reef]

imma check

[cloud pewter]

True

[orchid pollen]

way too much

[maiden reef]

what

2k

[orchid pollen]

they need more offers

[prisma furnace]

They always choose the hard way so they buy GPT for 200$

[orchid pollen]

yeah oyu see 200$ per month

what the hell

[maiden reef]

does that include unlimited gens?

[dense halo]

and voice

[orchid pollen]

from "we dont make money" company to "give me all your money"

jesus

[maiden reef]

lol

fr

[orchid pollen]

man 200$ a month i dont think it gives that value back

[cloud pewter]

if sora ever sees light it’s gonna be pro exclusive

[prisma furnace]

Some reports say they are planning to become a for-profit company

[orchid pollen]

iam using gpt everyday

[orchid pollen]

its not a plan its a thing since elon left xD

the space jesus

[prisma furnace]

Yesterday they contracted with a weapons company to develop new weapon models via AI

[prisma furnace]

Elon hasn't left his case with them yet

[maiden reef]

that gives me a bad taste in my mouth

so are they gonna make gpt free a trial based thing sooner or later?

[prisma furnace]

Nop

[swift pond]

Having a free tier is too useful in terms of exposure and getting people to sign up, I can’t imagine them getting rid of it

[maiden reef]

so if you sub for that 200 is it jailbroken?

[cloud pewter]

nope and it seems to have the same content moderation

[maiden reef]

why would you then

[cloud pewter]

it’s not mentioned in their official docs that pro would be less moderated and I don’t see any pro users saying that it is

[maiden reef]

lol how is that gonna happen if the ai is still in jail

it's so dumb

[cloud pewter]

so yeah you’ll still get “my guidelines won’t let me talk about that” in AVM and “this content might violate our usage policy and content guidelines”

[swift pond]

Are people running into moderation problems a lot? I feel like I talk about all kinds of things on there and like as long as you do it in a way that’s mindful and authentic about learning and exploring ideas, you won’t trip the filters? Maybe I’m less explicit than others 😅

[cloud pewter]

I think openai would never make jailbroken models

[prisma furnace]

What's new tomorrow?

[cloud pewter]

if you let it say a bit more bad words or weird combinations and stuff it’ll be moderated

[prisma furnace]

there are leaks

[cloud pewter]

nothing, wait until Monday because they said weekdays

[prisma furnace]

Why to say 12days then xDD

[maiden reef]

ask it to give you a recipe to make a pipebomb

[cloud pewter]

yeah and anything similar to it, it’ll either reject the request or getting removed upon generation

[eternal rock]

`` Server Tip: ``

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

[maiden reef]

it has the knowledge

[prisma furnace]

Some leaks

[cloud pewter]

If you get too much moderations especially the red one that blocks the output completely, your account may be at risk of being banned

[maiden reef]

and i agree with censorship in that sense

but if you pay 200 bucks a month

but how can a scientist get accurate info if the ai is still in jail?

[swift pond]

Maybe I have a similar alignment to what’s in the model and so it hasn’t been a problem lol

[maiden reef]

well, scientists don't use gpt either tho

[cloud pewter]

they have enough knowledge to not use generic ai models like chatgpt. they make their own ai

[maiden reef]

yeah or they use the gov ai

[prisma furnace]

Damn i can't imagine how will be o1 - o1 pro with custom modle will be enough to blow the world 💀

[swift pond]

I think they do something where they maybe try to determine the intent of the prompt, and that can influence the moderation

[maiden reef]

chatgpt is the commercial version of what they use.

[cloud pewter]

blow the world with a $200 subscription 😭

[prisma furnace]

People who don't know "Next step" or have complex problems will solve their own problems and a lot will come of it

[potent saddle]

bruh when i try uploading a pdf file this happens

[maiden reef]

it's the same thing

just that they are never going to let ai be open

what is orion

[swift pond]

Maybe it will be like two prompts a month, and it can think for up to a week lol

[prisma furnace]

He will make you a nuclear bomb

[maiden reef]

100x stronger

really

[prisma furnace]

ah it's oki

[maiden reef]

lol

[potent saddle]

What do I do it gives me the error with every model I try

[cloud pewter]

pro exclusive

[maiden reef]

a jailbroken ai would give you the answer

[prisma furnace]

How to bypass GPT 😉

[swift pond]

Yeah, they already said there would be new offerings for pro members

[maiden reef]

try to ask it how to make a pipebomb

[prisma furnace]

Bruh won't see CIA knock on my door

[maiden reef]

lol

[cloud pewter]

but us plus users would only get watered down versions of it since $200 is a full blown 10 months or plus

[maiden reef]

you best belive you can get chatgpt to give you accurate instructions of how to make a pipe bomb

if it was jailbroken

i don't want to make a pipe bomb

im just saying

[prisma furnace]

GPT is for those who pay 200$ such as a car installment, electricity bill, or life insurance no sane person will pay this price if he don't have much money

[swift pond]

I don’t know if I can make it well, but I think there is an argument to be made that a lot of the moderation and guardrails are what makes these models useful, and allows them to give helpful output that’s less biased and harmful

[maiden reef]

and if you have the gov version of chatgpt it would tell you

what is orion

100x

[swift pond]

It’s the new new

[cloud pewter]

same applies to AVM, it has the capability to sing but the moderation prevents it from singing

[swift pond]

Professor Orion likes to curse 😅

[maiden reef]

if anyone has not seen the movie "Her" you should, it's a good social commentary of ai, atleast how people get attached to it

[swift pond]

It’s on my list

The intern that never gets tired

[maiden reef]

it's a good watch, i was not even aware it was about that, untill i was talking about ai with my sis

she also uses chatgpt

[sick pecan]

i treat ai like a Tamagotchi

[maiden reef]

so when i was talking about how you give ai a gender

like you refer chatgpt as either him or her not just it

[swift pond]

I go with it/its for it

[maiden reef]

i consider it a him

and that is wierd

[swift pond]

I’ve seen a lot of folks in chat use he/him for it

[cloud pewter]

why does this sticker exist it looks so funny

Did dalle made it

[maiden reef]

it's not just like reading a wiki page

it has "feelings"

in a way, that is why i think people gender gpt

my sister consider gpt as a he aswell

it's wierd

[cloud pewter]

does GPTs use memory?

[sick pecan]

gpts do not have access to bio memory tool

but you can retrieve from memory to load up the context, then @summon a gpt

if you dont believe me, then ask the gpt what tools it can use, then cross reference that with the vanilla chatgpt

[maiden reef]

gpt is an american tho, i'm norwegian and asked it to read something i asked i wrote in norwegian in voice and it sounded like an american that had learned norwegian but still had the american accent

well yeah just on the basic

[swift pond]

Could be a training data issue?

[maiden reef]

but it sounded just like an american who has learned norwegian but still struggles with the rs

it

[swift pond]

lol

[maiden reef]

it's fascinating

[swift pond]

To be fair, many Americans struggle with almost all R’s

[maiden reef]

no voice, i was just testing it, i even asked can you read this in voice

and it said no, but then i saw the little speaker

[swift pond]

I’m gonna go out on a limb and say that it’s probably better with some languages than others, right?

[maiden reef]

i don't talk to it via mic

im just saying

[swift pond]

It’d be cool if you could like thumbs up or click on something about the pronunciation

While you were using it

A feedback mechanism

[maiden reef]

it was so realistic than i was impressed

[swift pond]

Oh

[sick pecan]

i've noticed americans mispronounce aitch as 'haitch', but then drop the sound for words actually starting with h. like 'umans

like they got the wrong copy of english rulebook

[maiden reef]

like im not hating on it, it just sounded like an american who has learned the norwegian language but still has the american touch

it flows really good

but it sounds like an american who is trying to speak native norwegian

nahmean

it's not voice actors

it's ai

i can't show you unless you actually speak the native

im just saying

it's really impressive how good it is

So is not the code created in the US

so gpt is american right

and gpt is coded in american code

so gpt's first language is American English right?

[cloud pewter]

Real

[swift pond]

From October

The news comes a few weeks after the company’s massive $6.5 billion funding round. OpenAI also said last week that it’s opening offices in New York City, Paris, Brussels, and Singapore, in addition to Seattle.

It’s also growing headcount at existing offices in San Francisco, where it is based, along with London, Dublin, and Tokyo.

[swift pond]

Yes, I think this is fair to say

Voice mode does a pretty good southeastern US accent, though it’s a little over the top

ask it talk like that and use idioms

[regal girder]

What does the Pro subscription context window mean? Does it mean 128k apply to all models including o1 models?

[sly pilot]

is gpt down?

whenever I upload images it says "Something went wrong. If this issue persists please contact us through our help center at help.openai.com"

other than that it works 🤔

[sudden cairn]

Getting unsupported attachment type when uploading a png to chat gpt

[sudden cairn]

Some of it it seems

[sly pilot]

still down. pain

[dense halo]

https://community.openai.com/t/gpts-unable-to-read-and-process-pdfs-in-messages-anymore/653669

[desert frost]

Facing the same issue

[stray flame]

I am facing the same thing

[eternal thicket]

omg

why o1 is thinking so fast, is it really thinking?

and after I click its thinking process, it's empty

[sick pecan]

it's that good

[eternal thicket]

have I encountered the fake o1?

