chatgpt-discussions

[young obsidian]

Bruh it doesnt work now lol

[verbal thicket]

Hey guys how do I get access to gpt 4o?

[modest cove]

it's free just go on to chatgpt site

[verbal thicket]

Its not their for me unforunately

[gleaming grove]

So I have 4o but I don’t seem to have the new voice and video, just the old audio I had before

[shrewd wedge]

@verbal thicket i subscribed to plus and immediately received it

[verbal thicket]

hm

[gleaming grove]

This is on iOS I’m talking about

[modest cove]

maybe they need more time to roll out the feature to your country or it's busy

[vestal hedge]

is dalle3 same power in 4o?

[shrewd wedge]

some people claim to have the desktop app already, anyone here have it yet?

[tight torrent]

No 😫

[verbal thicket]

im australian 😅

[inland quest]

i got it but not in the legal ways

[shrewd wedge]

how is it?

[vestal hedge]

wait what???? desktop gpt app?

[latent ibex]

only MAC for now

[inland quest]

pretty cool, but no new voice or anything announced, just easier chatting, and there is keyboard shortcut to chat right away

[vestal hedge]

is there a page on openai.com yet?

[shrewd wedge]

cool

[modest cove]

it says that gpt 4-o is not available when it's busy to free users

[plucky pelican]

and where to download?

[latent ibex]

https://openai.com/chatgpt/download/

[azure geyser]

through chatgpt.com

[latent ibex]

"ChatGPT Desktop app for Mac is coming soon."

[shrewd wedge]

k

[azure geyser]

i got a notif

[shrewd wedge]

cool

[plucky pelican]

i got a download link but I clicked it away

[shrewd wedge]

i’ll check that way

[modest cove]

" We'll roll out a new version of Voice Mode with GPT-4o in alpha within ChatGPT Plus in the coming weeks." well good thing i am in alpha team

[latent ibex]

[verbal thicket]

I think the vision is the m ost overpowered part of the update, will make maths tutoring a breeze

[tight torrent]

Please share details on how to get it.

[latent ibex]

good thing

[inland quest]

but "coming weeks" basically says nothing, its like "soon" on memory for europe users 😄

[latent ibex]

I mean usually teachers explain it better and more easy. GPT-4 often explains it to complex

[modest cove]

xD

[verbal thicket]

nah i think chat gpt maths explanations r better

[latent ibex]

memory still isn't in EUrope yeah

[verbal thicket]

u can ask it to go in depth too whic his nice

[modest cove]

i tried it with gpt 4-o it has so much spelling mistakes

[latent ibex]

teachers do the same lol

[verbal thicket]

ye but u wouldnt ask a teacher to go in depth on the same question 10 times

[latent ibex]

I mean if you pay a tutor they shouldn't mind

[verbal thicket]

even then lol id personally feel embarrased asking the same thing 20 itmes over

[latent ibex]

the new update where gpt has more emotions gpt might laugh you out

[analog sedge]

Hi. So, I have a custom GPT i created a few weeks ago and want to check out 4o. Is there a way to convert or clone an existing custom gpt to the current model version? Thanks.

[hot steeple]

Hi

[verbal thicket]

fair

[inland quest]

but that laughing part i dont know.. it didnt make me feel comfortable 😄

[hot steeple]

I have a free account, I had 4o access but then I logged out and back in again and now I lost access

[inland quest]

i dont know if i want laughing robot, a bit creepy

[modest cove]

did anyone see the demo in openai blog post about gpt 4-o?

the robot typwriter thingy

[inland quest]

ye

[modest cove]

i would say the demo was fake

[verbal thicket]

bruh wish we could pay with crypto for premium

[modest cove]

i put the same prompt in demo to gpt 4-o the image was similar but the words had so many grammatical errors

[inland quest]

i think its still using old model for generations

[modest cove]

yeah but my question is did they improve the dall e model or gpt model? or both?