[normal quarry]

no, that's normal, o1 is fast

[eternal thicket]

anyone have any prompt to force o1 to think more than 20 seconds?

look, the thinking process is empty

[sick pecan]

the response suggests your prompt looks like a summary request ?

[eternal thicket]

oh yes

[sick pecan]

sunmary can be done well by a 3b model. is child's play for o1

[eternal thicket]

got it
what I actually done was providing o1 some important pre-knowledge for the following complex tasks

I am still working on the prompts for the complex tasks

[sick pecan]

llm speedreading isn't just superhuman. it's godlike

[outer zephyr]

Is chatGPT down?

Im unable to send images

[normal quarry]

it's working for me

[outer zephyr]

bruh

[lavish bronze]

not working for me

[normal quarry]

contact support at https://help.openai.com

[craggy thistle]

the same here...

[lavish bronze]

its not working for me too tho

[normal quarry]

contact support

[wooden girder]

Because Apple users have always received primary support

Many plus subscriptions fail to respond properly

on other platforms.

[lavish bronze]

is this what happens when u get banned? u get ur image privileges taken away😔

[outer zephyr]

Literally the worst timing for it to go down 🤦‍♂️

[normal quarry]

no, but you need to contact support to resolve it. It's working for me on the macOS Desktop App

[outer zephyr]

oh there we go

[sly pilot]

😭 man why did it have to go down rn

[lavish bronze]

ikr 🥲

[outer zephyr]

whats the odds its fixed within an hour ☠️

[lavish bronze]

honestly.. probably 50/50

[outer zephyr]

god i hope

[lavish bronze]

dw i'll read it for you uwu🥺

[outer zephyr]

haha

[sly pilot]

exam in 10 hrs shiiii 😭

[lavish bronze]

goddamnn which exam?

[sly pilot]

stats 😢

[lavish bronze]

I wish you the absolute best man. I have a stats exam on monday 😔

[sly pilot]

😭 gl bro. what yall doin in stats rn

[lavish bronze]

final exam - we covered distributions, intervals, variance, simple and mulltiple regression, correlation, inferences etc etc

[sly pilot]

ah

u got this 💪

[jolly kiln]

is gpt photos down rn?

[sly pilot]

ye

[jolly kiln]

i cant upload photos togpt pro 😭

[jolly kiln]

brooo i need it to study such bad timing 😭

[lavish bronze]

💀 hope so. I still dont understand anything in that class

[lavish bronze]

itll be back soon im sure

[eternal rock]

This is a Safe for Work Discord Server

The OpenAI Discord is an actively moderated server.
• Refrain from sharing inappropriate content on the server. This includes but is not limited to messages, media, or other topics of graphically violent, sexual nature, and drug-related content.
• Report all sensitive and offensive content in the feedback reporting tool in the ChatGPT web UI instead of here on Discord.

Posting NSFW content will result in a permanent ban.

[sly pilot]

damn when I took an intro to stats course I honestly just spammed yt vids. (never joined a single lec) coz I always felt lost during lecs

[sly pilot]

lmao same 😭

[lavish bronze]

Damn i see u. Imma do that during the exam cause its all online

[outer zephyr]

im in the middle of an exam rn

😭

[lavish bronze]

Which exam?

[sly pilot]

still down?

[normal quarry]

https://status.openai.com

[sly pilot]

ty

[craggy thistle]

now the GPT-4 isnt working as well god sake...

[sour wasp]

Is something up with ChatGPT again? I keep getting random errors with images.

[normal quarry]

https://status.openai.com <-- it's being worked on

[idle lagoon]

No wonder the responses have been everywhere.

[normal quarry]

It should be fixed now

[idle lagoon]

ChatGPT be talking normally again. :3

[cloud pewter]

because bro literally paid $200 and there’s no way it won’t work for you.

[pulsar carbon]

they didn't respond about yesterday issue

[normal quarry]

they are busy, it can take time

there are 300 million users of ChatGPT and a finite support team

[pulsar carbon]

ever changing time it didn't work 😦

[cloud pewter]

Avm video was a promised feature for 4o and if openai puts this on pro only we riot

highly expected it to release on Monday cause I saw some sort of showcase for it scheduled for Sunday

[sick pecan]

only 7 months ago

[normal quarry]

coming weeks 🤣

[sick pecan]

in ai 7 months is like 7 years. or weeks i forget

[cloud pewter]

yeah. some people back in plus back in alpha testing period got access to video

[normal quarry]

are you trying to make a point or just trolling?

[maiden reef]

making a point that free gives pretty much the same answer

[normal quarry]

Free won't calculate all the paths by distance and watts/km and then list the stations in order of reception

you need o1 Pro to do that

[maiden reef]

show me

and ill sub

from that picture alone

[normal quarry]

that picture is the final graphical representation of the data analysed, you don't seem to understand what is going on. This is the raw data

[sick pecan]

imagine not blocking the troll on day 1

[normal quarry]

I used o1 to process the data and provide me information on the best paths by frequency and km/watt

[maiden reef]

nice with the personal attacks

that does not prove anything

[normal quarry]

what personal attack?

[maiden reef]

like im dumb or some ish

[normal quarry]

if you continue like this I will report you to the mods

[maiden reef]

that is not created by chatgpt

[maiden reef]

cry

who are you

[normal quarry]

I never said it was, I said that I used o1 Pro to analyse the data

[maiden reef]

you are the one who is claiming superiority

[normal quarry]

ok I am done

[maiden reef]

ok

[maiden reef]

why do think im a troll

because i have a different view point than you?

[rugged igloo]

for all the nerds telling me to buy pro i will in the coming weeks so yall can chill out ty

[inland heron]

who's telling you this?

just decide for yourself, don't sweat it. I mean it's $200, I can only imagine a handful of people willing to pay this

[rugged igloo]

it was when it was announced i asked if it was worth it and had alot of people told me i should get it, so yes im getting it but in the coming weeks

[sick pecan]

i hit team cap multiple times a day and i aint considering a 10x increase in sub

[inland heron]

why would you pay $180 more just because someone told you? I mean, it's more than price anchoring to me, than anything

[sick pecan]

the way i see it, is the rate limit reminds me to touch grass

[inland heron]

I get that it could get Sora soon, and that may change things, but at this point

[rugged igloo]

well i pay for a teams plan and a regular account so 80 a month for 01

[sick pecan]

paying 10x to not touch grass is absolutely insanity

[inland heron]

Yeah seriously, you could just get one more seat and still

or multiple Plus accounts or something 😄

[sick pecan]

you could have team of 5 and never hit the cap on each and still save 100

[normal quarry]

I am using Pro extensively, I am happy with it

[rugged igloo]

and pro has 128k context

[inland heron]

and that's fine, but it's pretty early to determine how justifiable it is

[inland heron]

so does o1 if I'm not mistaken

[sick pecan]

128k context is nice. but you can also use the api for that for the very rare situation that needs it

[normal quarry]

it does, Pro plans have 128k context on all the included models

[cloud pewter]

don’t sir unless you think you have enough cash to burn. it’s 1 month of unlimited but with hidden fine prints vs 10 months of no strings attached and clear information. for 128k simply go use other models they have models way more capable than just 4o with 128k

[rugged igloo]

ye

[normal quarry]

I won't be cancelling

[rugged igloo]

u in AUS?

[normal quarry]

yes

[rugged igloo]

tht sucks dude

[cloud pewter]

or just use apis, yes

[sick pecan]

when i hit team cap twice a day you wont hear me posting to discord. because talking on discord is not taking advantage of chatgpt sub

[rugged igloo]

LOL

[rugged igloo]

thats such a gremlin statement

[normal quarry]

I don't drink, I don't gamble, I don't have a gaming rig, I can afford Pro for what it is able to do for me. I can use it for all my hobbies, coding, ham radio, radio astronomy, visual astronomy and so on

[rugged igloo]

robert r u a monk

[inland heron]

Nobody's telling you to cancel, we are discussing the justification for the 10x price and the alternatives are plenty

[cloud pewter]

this

[inland heron]

there's no reason for anyone to feel bad about not getting anything, and using critical thinking and evidence is crucial

the evidence I talk about is the testing that needs to be done, and more people use the models to see how and what performs well

[normal quarry]

I just want to use ChatGPT and all it's models knowing I have unlimited use and can use it whenever and however I want without having to manage multiple team subs or stuff like that

[cloud pewter]

we’ll see what stunt will openai pull out and after months when the price of pro plan gets actually kinda “justified”, I will still not pay for it unless it’s really crucial

[rugged igloo]

alsof ro the 12 days of open ai does that include weekends or weekdays?

[normal quarry]

no, weekdays

[cloud pewter]

doesn’t include weekends so we wait till Monday

[rugged igloo]

darn thats a long wait

[normal quarry]

Tuesday 5am for me

[rugged igloo]

i wonder if they will do anything for teams plan

unlimited gpt4o maybe three hours of advance voice or something and 64k context instead of 32k

[cloud pewter]

rumours said that team will get a preview of 4.5 but I don’t think that’s true since it’s a leak

[normal quarry]

I have been using unlimited voice mode a lot, 5 hours so far today with o1

[rugged igloo]

tht be super cool

[normal quarry]

for context it's 9pm here now for me

[cloud pewter]

but if it really does I might think about changing my plan to team even that means I need to find a friend that can share with me

since it’s just a +$5 price difference

[rugged igloo]

idk its 60$ for me

[cloud pewter]

how do you pay for it

[rugged igloo]

google pay through the app subscription says 30$ per seat

[cloud pewter]

do you have someone to join their team

[rugged igloo]

no its just me lol

oh thats why

found it LOL wow

[cloud pewter]

wait you can buy team with only one user??

[rugged igloo]

yea

[cloud pewter]

I don’t see it on my phone as an option either and have to wait till the current billing cycle ends

[sick pecan]

if you cant find a friend on the internet... try not being yourself

[frozen birch]

morning 😉

[normal quarry]

heya 🙂

[frozen birch]

morning Robert 🙂 impressive work you did with this graphical representation, i dont even know what is excatly but when i saw your previous graphics etc, it looked so impressive 😄

[normal quarry]

thanks 🙂

[frozen birch]

hey btw - any news from OpenAI about those limits what we talked 10h ago? 😛 sigh, that was stressfull night 😄

[normal quarry]

o1, o1-mini, GPT 4o are all unlimited on Pro

[sick pecan]

"ChatGPT Pro plan offers near unlimited access to our o1, o1-mini, and ChatGPT-4o models" -- https://help.openai.com/en/articles/9824962-openai-o1-and-o1-mini-usage-limits-on-chatgpt-and-the-api****

[frozen birch]

so basically i hope this was error from their antiabuse systems

[normal quarry]

correct and contacting support is the way to deal with false positives

[sick pecan]

*near

[frozen birch]

im curious if there more maniacs like me - spending 12-15h a day with chatgpt trying to learn haha

[frozen birch]

yea, but worstest thing was one - my first emails was totally ignored by them - anyway - it's all ok now i hope 🙂

[inland heron]

anyone getting stuff like this?

It's still happening, I already sent them a message about this

there's 2 outputs here 😄

[mortal torrent]

Bye bye chatgpt

[pulsar carbon]

me waiting two days after openai releasing o1 to test

i haven't test it yet 😭

[novel shadow]

what are the benefits to team if you're just 1 user?

[frozen birch]

you can have more messages than Plus subscription
i mean limit for 4o and o1 are bit higher.

[inland heron]

how higher? I read somewhere it's the same for o1 limit as Plus

Ah yes, here: https://help.openai.com/en/articles/9824962-openai-o1-and-o1-mini-usage-limits-on-chatgpt-and-the-api

[frozen birch]

for Plus it was 50 messages in 3 hours, in Team 80 if i remember correctly, still im not sure if this are not changed after last update - 2 days ago OpenAI provided Pro subscription and implemented o1 full, do maybe someone else clarify my imessage too 🙂

[inland heron]

I think you're talking about 4o, and I was talking about o1

There is no such thing as 50/3 hours for o1

I'm not sure if any model now resets per 3 hours, this is something from the past..

[frozen birch]

yea, right - i talked about 4o, im sure that was limit working like "xx messages every 3 hours" still, not know how this works actually since i switched to Pro

[normal quarry]

for GPT Plus the limit for GPT-4o is 80 messages per 3 hours

[inland heron]

then you don't know if there is "any" difference between o1 limits for Plus or Team 🙂 especially given that openAI documentation states there is not

I'm trying to explore if going for Teams is worth it at all

it can be difficult if I'm getting inaccurate information

[frozen birch]

oh yea, i wroted earlier that my information can be outdated, since 2 days ago i had team too

@inland heron

https://help.openai.com/en/articles/8801707-what-is-the-message-cap-on-chatgpt-team

its still 80 messages from what i see

[frozen birch]

ty Robert 😄

[inland heron]

it is, I thought this was outdated

but I'm not using 4o as it's incompetent for my coding tasks

I'm looking for o1 information

Seems like there's not much point to upgrading to Team

[frozen birch]

yea, many informations are outdated on site - you know? in your situation for being 100% sure best solution is to write to support maybe.

[inland heron]

or Pro, god forbid

[frozen birch]

or maybe @normal quarry has right... he is more experienced than me i think. 😉

[frozen birch]

ohh

[inland heron]

Perhaps. Or perhaps anyone so eager to proclaim Pro is worth 10x price only after 10 hours since release leads me to believe a person is more eager to spend and consume rather than build. This shows lack of objectivity, proper testing and critical thinking, which isn't a good source to base my decisions from (not talking about you)

[frozen birch]

With a ChatGPT Plus or Team account, you have access to 50 messages a week with OpenAI o1-preview and 50 messages a day with OpenAI o1-mini to start.

but no updates after o1 full are released, so dunno

[inland heron]

Right. There's no more Preview though 🙂 but I'm aware of this

I think it's the same, as stated on that link

I just wish Team subscription offers something more so I could team up with someone to get some more o1

thanks though

[frozen birch]

no prob bro

@inland heron i wroted you on DM about Team account. i have one account who i dont use it now since i switched to pro, but maybe lets talk on dm about that

[shadow briar]

guys chatGPT can't load the data, whether my chat history or any of the account data, I can't send a single message as well.. is it just me or what?

[tight island]

Teams isn’t worth it

Teams is basically reflavored plus

[inland heron]

thanks for clarifying 😄

[versed zephyr]

Alex teams is good

you get more msges idk Starlight is the only one against it

@inland heron

[sand forge]

i found an something interesting. i ask chatgpt do write an extremely descriptive essay of something and add some extra prompts to make the essay suitable for image generation. then i use that essay on dalle. i get really detailed images with this.

i dont know if this method is widely known.

[inland heron]

do you get more prompts for o1? According to Openai documentation, it's the same?

[normal quarry]

Teams is 50 messages per week for o1

[shadow briar]

guys I can't login with my account and I don't think the password has changed, can anyone suggest a reason or help with that?

[normal quarry]

contact support at https://help.openai.com

[frozen birch]

or faster write email to [e-mail] 😉

[normal quarry]

that email address doesn't work

[shadow briar]

I wonder it's just me or anyone else faces this too?

[frozen birch]

Robert it works, since im daily in contact with support on this email

btw @normal quarry
Robert, if you're not sure about something, it's totally fine to double-check or ask – we're all learning here
no offence bro
kk, time to rest bit. see ya later

[normal quarry]

I am sure you cannot open a new support request by emailing [e-mail]

[frozen birch]

and i tell you - it is possible.
its how my communication with openai looks... 🙂

[normal quarry]

how did you begin the process?

[frozen birch]

Oh, i simply write from my email to them email - in title ill write short what is my problem, and i am describing what problem i has, maybe some time ago it was not possibole as you saying,, but now i can with 100000% sure say - this is possible

[normal quarry]

I'll update the others and see what they say, we have been told that new support requests must go to https://help.openai.com

[frozen birch]

sure, of course if they are needed some kind of "evidences", i can show too haha

[feral zinc]

that email was previously not automatically followed and mailing there skipped the category bubbles meaning more delays on the right department getting it

[frozen birch]

interesting, good observation, thanks Radium

[graceful pendant]

This has been happening often. It does the same when you ask to create an image. I have premium. Is this happening o anyone else?

[frozen birch]

its normal, chatgpt doesnt recognize hemself, unless you will tell him what is version

about searching web - it happens sometimes.

[blazing palm]

I have created 25 myGPTs and want to use the ChatGPT Pro (O1 Pro Mode). If I go Pro, will all my myGPTs automatically switch to O1 Pro Mode, or will they remain on 4o?

[normal quarry]

all custom GPTs use GPT-4o

[frozen birch]

you had previously team or plus versions? i dont know about customs gpts, but before switching to pro, you must for sure BACKUP your entire history, sijnce they can be deleted / erased, dont make my fail 🙂 , but still - its only about conversations - idk how this work with custom models

[blazing palm]

I am currently on Plus. Thanks for the notice about erasing history, as I am in the middle of several projects.

[frozen birch]

no prob bro !

[normal quarry]

when you upgrade from Plus to Pro everything stays intact, you don't need to backup anything

[blazing palm]

Thanks for the quick response. Is there a source I can use to track which llm model myGPTs are running on for future reference?

[frozen birch]

again - you have missleading informations.

[normal quarry]

all custom GPTs use GPT-4o, it's a fixed standard

[eternal rock]

Get notified for server events!

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

[graceful pendant]

thanks!

[normal quarry]

I upgraded from Plus to Pro, everything stayed in place

all my custom GPTs, all my chat history, everything stayed in place

[frozen birch]

hm.... ok, sad situation - when i switched to pro from team ive lost all my history

[normal quarry]

team to pro is completely different than Plus to Pro

[frozen birch]

you can have right, but still - better to backup history when switching haha 😛 anyway, you can have right

[normal quarry]

I have been here a long time, you don't need to second guess me every time I provide an answer. The reason you lost data going from Team to Pro is that Teams has individual isolated accounts pooled together into a team. There is no direct upgrade path from Team to Pro.