[inland quest]

for me the voice part was crazy, its too much human for me, i want to feel like im talking with a robot when im talking with a robot

[modest cove]

yeah, that was impressing

[inland quest]

now it seemt like they were talking to some AI girlfriend on announcment

[crystal tapir]

Not available in browser yet?

[inland quest]

i guess its the gpt model itself, they said it was reworked with all modalities

[young obsidian]

Bruh

[young obsidian]

It's a showcase

[hot steeple]

for me i had it in browser on free, then logged out and back and it went away, its a bit buggy, i was told to just wait

[modest cove]

they alteast had to mention that it's coming out in a few weeks, or it's modified

[grizzled badge]

Bring back the thumbs up button for the prompts! I like to praise my AI overload. This GPT4o is SICK!

[inland quest]

i wonder how it compares to turbo madel

[crystal tapir]

I was testing sending files on Android. Could only send 3 pictures.

[sacred bloom]

How use GPT 4o ?
Pls help guys )

[sharp pilot]

The GPT-4o blog post seems to suggest that the text roll out is for both free and Plus members. I haven't received it on my free (expired Plus) account. I'm tempted to jump to conclusions that there is some order to the rollout.
Can anyone confirm if they have received access yet on a free (or expired Plus) account?

And as a technologist I am curious if there is some reasoning or rationale behind staying completely quiet about rollout strategies. Is this one of those things where there is an intellectual reason they don't comment while they are doing rollouts even though it causes confusion?

[cedar glade]

The functionality of real-time conversation hasn't been rolled out yet. We just have the new model.

[inland quest]

staying quiet, makes you wonder when you will get it, so you check more times, visit the website more, ask about it etc..

if you knew exactly when you will get it, you wont be checking it on website, asking about it here, i guess its psychological factors

[keen meadow]

The point of the rollout is to monitor their systems as they add more users. If they notice something they don’t like, they can pause and get on the case. Publishing a rollout schedule like it’s a video game release wouldn’t make sense.

[sharp pilot]

I don't understand how it would confuse anyone for them to say how they are doing their rollout, or literally even if they are still deciding.

[low pewter]

does anyone know if our custom gpt's auto update to 4o?

[sharp pilot]

I know things can have weird repercussions so maybe there is a rationale for their silence here that I am just not seeing, but most companies do some kind of updating with their communities.

[indigo charm]

Is GPT4 Free available now?

[sharp pilot]

By appearances no. By the announcement it is rolling out but it sounds to me like anyone who has confirmed back that they have it has Plus.

IMO they have only scheduled the roll out for free users even though they announced it as happening.

[tranquil gulch]

Is it just me that cant access gpt 4o?

[keen meadow]

Almost no companies publish how they’re rolling out new features. It just comes when it comes. Having to come up with some crazy table showing regions, user tiers, early access privileges, timeframes, products etc would be madness. Their staff don’t want to draw that up, nor would they want to stick to it.

In practice, you usually take it slow initially, add one or two large user groups, and if everything is golden, just open the floodgates. If there’s an issue, you usually catch it quickly.

[inland heron]

gpt4o slowed to a crawl

[haughty forum]

Have you guys t ad the fine print yet?

[inland heron]

looks like it's speeding up

[haughty forum]

*read

[sharp pilot]

I don't think we disagree on the rollouts but perhaps we do disagree whether the standard is to announce them as happening without giving status updates. Circumstantial "users confirming back" evidence shows that for free users it hasn't started yet, honestly.

[haughty forum]

We recognize that GPT-4o’s audio modalities present a variety of novel risks. Today we are publicly releasing text and image inputs and text outputs. Over the upcoming weeks and months, we’ll be working on the technical infrastructure, usability via post-training, and safety necessary to release the other modalities. For example, at launch, audio outputs will be limited to a selection of preset voices and will abide by our existing safety policies. We will share further details addressing the full range of GPT-4o’s modalities in the forthcoming system card.