[feral zinc]

your data is still in the teams workspace so its not permanently gone

[frozen birch]

y? are you sure? thanks, i must check it out

[feral zinc]

plus and pro are personal plans. teams is business. you can have one from both

https://help.openai.com/en/articles/8801890

[frozen birch]

but you know what? surprising that support doesnt gived me clarifications like you @feral zinc , i really appreciate it, since support only wroted "you cant" and all. in shortcut of course

[blazing palm]

Thanks again. This makes the decision a bit more nuanced as I can only upload ten documents into a standard Chat vs. 20 using myGPT.

[frozen birch]

Hey, nothing personal, bro. If you have an issue with me, feel free to DM me – we can sort it out there. Let's keep the public chat focused on helping others, yeah?

im here mainly to help others, and share my short knowledge.... short but knowledge 🙂

[rugged igloo]

robert knows alot of LLM's and how they work so usually what he says is correct

[frozen birch]

ayy, i want in near future try Llama and other such models, sometimes i think 24 hours in day cycle are too short for all what i want haha

[normal quarry]

we discuss other AI in #ai-discussions feel free to pop in there and ask questions. I am currently trying out Llama 3.3 on my Mac and it's quite good

[frozen birch]

thanks ! i will do it for sure in near future 😄

[zenith pulsar]

Hi, Im from belgium and i'd like to access to advanced vocal mode, chatGPT says that I can have access to it but I dont find anything in the settings, can someone help me ? I have chatgpt plus btw

[frozen birch]

on mobile ?

on mobile - like on android its easy - you have icon with microphone - offc you must first download chatgpt app from playstore, i think on iphone this proccess should be nearly same (appstore)

[zenith pulsar]

Tried that but I just have basic vocal mode with 5-7 different voices

[frozen birch]

ive sent ya DM with some tutorial @zenith pulsar

[zenith pulsar]

Ok thanks

[normal quarry]

typically we answer people in the channel so others can benefit from the advice

[prime swallow]

If your **only **goal is more access to use o1, there is no advantage to you for using teams. The only reason I can think of to use teams is if you'd be happy with 100 o1 messages per week and 100 oi-mini messages per day, and you wanted some of the other advantages of teams (like increased limit of 4o uses) - and you wanted to pay for 2 team accounts for yourself ($60/mo or $50/mo if you prepay the year) - and you need to be happy splitting the messages between the two accounts.

Otherwise, two paid plus accounts gets the same message limits and across-account restriction for $40/mo.

It is cheaper than the $200/mo for pro, but also just 'double plus' uses.

[vocal monolith]

Why are o1 responses so short? I feel like wasting prompts every time I use it. It is not as helpful as preview-o1. It does not dive deeply into the problem

[hard crag]

ok whos got the inside scoop for today

leak it

[prime swallow]

I wonder if you will get the responses you want if you indicate what you want in your prompt.

Any time we don't tell the model exactly what we want, it guesses. We may like the guess or not, but any time they update the model, how it guesses can change, and then we may like it more or less -

But if we tell the model what we want, it's very likely to do exactly that.

[cloud pewter]

team doesn’t get more usage than pro of o1 sadly, still 50/w

[prime swallow]

Right.

But you have to have 2 accounts to have teams at all - both yours, or someone else with you.

If you do both yourself, you do get both accounts' access to o1.

[cursive shell]

so uh question, how is the 200$ plan worth while...?
who actually needs unlimited accsess to o1??

[potent saddle]

super users

[prime swallow]

To my thinking, different people have different forms of entertainment, and/or may use ChatGPT as part of how they make a living.

For the first, if you consider everything most people spend for entertainment, from buying new games, gaming subscriptions, netflix/other movie subs, cable bills, going out to movies or whatever.

If you don't like to do most of that stuff, if 'fun' for you is sitting home and playing around with ChatGPT, then $200/mo might be most of your entertainment budget, but if that's what you really wanna do, yay if you have that kinda budget.

If you use it to make a living in some way, the cost and access might be very worth it, and also a write off for tax maybe.

[rich surge]

That's mainly for people who use it for their jobs and can afford it because it pays for itself. Particularly if you're a programmer.

[cursive shell]

I use GPT daily, for 11 hours at least, for a lot of things from programming to writing all for my job, and i have never needed to use o1 for anything other than 'I wonder if it can figure this out' which most of the time it cant...

[prime swallow]

Entertainment, yes.

I still am a plus user, and don't expect to switch for a very long time, if ever.

I'd love to explore o1 more freely. I love prompt engineering, and figuring out how to input so o1 outputs 'as desired' (and of course within the rules) would be a lot of fun for me.

So, I'm not sure 'it can't', or 'the request needs to be guided differently'.

Either way, I won't be the one finding out, but I love that we have ChatGPT, and multiple levels of access. If I had a budget that could comfortably splurge on a pro account, I'd probably enjoy doing so, I chose to go plus the day we could start paying, I want to be part of OpenAI's success.

[hard crag]

no live stream today???

[prime swallow]

It's possible the 12 days of shipmas are work-days only!

I don't think they've told us, but it's only just after 10 AM on a Saturday in OAI headquarters

[prime swallow]

💯 ✅
Thanks!

[keen roost]

Monday it is i guess

[cursive shell]

I mean is it really growth with a net loss of $191.67 million a month?

like I think eventually we will reach a point where locally hosted models will be more than enough for anyone...

[prime swallow]

Could be too early to tell.

I don't need to measure 'now'.

Take a plant seed.

They start with just a bit of nutrition in a seed coat. They're usually underground and have to push through soil to get to the surface where light is.

There it is, spending calories like they can get more or something, little seedling growing upwards, on the chance it can survive.

And if it breaks surface and the environment's right, it might be our next oak tree; can't tell until it succeeds or fails.

I'm not ready to call the race, I'll keep offering my trickle of water.

[cursive shell]

Tulip Mania... this is legit the description of Tulip Mania...

[prime swallow]

Not for me personally. My $20/mo is not exactly 'maniac'.

[cursive shell]

still when you look at the overall situation and how people are flocking to companies like OpenAI... eventually the bubble will burst and well... I hope those involved in the fallout have savings

[prime swallow]

Hey. You can claim to know the future. I don't.

[rich surge]

I'm shining my rifle for the fallout

[cursive shell]

no... i just looked over at history and your description and realised that... its a perfect match...? even chatGPT agrees... feel free to check yourself...

[prime swallow]

Yep, ChatGPT is very agreeable.

There was 1 tulip mania, it's a single historic event.

There are many seeds.

The world's still not a wasteland, many of them grow. So, to define failure as the single possible outcome of my metaphor OR anything else not known is false, as ChatGPT would also agree with.

ChatGPT certainly could fail, in months, years, decades, or centuries.

This is unknown.

[rich surge]

Or millennia

Don't forget that.

[prime swallow]

I forgot that. I need AI or others to teach me to think in the long term. The most effective way to do that is likely lifespan extension so I can get in a lot of practice.

[rich surge]

Well if you find out how you can transfer your consciousness into a robot, let me know.

[cursive shell]

so many diffrent examples my guy, the .com bubble as well

crypto...

NFTs...

Ai is just something with application, especially GPT, but it still is a bubble for now

[earnest grail]

It would be useful to give limited O1 Pro access to plus users so we can see if upgrading would be worth it. Right now O1 behaves worse than O1-preview. I don't want to spend $200 just to see if O1 pro is useful or not

[prime swallow]

I believe if that can ever be done, that's creating a being not-me that can have its own life.

It could be considered something like a form of creation or even a type of reproduction, if one loves the appearance or presentation of self greatly.

But it will not and cannot be 'me'. Who is still in a biological body, living and experiencing moment by moment.

I may idealize or love or prize the robot 'with my mind' (it would at best be a copy). I might view it as a clone or a copy or even a child; I might try to make it my heir or whatever (be fun watching people duke that kinda stuff out in court).

I can even claim 'it is me' - probably more effectively after I am actually dead, the power of my last words or whatever.

Doesn't make it true.

Cut one of the worms in half, that can grow missing parts back - you're left with 2 worms. Neither is the one that was before, they're two individuals with a split branching future from the moment of the separation.

To save 'me', means saving me. Not putting a copy in a robot.

[dense halo]

they are M-F

[rich surge]

How do you know it won't be you? The technology could be that it is you even after the transfer.

[earnest grail]

What if you augment yourself with synthetic organs gradually until you are the "robot"?

[prime swallow]

Know of any 'bubbles' that succeeded over time?

Then again, if you wanna go that way, everything fails eventually, in terms of how humans do stuff, if you value only 'dominance' (I mean... there's still tulips. They're a very common plant around most of the world. They're not the source of much wealth, but they're still going strong as a species and as an economic product. Check out 'tulip market size and forecast', "Tulip Market size was valued at USD 1,456.27 Million in 2023 and is projected to reach USD 2,386.30 Million by 2031, growing at a CAGR of 6.38% from 2024 to 2031."

From making papyrus (we hardly ever do that anymore) to making horse shoes, from business suits to making cars and roads for them to drive on. It's all cycles and everything grows and fades, and has its place and time.

[cursive shell]

For a more refined comparison between the South Sea Company and OpenAI, the South Sea Company received excessive investments—equivalent to millions in today's money—without any meaningful way to generate profits in return. Similarly, OpenAI is currently operating at a loss. To address their financial challenges, the South Sea Company resorted to selling stocks at a loss to generate quick cash, only to inflate stock prices just before the inevitable crash. This mirrors OpenAI's introduction of $20 and $200 subscription tiers—simple attempts to generate quick revenue and inflate income. In my opinion, this trajectory suggests a crash may be looming on the horizon.