GPT-4o’s text and image capabilities are starting to roll out today in ChatGPT. We are making GPT-4o available in the free tier, and to Plus users with up to 5x higher message limits. We'll roll out a new version of Voice Mode with GPT-4o in alpha within ChatGPT Plus in the coming weeks.

from the gpt4o page

[keen meadow]

I’ve been doing software dev a long time and never heard of publishing updates in such a manner. We have enough to worry about without treating users like they’re a department manager. The initial rollout could have been to OpenAI and Microsoft employees, whereupon an issue was found, and now they’re dealing with it. The rollout is likely underway in some shape or form.

[glossy fjord]

Hi, I want to ask a question
I don't understand the thinking

You have exhausted your GPT-4o limit. Replies will use GPT-3.5 until your limit is reset at 16:54.

and I have to understand how
like 16 hours and 54 min
or mine time in 2 hours and 54 min
or 16 min and 54 sec?

[sharp pilot]

Lots of plausible internal rollout stages or active work being done like you say, or in the spirit of your good example. But I don't think it's so unheard of to have a community manager or some form of updates happening alongside the deployment. Clearly that's not how they do things.

[inland quest]

in 2 hours

[glossy fjord]

thx ^w^

[molten tundra]

I give credit the 4o for coding work today is using a lot less filler words and just giving me good information

[glossy fjord]

yes, I also tried 4o entering stories here and something came up that GPT 3.5 couldn't summarize right away, but it had to be done after about an hour of writing

for me, GPT is 4o better and can understand more

[dreamy fiber]

So gpt4o is free for everyone?

[keen meadow]

Well it’s just a matter of opinion, but personally I wouldn’t want to work for a company that treats a feature rollout like a sacred pact with our customers and has community managers pinging us non-stop on Slack. Absolute hell imo.

[dreamy fiber]

As of today?

[latent ibex]

soon yea

[dreamy fiber]

Oh okay

So we still have to pay $20

[molten tundra]

Going to try it for some super advance things for sure, but so far very pleased at the outputs

[latent ibex]

nah but plus alr has it, probs comes over the weeks for free usedrs

[glossy fjord]

yes im free plan GPT 3.5 + 4o

[sharp pilot]

I've seen large communities have no-ping type rules for the visible staff, which would make sense here. It's an alternative to the radio silence.

[glossy fjord]

yes, I'm already looking forward to seeing what will disappear as interesting things or what (4O) will produce as interesting things

[latent ibex]

do you have gpt's (store) too?

[keen meadow]

What’s the point though? They said rollout would start yesterday and last up to several weeks across user types and features. So nothing has gone wrong. We’re just whining because we want instant gratification. GPT-4o was available to all paid users very rapidly. They’re doing great so far.

[haughty forum]

Weeks or months

[latent ibex]

weeks tehy said

but probs months for Europe

[glossy fjord]

it behaves like GPT 3.5, only with the functionality of GPT 4.o

Help with writing, troubleshooting and much more
Access to GPT-3.5
Limited access to GPT‑4o
Limited access to advanced data analysis, file uploads, image processing, web browsing and custom GPT

[latent ibex]

o right

[sharp pilot]

This is where our subjective difference is, right now all the tech journalists are giving their lazy write-ups that ChatGPT has the features that they have announced. So they quickly are in a murky area where they have presented pending features as existing.

[latent ibex]

i am plus user, but what they refer to with limited access? lol

[haughty forum]

Idk there's something apple exclusive feeling about this.

I also wonder how willing they are to launch what was demoed.

If you re-read the announcement it's also just voice that's coming in a few weeks, probably not voice and vision.

[glossy fjord]

im europe - czechia

[latent ibex]

I know but I meant the vision tho

whenever that comes out

probs months delay

because of Laws

[keen meadow]

Easy solution, just ignore bad or lazy journalism. People who really care about the power of ChatGPT have a Plus subscription and are too busy enjoying GPT-4o to care about a slow app rollout, vision capabilities etc.

[loud vessel]

Chatgpt 4o feels smart again but remember when chatgpt 4 was smart? Then everyday it got dumber and dumber, stealth nerfs.

[haughty forum]

Also caution about 'ethics and safety' has a lot more to do with fear of damaging IP than anything else.