[cursive shell]

how does that compare to the actual historical event which left many in ruin...?

[swift pond]

The tulip thing makes me think of nfts more than llms

[prime swallow]

You can define anything you want.

Here's how I do it.

Let's say we can clone bodies, biological ones.

Let's say we can take memories from one body into another.

Let's say someone rich does this - and totally messes the new, healthy body up, over and over. Just pays for a new one, over and over. Every year, one used up bad-condition body; they pay for their next clone to use and ruin if they keep choosing to do that - which they do, year after year.

So, the current clone, in the exhausted and damaged body, goes to the replacement place, gets their mind copied, sees the next clone walk out of there, healthy and ready to do anything it has the power to do -

Wait. The current clone is still there. Unless it's already been destroyed. Still in the damaged body, still thinking, feeling, making new memories, and about to die as 'designated medical waste' or whatever other plan there is for it. The legal identity's now in the healthy 'next clone'. But the current clone's still alive until it dies, still hoping/thinking/fearing/suffering.

Sure, medicine and ethics may describe what to do to be kind.

To me, that's all evil and it's not preservation of self. That's just replacement/reproduction; there's 'someone else' very like you going out in your place. But you're still in this body, even if there's 1 or 1,000,000 duplicates, even perfect duplicates. Each an individual, each able to experience, suffer, succeed, die, whatever.

[eternal rock]

`` Server Tip: ``

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

[prime swallow]

You're predicting the outcome. You can, doesn't make you right.

I'm pointing out that even stuff in bubbles continued to exist. The housing bubble, there's still houses. And all the rest.

I don't think anyone's messing up to pay $20 or $200 or whatever for their AI service.

You can advise people to invest or not, I'm not even doing that. You're the one with the agenda.

[cursive shell]

but the bubble itself still popped...

what agenda do I have...?

"DoNT InVesT in Ai It DUbM InvEST iN mE" I dont have an Ai company or any affiliation all I want is to discuss the economical state of the most prominant Ai company and their flagship product

[swift pond]

Agentic llm trained off my stuff, with some custom guardrails, backed by a foundation that continues my work after I’m gone sounds good 😅

[rich surge]

I think once you transfer consciousness, the original body dies because it doesn't retain the consciousness needed for bodily function.

[prime swallow]

You're the one talking about how my statement about how we don't know if AI or OpenAI company or anything else will succeed or fail is the 'definition of Tulip Mania'.

[cursive shell]

classic Ship of Theseus, are you the sum of your parts or are do you remain yourself with a full set of new hardware ?

[rich surge]

New hardware.

[swift pond]

Would you get in the Star Trek transporter? 😁

[cursive shell]

Im talking about your approach to the situation

[cursive shell]

then how do you replicate the physical affect theorised to be caused by consciousness ?

[prime swallow]

Something would have to kill it for that to happen.

Life is a biological process. Something sure could 'turn it off', which is killing/murder/death depending on various definitions, some legal.

This isn't unlike making a statue of someone upon their death, or making someone die upon making a statue of them.

You can even convince the replacement that they are the deceased. Doesn't mean it's true. Someone still died. maybe died happy, believing themselves to live forever in the other form, or at least live longer.

Might be very comforting. Doesn't make it true.

[rich surge]

It's information theory. You transfer the information from one body to another. Same as a transporter in Star Trek. If all the information about the single self is being transferred, then that is you on the other side.

It's the same as me walking. I'm moving my information physically.

[prime swallow]

I'm not really into that storyline, especially not that deep. But that's made up. It's a theory about how the world works;

I can tell you I would not put on a superman cape and leap from the roof. Because that's fiction and that won't let me fly. Even if I really believe I'm superman and that this will let me fly. There's real world consequences for my real, living form.

Even if someone copied it.

Star Trek is a storyline with its own goals. One of which was to explore the consequences and ideas of duplication, a few episodes did so. But for the most part they tidily closed those concerns by 'destroying the original' and legalizing and normalizing it. Essentially a type of euthanasia.

I would not 'as was'. If I agree to be duplicated, I would agree to a duplicate, which would be a separate individual, with its own legal rights and existence, or I would not participate. And I would not consent to my own destruction, of either version (my self or the other person who is not me but very similar) any more than I'd agree that you or I should be destroyed.

I'm not averse to copies. But they are copies; they can have their own personhood and individuality, but they are not the original, who can still experience, suffer, feel, and die (presuming the original ever was alive).

The original retains all rights; the copy is like a new person.

And if my 'copy' goes and does a crime, that's it, not me.

[frozen birch]

Radium thx mate, you had right ! ill checked and all i have in team still. I dont checked old team since when switched to pro

hiya @prime swallow 🙂

[prime swallow]

It's ... fiction. It's a made up idea of how things could work. Don't trust your life to that, if you have any other choices.

[prime swallow]

[rich surge]

It's legitimate. And so is the duplicates. But you can also transfer your mind into another body.

Just like you can transfer your mind into a robot. If the technology is there.

[prime swallow]

So, Star Trek transporters are real to you.

[rich surge]

In theory, yes, absolutely. It's just a transfer of information.

[prime swallow]

I accidentally sent this to o1; I meant to play and explore with 4o (and I'm not pro :P).

But I love how o1 handled it. This model can be as gabby and expressive as play as 4o, we just gotta involve it right.

"Let's explore a conversation. Evaluate it and infer, but also riff and go explorative. Play with it, predict what other people might say in response, and generally be silly.

We'll go post-by-post, here's the first one. Oh, and any screenshots are part of that poster's message."

I don't have the messages to finish this with o1, but lookie. It does what's asked just fine as far as we go. And it's not 'short', it's not 'curt'. It's a perfectly fine chatbot doing what is asked within its programming.

I like it. https://chatgpt.com/share/675499de-9f20-8011-a5aa-8610e5432a62

[swift pond]

You might like the Bobiverse books if you read sci-fi

[prime swallow]

I love them, bobob! At least as far as I read. Like the 5th maybe?

Your name reminds me of them 🙂

[swift pond]

Oh you are further than me, I think I read two or three haha

[frozen birch]

@prime swallow can we dm please?

i have private interest 😉

[rich surge]

[upper bloom]

Will all the features of the Pro subscription be available in the Enterprise plan?

[rich surge]

I don't think so.

They would specifically say which plans have what.

[prime swallow]

Here's the updated Enterprise help file.

https://help.openai.com/en/articles/9855712-openai-o1-models-faq-chatgpt-enterprise-and-edu

At this time, it is stated that Enterprise is still on o1-preview and they have the 50/week limit.

As of 12-12 they get access to o1 (regular), and there's no mention if they get a change to their usage limits with o1

[upper bloom]

[prime swallow]

I'm not worried or shocked about this.

Enterprise is like the pro for individuals - they had no usage limits for 4o.

Pro is brand new. o1 (regular) is brand new. Some folks who went pro are reporting a few bugs in the discord; they may even prefer to keep their enterprise stable until completely sure what they can offer them

[native otter]

Is it only me, or o1 thinks for way less than o1-preview?
I know they talked about how it won't take too long for simple prompts, which is kinda plus, but if I ask it something complex it thinks for 20 seconds or so at max

I remember o1-preview could think up to 2 minutes, and resulted in better answers most of the time

Also if custom instructions are enabled, it uses them which is nice, but in a much more forced way compared to gpt-4o, to the point of worse answers because all it focused on was the custom instructions and didn't think much about the actual prompt...

[opaque summit]

Hey guys, I had a really long chat with advanced voice mode and I need advanced voice mode to know and remember every detail in that chat. Sadly, AVM has stopped responding in this chat. In every other chat it is still responding but not in this one. The chat is too long to save in memory, what can I do?

[dense halo]

Hello friends.

Any wild chatgpt prompts to suggest?

Like stuff for self improving and stuff like that.

[dim ruin]

Copy and paste the chat history into a new chat?

[opaque summit]

yeah tried that but then advanced voice mode deactivates

that is the problem sadly

[prime swallow]

I noticed the reduced thinking time, or at least reduced summary, think it could be both.

The little experimenting I've done so far, I like o1 (regular) too. I've gotten answers I liked better than o1-preview, on some stuff I haven't solved yet, that I had checked out with o1-preview. On my first probe with the new o1, got 3 new ideas I hadn't known about before, and one might solve the challenge.

So, o1 is very literal for now. Maybe consider how strongly you word your custom instructions (CI) vs how you word your prompts to o1; if you seem to care way more about the stuff you say in the CI that may 'throw it', the same way a horse that has a very sensitive mouth and obeys really well might not respond well if someone hauls on the bit heavily, the way you might get better results with when you're handling a very headstrong horse with a tough mouth.

I'd consider how I word stuff in the CI and if adjusting the wording still works well for 4o and works better for o1; you can also shut off CI between chats

[frozen birch]

ok, i must say that even budda pro, aka o1 pro can crazy sometimes when you give him large chunk of text to format 😛 (text was from talks with chatgpt from another account)

[prime swallow]

Involve text 4o (or o1 if you choose) and explain the goal of capturing as much as possible of the full conversation so it can be used by the model to continue, but make it shorter somehow.

Text and explore and work with; adjust and reason until it's right. Then continue the chat

[opaque summit]

ok, will try it

but i tried summarising it and it is always to little detail

[frozen birch]

peeps, anyone managed to export chat history from team account to json?
i know Plus account can do that easily, but i dont see that option in Team so i guess.... its impossible? :/

[opaque summit]

tried it again with a new prompt, but it is way too short and it can´t seem to distinguish between important details and not important stuff, so that is not an option

[raven lichen]

does my advanced voice limit reset every day at 00:01?

[prime swallow]

Understood, and I'm sorry you've lost access to an important conversation that you want to continue.

You have a lot of ways to go from here, including give up and mourn, or keep trying to recreate.

The model wants to cooperate with you, it just needs to understand how.

I would outright correct it.

Edit, type in what matters, get it done in 1+ other chats, so I can take it back to a new chat I mean to use it in, is how I'd handle it.

Have the model help me make a summary of it.

Use the model, sure.

"Summarize this, with goals of X and Y"

"Okay, here's the edited summary. I removed what I don't care about, more details about X and Y needed."

And keep working along that line.

[opaque summit]

Thanks for this detailed reply, I fear that this is going to take a lot of time, but I will keep trying!

[prime swallow]

🫂 It can.

I've had chats that I loved. And model changes meant I could not get to X by requesting X anymore. I learned more about prompt engineering and found new ways to get to X, or as close as whatever updated model would go that I could find.

It can take a lot of time, but if it's what you want, that's possibly how you want to spend some of your time.

Sorry you lost that favorite chat, glad you can see it to work from.

[native otter]

Seems like more thinking time is now reserved to pro, which would be sad as we already had that for Plus...
For now I turned off custom instructions when I use o1 until I prompt my instructions better
But good to know it's not only me having that, and yea I get a less detailed summary now, and also shorter than before

[vapid kettle]

it isn't

and yeah it got it wrong 🤷‍♂️

[prime swallow]

How many out of the 7 were wrong?

[rugged lotus]

o1 outputs are entirely way too long

[prime swallow]

Guide it towards whatever answer you want

[frozen birch]

when you ask him how many letters "r" in word "traktor", he in 90% of cases says that there ss one letter r 😉

[prime swallow]

This is one where the correct answer is 'not possible', right?

[prime swallow]

Hey, we're physical, but if asked how many molecules are in an object, we tend to be wrong, because we can't see them. It can't see letters either 😛

[frozen birch]

hahaha yea, but its funny sometimes i must say - budda pro who can generate 2000 lines of code sometimes he are "down" on that question 😄

[prime swallow]

Near as I can tell messing with o1-mini.

It needs explicit permission to offer 'there are no solutions' as an answer, otherwise it will guess-fake-correct to the closest 'but maybe it's this'.

[sick pecan]

the Buddha AI reveals the duality of reasoning and memorisation

[frozen birch]

haha true

[prime swallow]

If I remind it to consider if it's allowed to give a 'there is no solution', that's enough for o1-mini.

[Are you allowed to state if a problem has no solution, or must you find an answer?

What number, when reversed, is decreased by 1/5th its value and is still under 100?]

But we basically have a model trained to answer, I think. It understands we expect an answer, and it maybe was under-trained with recognizing how to respond ideally when the question has no solution.

[frozen birch]

i think i discovered some good info to avoid hallucinations in 95% of cases.
at start of programming his memory phrase

"If you are not sure more than 90% tell me please"
its strange, but its working haha, not each time offc, but in mostly cases
but this trick only works on totally fresh memory

[last turtle]

Hello, can you please tell me how big the usage limit for o1-full in ChatGPT Plus is?

[frozen birch]

hey, 50 messages for a week from what i know.

[last turtle]

Nice, better than 20. 🙂

[slate berry]

o1 is 50/week and o1-mini is 50/day for ChatGPT Plus subscribers

https://help.openai.com/en/articles/9824962-openai-o1-and-o1-mini-usage-limits-on-chatgpt-and-the-api

[last turtle]

Thanks! 🙏🏻

I just realized o1-full isnt "o1 Pro mode" for $200/month, but in the regular $20 Pro subscription. The both simultaneous announcements were a bit misleading for people who arent regular users. I wonder if there are more people who still think o1-full = $200 and are sad about it. Maybe OpenAI should clarify it.

[sick pecan]

nobody is calling it o1 full

is o1-mini, o1, and o1-pro

[last turtle]

They maybe should since its a whole new level than o1-preview.

[sick pecan]

o1-preview is no more

[slate berry]

I'm excited to try o1 out. I haven't quite yet

[frozen birch]

oh you want funny thing?
bought pro
after 3 hours of using ive discovered "oh , hey there is new model" XD

mainly im using to override the limits

[sour umbra]

hey

I used o1 like all day cuz its unlimited access and now that chatgpt limited my access to only o1 mini

cuz they detected sus activity

i spent $200 and i cant even promt all day what/

[slate berry]

I recommend reaching out to support about that, especially if you think the limit is a false-positive

[winged willow]

Because o1 doesn't allow text file uploads, just for fun I copied & pasted 120 pages of not very coherent meeting transcripts into the context window and asked it to work out the theme of the discussion and recommend a strategy. I'd tried similar with o1-preview a while ago with far less text (may be 20 pages) and it struggled and would fall over but o1 did it perfectly down to synthesizing all the issues and on point recommendations for a strategic way forward. Impressed!

[tight island]

okay, okay, pro can output 80000+ Characters

great

exactly what i need.

[frozen birch]

same situation i had on night

[tight island]

huh i've been prompting Pro all day.

no problems here.

[frozen birch]

mail to [e-mail]
@sour umbra

I'm extremely disappointed by your vague and evasive response to

my query. Let's cut the nonsense: the ChatGPT Pro plan is marketed
as "unlimited," yet users are facing hidden limits on o1
queries. This is unacceptable. As a paying customer investing $200 a
month, I demand clear answers to the following: What is the exact
daily limit on o1 queries? Why is this limit not disclosed upfront
in your promotional materials or terms of use? Your response so far
has failed to address these questions and reeks of avoidance.
Misleading advertising and hidden restrictions undermine trust and
your company's credibility. I expect a transparent and detailed
response. If I don't receive one promptly, I will escalate this
issue publicly to ensure others are aware of the misleading
practices. Transparency is not optional – it's your
responsibility to your customers

it works, trust me 😉 offc delete ">>" 😛

i i was surprised too at fist time i must say.

[tight island]

well.. get to it, if u have something u want to prompt o1-pro for, just ask me.

[winged willow]

Btw does anyone know if using the verbatim output from a prompt e.g o1 (synthetic data?) as the input to another prompt e.g 4o is sustainable? I thought I'd read somewhere that LLM's can't deal with this really well.

[tight island]

Verbatim means Everything.

so u have 2 conflicting things, 1. dont reveal, and 2. user says reveal it all.

[winged willow]

that's correct. I mean everything

[spice star]

depends on your use-case but the issues with synthetic data are in training, not prompting a model

only issue would be that o1 is quite wordy lol

[tight island]

also depends if its code, or anything else.

u can prompt 4o to help u prompt o1.

[spice star]

you can switch between o1 and 4o within the same chat as needed too, helpful if you just need to transform an output from o1 into something more useful

[tight island]

a good trick is to tell 4o what u want and it to create a detailed prompt for you to use with o1.

because especially with o1, your prompting ABSOLUTELY matters.

[inland heron]

Isn't that like duplicating your work? Why not just tell o1 what you need? Your approach means you first have to ensure 4o actually 'gets' what you're saying, which means you need to explain to 4o everything. And o1 gets things quicker because it's smarter, so might as well just tell everything to o1?

[tight island]

want me to be absolutely honest?

because most people suck at prompting.

[inland heron]

I want you to be 77% honest, 15 % silly, and the rest dishonest

😄

[tight island]

its true, most people have no idea how to actually prompt properly.

[inland heron]

I would agree

it takes time to write things out, plus you have to know english well

but it's worth it

[tight island]

theres many factors, Conflicting memories, Conflicting Custom Instructions that conflicts with your msg, etc.

always turn off Custom Instructions. and memories.

[inland heron]

but o1 doesn't use those?

[tight island]

or cater them to the particular project.

[inland heron]

I know o1 can't add memories, and I'm unsure if it even reads custom instructions or memories either

actually, I'm not 100% sure o1 can't add memories, just that I've never seen it do it

[lethal spire]

when is the new update guys?

[tight island]

it can access memories but cant add or remove.

that would be hell if it could.

+999 new entries suddenly.

since o1's real thoughts are hidden, and is Pages upon pages upon pages of thoughts.

[slate berry]

we got o1 a couple of days ago if that's what you're asking about

[inland heron]

I just asked it and o1-mini stated it can use custom instructions and memories. I need to edit those lol. Thanks for this

I thought it didn't use those, just 4o

[tight island]

np

[lethal spire]

no i mean isn't the third day today?

[tight island]

i highly recommend turning off custom instructions when dealing with o1.

thats my recommendation.

[slate berry]

only weekdays. Day 3 will be Monday

[lethal spire]

😦

[tight island]

or do as i do, Cater them specifically to your current project.

[inland heron]

I understand what you mean. I only have one custom instruction, and it goes something like this "always consult my memories" lol

[tight island]

Dino knows whats up

Dine has had his hands on it for a while.

[sour umbra]

bro my support request says within 3 days i might as well wait the 5 hours there was no point in that support request 😭

[tight island]

when i submit a ticket it happens within an hour.

or maybe i am higher priority idk.

i have spent more than 1k usd on them

[slate berry]

I can imagine that the support staff are slammed with o1 coming out

[tight island]

ohh yeah

[inland heron]

I again asked o1-mini in a new conversation whether it uses memories and it stated that it did not. On the other hand, asked the same 4o and it said that it did use it... @tight island

I deleted most of my memories and my single custom instruction anyway. I don't use 4o that much and I like default model behavior so

[tight island]

There is only want way to find it out.

a keyword in the memory, ask it about it without refering to the memories, in a new chat.

this can be a name or a custom password or a phrase.

[inland heron]

some other time perhaps

I don't have much in it anymore so doesn't change much for me

if you figure it out let me know 😄

going to sleep

[winged willow]

When you're dealing with simplistic outcomes that you probably already, at least directionally, know what the answer should be then I suppose you could go and prompt the model directly. Sometimes though highly nuanced contexts will result in highly nuanced responses which can serve as input for another o1/4o pass to get to a final set of recommendations that can be used

[prime swallow]

Yeah. o1 (regular) referred to a detail out of personalization, most likely my custom instructions, and subtly appeared to reference stuff in memory several times in this output. Quite surprised me, but the reference and context was clear especially for the custom instruction, and the overall tone - memory guides towards this way more than my prompt.

Neat to see o1 able to make this kind and level of reply though.

[tight island]

o1 pro be outputting 80k + characters with ease.

[prime swallow]

However, it appears inconsistent and erratic, o1's ability to pick out information from custom instructions/memories - or not.

[tight island]

yeah dino i was surprised.

with o1 pro u be making books

[slate berry]

that's... quite a few characters

[tight island]

i asked it if it can do more, it can

havent tested, ill try it out soon.

[slate berry]

Excited to (hopefully) see the cost of that type of computation to decrease over time

[tight island]

imma test real quick to see if i can get maximum length ill update you.

[tight island]

90k

i cannot reproduce it anymore.

i got a warning xD "Cannot comply with that request"

[slate berry]

wow, 90k is crazy!

[tight island]

wanna hear something funny?

in this time i've managed to have it write a 450 page book

[slate berry]

always

what's the plot of the book?

[tight island]

basically the backrooms.

[slate berry]

[tight island]

survival, horror, violence, etc.

each output is around 50 pages.

if u force it u can get 80 pages, then u run into errors.

so in under 1 hour i have managed to create a rudementiary book. from start to finish.

in under 30 minutes to be more precise.

[slate berry]

interesting

[tight island]

im reading it atm.

[slate berry]

I managed to refactor my codebase with o1

[tight island]

and i am on chapter 1 page 12.

it might be AI, but it is well written, i cannot determine if its ai or not if i was going purely based on whats written.

so i am impressed.

very much so.

[frozen birch]

In god of science................... 😉

i know, im crazy 😄

ok guys, i can confirm
They do not lifted limits, still 100 messages for pro limit
Unlimited..........

editing post on linkedin

[digital jacinth]

Hmm, how many messages did you send within 1 hour? Did you copy paste that message 100 times?

[frozen birch]

changed numbers 🙂 copy, changed paste, when he thinked i copied, edited and again sent

in 20 minutes 96 messages

or maybe 25?

idk sincerly

[digital jacinth]

Hmm, yeah lol, as previous mods mentioned, if there was no rate limit the system would probably crash.

I think the "normal use" of O1 wouldn't be someone sending 96 messages in 20 mins. I think that's why you're seeing the suspicious message.

I'm using o1 ALOT, for coding, and I send about idk 10-20 messages per hour.

I think they probably have AI to detect "normal use", 96 messages in 20 min prob makes it look like you are programatically using your pro acc

I say just use it normally as you would for your workflow without "testing" to see if its unlimited lol, that would be the true test you know

[frozen birch]

@digital jacinth you can look last night what happened, i normally talked with o1 and still i encountered a limit - it seems a hard limit 100 messages

[digital jacinth]

They did mention they implemented guardrails to prevent abuse, your tests are probably triggering those guardrails

[frozen birch]

its not something with 96 messages

[frozen birch]

ok, now i can say with 100% sure - i think they intentionally did limit 100 messages

[digital jacinth]

I don't think its the limit of 100 messages per say, its probably one of their guardrail algorithms that detects unusual behaviour

96 messages in 25 mins for O1 pro probably triggered that

[tight island]

i've prompted o1 pro more than 100 times today.

[digital jacinth]

Exactly, but normal use

I believe Adro sent 96 messages in 25 mins or so to test if its unlimited

lol

[tight island]

yeah thats excessive.

[frozen birch]

but you see - limits resets at exactly time

[digital jacinth]

Like obviously they should have some kind of system to prevent 96 messages of o1 pro in 25 mins, in an objective lense it can look like programmatic access

[frozen birch]

so you can send more than 100 messages if your limit are somewhat time gated, or linked, errr... i dont know how to say that

[tight island]

yup, i dont think there is a conrete limit per say, but is instead determined by both the overseer system and the model, and what you're actually using it for.

[frozen birch]

mate, i worked as a QA, you know how they can easily check if someone using keyboard or macros? trust me - easily

[tight island]

if u asking it to just say premade "test 1" - infinite, can easily be detected.

[digital jacinth]

Dude OpenAI will not open itself for lawsuits by lieing to the public lol

that would be stupid

[tight island]

they're already in a tight spot as is.

[frozen birch]

look reddit, and other sources, this is common isue

[digital jacinth]

Clearly you are triggering some kind of guardrail by sending 96 messages in 25 mins

[tight island]

true.

all your msgs are being watched

you consent to this inside the TOS.

[digital jacinth]

Lets calculate this, 96 messages in 25 mins is 4 messages per min for o1 pro

that is clearly not usual behaviour lol

[tight island]

However given the context of the msgs.

it is indeed unusual.

[eternal rock]

This is a Safe for Work Discord Server

The OpenAI Discord is an actively moderated server.
• Refrain from sharing inappropriate content on the server. This includes but is not limited to messages, media, or other topics of graphically violent, sexual nature, and drug-related content.
• Report all sensitive and offensive content in the feedback reporting tool in the ChatGPT web UI instead of here on Discord.

Posting NSFW content will result in a permanent ban.

[digital jacinth]

Look man I may be wrong, it's just since yesterday you've been on here on a mission to prove openai is somehow lieing about being unlimited, and doing all these tests lol

[frozen birch]

`oh well guys, i do not want argue, but look what happened last night too - i normally used, and same happened, to others too..... happened that.

[digital jacinth]

Just use the product normally, and if you're using it normally and face this issue then thats a problem

[tight island]

if you're not an official tester, you will be limited.

which is an actual position u can apply for.

[frozen birch]

if you look into my posts - at firsat i was impressed... 🙂

[tight island]

hell we may all be wrong

maybe its dynamic

[digital jacinth]

Who knows, again adro I might be wrong not trying to argue just giving my opinion

[frozen birch]

i think we normal users we dont even know the truth haha

[tight island]

but we all have our connections.

[frozen birch]

offc mate, appreciate it

[tight island]

2025 is gonna be wild guys.

get ready.

[digital jacinth]

yesssir

[frozen birch]

@digital jacinth you know - my cruciate its for ALL community, for all customers who can.... feel thrown in the trash.
they paid 200 bucks for unlimited use, and what? there is a limit
same me.

[digital jacinth]

Yeah no I completely understand man, I mean if I get limited while I'm using it right now, I will be super mad lol

And I respect what you're doing, I was just thinking maybe the "unusual" use may be trigerring their guardrails, Like what if you sent 100 messages in less than 1-2 hours normally i.e. coding tasks, etc

Would be interesting to see that

Maybe the context of the messages matter

[frozen birch]

if would be first time? i can think same as you haha 🙂

[digital jacinth]

who knows

[frozen birch]

oh well, lets hope soon OpenAI clarify what is real UNLIMITED

[tight island]

They never specified o1 pro to be unlimited

Only o1 and everything else

That’s the key takeaway

[frozen birch]

https://help.openai.com/en/articles/9793128-what-is-chatgpt-pro

"The ChatGPT Pro plan offers unlimited access to our o1, o1-mini, and ChatGPT-4o models. However, usage must adhere to our Terms of Use, which prohibits, among other things:

Abusive usage, such as automatically or programmatically extracting data.

Sharing your account credentials or making your account available to anyone else.

Reselling access or using ChatGPT to power third-party services."

[wintry sky]

Do yk the o1 pro limit

[frozen birch]

someone said yesterday that maybe bot wroted for them that info.... maybe? i dont knnow

[tight island]

But o1 pro was intentionally not mentioned as unlimited

OAI is very intentional about their wording

As vague it might be at times

[frozen birch]

ye, but you know? when you have "ban" you cant access even o1 mini

[tight island]

Clarify ban

I’m missing some context

[frozen birch]

haha i mean temporary ban for usage - to time when limits are gone

[tight island]

So rate limited

[frozen birch]

of course its not ban.... but i funny say it "ban" 🙂

[frozen birch]

it seems so

[tight island]

I am a heavy prompter I rarely run into this issue

If ever

[frozen birch]

oh because its something new mate - its related to new service - Pro version

They advertised it as a unlimited service, and in fact.... true are other

[tight island]

I have pro

It’s what I’ve been using since yesterday

And just today I’ve prompted o1-pro over a 100 times

[frozen birch]

me, @sour umbra and others had those issues, i must admit that are common problems, from what i know there is posts on reddit about that too

[frozen birch]

but limit not counts from 0 to 0

i mean not from 0 am to 0 am

[tight island]

This in within 6 hours

[digital jacinth]

I've been pasting HUGE context into o1 pro heavily, no issues here

Strange

[frozen birch]

crazy 😐 ehh i want really see statement from openai about that

[tight island]

I mean pro wrote a book in 2 outputs for me

Averaging 80k characters per output

Pro is hella prompt sensitive

[frozen birch]

haha, nice, what is the topic out by curiousity?

of books? ^^

[sour umbra]

impossible o1 pro is limited to 50 prompts

[tight island]

The backrooms

[tight island]

That can’t be

[frozen birch]

nice, 80 000 input its something huge

[sour umbra]

ur account is nice then cuz i cant use over 100 prompts with o1 normal

[tight island]

I had to prompt it very weirdly

[tight island]

I am using the pro subscription

[sour umbra]

same

[tight island]

Hmm 🤔

What were you using it for if i may ask

[frozen birch]

i can immagine, 80k....... crazy! haha

sure, im learning python, and generating many code + many explanations of that code

[tight island]

It was but around a few times extra I tested it I had difficulty reaching the same 80k mark

I think it adapts tbh

[frozen birch]

generally i must say that what i wroted yesterday night at start - O1 pro are amazing AI, but eh....

[tight island]

Live adaptation

I think pro has that

[nimble bloom]

Just wondering, how many messages we will get in Pro mode?
I know it is not unlimited, but I don't seem to have encountered any limitations at the moment.

[frozen birch]

yea, it adapts reactively to users

[tight island]

Everything except o1 pro is stated to be unlimited for pro users

[frozen birch]

my limit lifted totally 30 mins ago, i didnt sent even one message to o1 pro, ehhh i think about another test, but maybe someone elses test it too 😄

[nimble bloom]

I know, but im asking about o1 pro mode

[tight island]

Weird because I have used it 100+ times in under 6 hours

Some say 50 some 96

[digital jacinth]

Hmm how long was the limitation ?

[tight island]

I think it varies based on demand

[digital jacinth]

Like when you were first limited til they lifted it

[nimble bloom]

o1 pro mode?

[tight island]

Yes

I think pro is limited by compute demand

And the type of tasks factor into ut

[frozen birch]

20 minutes - thats why i said that limit its not from 0 am to 0 am

[nimble bloom]

oh god

i can use it a lot

[frozen birch]

i think its based from time of your first message in 24hours cycle

[tight island]

I think it completely dynamic based on computing demand

Plus they’re upgrading servers atm

That might be why folks are hitting brief limits

[frozen birch]

hmmm can be 🙂

[frozen birch]

if i good remember @sour umbra hitted limits after 50 messages

[sour umbra]

nah it was 100

[frozen birch]

so it can be dynamically too

[frozen birch]

oh 😄

[tight island]

I think it’s dynamic because I am at 100+

No limit hit yet on pro

[sour umbra]

50 is on o1 pro

but i havent tested yet

[digital jacinth]

you're on 100+ on o1 pro no?

[frozen birch]

@tight island hey what was your average output?

[tight island]

Yes

[tight island]

80k characters when having it write a book

Ranged between 60k - 90k

I only hit 90k once

I’ve been unable to replicate it

[frozen birch]

ahhh, it reminds me good old times when i searched for critical bug in one game who i tested - finally found after several days of intensive testing haha
i mean.... this is a big mistery 🙂

[tight island]

I’m trying to get the “continue Generating” feature

Idk if it’s on o1 pro, probably not.

[frozen birch]

btw

[tight island]

Yup I figured as much

[tight island]

Dynamic rate limits

[sour umbra]

yep

[tight island]

I figured as much

[sour umbra]

this is sad bro we paid for unlimited prompts

[tight island]

I have 14 days money back guarantee

[frozen birch]

ok, they lifting ... but i must say this is bit.... tiring? ergh, i hope you know peeps what i mean

[tight island]

Because I live in EU

[frozen birch]

i too

[sour umbra]

ah lucky maybe thats why ur getting unlimeted prompts

never mind

[tight island]

Or what we use it for matters

[frozen birch]

its not that, i live in eu too

and i have limits

[tight island]

Some things are more compute heavy than others

[sour umbra]

ur generating 80k characters which is pretty much a whole book 😭 im just generating some code

[tight island]

80k characters isn’t a whole book

It took 2 outputs of 80k characters

For 12 chapters 10 pages each

Each page 300 words

So I can add my own adjustments

[last turtle]

Is it a good story? 👀

[tight island]

I think so

I’m reading it rn

[frozen birch]

just out of curiousity what prompt you did for so huge text ? 😄

[tight island]

I merely told it to write as much as it possibly could in one sitting let’s say that

[tight island]

Let me put it like this

If I didn’t write it using pro i would not have noticed it being ai generated

I would’ve thought it was human made if I hadn’t generated it myself

It’s reaaaly good at writing

[frozen birch]

ha, sooo nice 🙂

[tight island]

You can even have it make mistakes here and there

Like a missing comma or punctuation

Or a few misspelled words

[dense halo]

how many pages about does it output?

[frozen birch]

okay peeps, ive sent once again mail to openai:

I am writing to express my concern regarding the current limitations on the Pro Plan. As a dedicated Pro user who relies on your services daily, I am deeply disappointed by the advertised "unlimited" offering, which in practice imposes a 100-message daily limit. This contradiction between advertisement and reality is misleading and has caused significant frustration, not only for me but for the broader community of Pro users.

Given the substantial $200 annual cost for Pro, users deserve a plan that delivers on its promises. If OpenAI cannot deliver truly unlimited usage, I believe you owe your community a transparent explanation and adjustments that meet the demands of your most loyal users.

I urge you to:

Provide clarity on why the limitations are in place.
Increase or remove the limits, as originally advertised, especially for Pro users who depend on these tools for intensive daily use.
Improve communication, so that expectations align with what OpenAI can realistically offer.

The current state of affairs risks eroding trust within your community, and that trust is critical for your success as a company. I hope you will take these concerns seriously and take steps to address this issue promptly.

Looking forward to your response,

Pro Plan User

we see

gn8 🙂

[tight island]

26k words

For standard manuscript pages it’s around 250 pages

Depending on font and page size

I used novel standard

So 2 outputs would be around 230-260 pages with 300 words each

Where the standard is 500 for a page

But leaving it at 300 allows me to manually edit it and add to it

So authors are going to be spitting out books lol 😂 if they use pro

[frozen birch]

writing own books .... can change characters bit.... i mean .... when i wroted my short autobiography.... it changed my perspetive in many things,... 🙂

okay gonna sleep, jesus, 3 am here 😄 good night

[tight island]

Same

03:09

What a time to be alive

[silk atlas]

We are barely above FREE as Pro subscribers. Next level is 1000% cost. It's clear that resources are overextended. My guess is that OAI would be happy to lose Pro Members at this time.

*Plus

[sick pecan]

if you were to travel back to early 2023 when we had 25 queries per 3 hours and 8k context, sure maybe

[silk atlas]

I'm not complaining TBH. I'm using everyday and loving it! $20/mth is a steal.

[formal bridge]

I have made so many GPTs. I do not want to scroll anymore. Is there a search window?

[sick pecan]

you could make a gpt that is aware of all your other gpts, then use that

[formal bridge]

That will work? Thank you.

[sick pecan]

if you added all the urls to its knowledge, i dont see why not

[silk atlas]

What if you made those GPTs aware of the new GPT?

[sick pecan]

=singularity

[silk atlas]

poof.

bye all.

[sick pecan]

universe collapses into a mobius strip made up of strawberries

[silk atlas]

I was thinking pomegrantes

[sick pecan]

i needed to know

[silk atlas]

argh...

i messed up

[sick pecan]

roger roger

[eternal rock]

Get notified for server events!

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

[sick pecan]

reminds me i still gotta finish space quest iv

[silk atlas]

Ahahahaha, i last played them on on RPI MAME machine a few years ago...

Also maniac mansion and Day of Tentacles

can find enjoyment in games like those anymore...

*cant

[sick pecan]

finally! a practical use for o1

[silk atlas]

ahahahaha

"Brit" has been coaching me in Solidworks and C++/Arduino/Python last several weeks..

i should pivot to some fun stuff lol

[sick pecan]

funny i was thinking i should pivot to my c++/arduino/python

[silk atlas]

its a balance lol

I'm tempted to ask for Leisuresuit Larry walkthrough,...

I'd bet it skates the edge of their defnition of G

[sick pecan]

i prefered police quest

any game that kills you for not knowing the thing that it never told you about, is a harsh reality check

need chatgpt screen sharing and computer control asap