I imagine voice mode will have a separate rate limit to start.

[glossy fjord]

yes Every news is not immediate, but we are surprised that I have it right after a few hours after the public

[keen meadow]

Doubt it. From my experience with US tech companies, the overwhelming majority of devs are on Macs. Prob just made sense for them to prioritise a Mac app.

[sharp pilot]

This is a slippery slope to a haves and have nots argument. Which is certainly and unfortunately adjacent to their stated goals of opening up accessibility, it's just we disagree on degrees of transparency as it happens and I don't think that gives them a free pass when they enter a murky region of features as "available" rather than "will happen".

[haughty forum]

It's a pretty well funded company to only support Mac for what's basically a web app.

[glossy fjord]

many others prefer windows or linux

[high sphinx]

I guess they put their money into other stuff

[tight sapphire]

One feature I would love they would implement would be the option for chatGPT to extrapolate information from a previous discussion if we can provide a link to the discussion.

Why would I want such a feature.

Sometimes in a convo that model gets to know the subject and style I like to communicate and I wouldn't have to do everything anew. Plus it would help battle the lag that comes with long conversations.

[high sphinx]

but windows will come later this year

[haughty forum]

I think the timing of their talks with apple imply more than just ChatGPT as a Siri alternative in ios18

[latent ibex]

source?

[feral zinc]

microsoft might also be one factor here due recently widely rolled out copilot

[haughty forum]

Yup

[latent ibex]

don't say lies if no official announcment @high sphinx

[keen meadow]

But where is transparency lacking?

https://openai.com/index/gpt-4o-and-more-tools-to-chatgpt-free/

We are starting to roll out more intelligence and advanced tools to ChatGPT Free users over the coming weeks.

https://openai.com/index/hello-gpt-4o/

GPT-4o’s text and image capabilities are starting to roll out today in ChatGPT. We are making GPT-4o available in the free tier, and to Plus users with up to 5x higher message limits. We'll roll out a new version of Voice Mode with GPT-4o in alpha within ChatGPT Plus in the coming weeks.

Seems pretty clear to me. Not OpenAI's problem that low-end tech sites are constantly chasing click-bait.

[sharp pilot]

I don't have it handy but it was in the app announcement webpage. "Later this year" is a bit disappointing, I know app porting work isn't trivial but I guess we'll see.

[loud vessel]

cant believe how smart chatgpt 4o feels... they are going to dumb it down like they did with chatgpt 4 so enjoy it while it lasts

[high sphinx]

it says on website

[latent ibex]

for sure

[keen meadow]

And I'm strongly in favour of supporting as many platforms as possible. But for all we know, their metrics show a much higher conversion rate for Mac users than Windows or Linux. I've managed analytics for mobile before and for us, iOS users spent astronomically more than Android users. Perhaps that applies to Mac as well to some degree, and OpenAI is chasing the easy wins to start.

[haughty forum]

It makes zero sense from a marketshare standpoint

[latent ibex]

now MAC can test and Windows will have a perfect version hehe

[muted jolt]

Have to think of development too. The reality is, it's easier to dev an app on mac these days than windows that will work reliably accross all devices.

[haughty forum]

I guess they can assume the majority of windows users are idiots.... that's fine I suppose.

[latent ibex]

unless you want to use vm's 💀

[haughty forum]

Compatibility issues with a web app?

[sharp pilot]

Porting apps is non-trivial but "later this year" seems a little too conservative. I'd just like screen sharing without funding Apple with dollars I can't afford to spend.

[haughty forum]

Oh no, it's so hard to access the webcam and desktop and microphone, how will people ever work remote? 🤯

Better develop for Mac only since it only has one model of webcam and microphone.

[glossy fjord]

but in Europe it's the other way around, here there are a lot of people on windows and linux and Android and also if someone has a MAC/IOS / apple, it used to be that a lot of people had a USB C mobile instead of a port for an Apple phone, so these people here were driven away
even though no one could help them borrow a cable, because the cable was an exception for 1% of Europe, here it is the opposite,
Apple is expensive and capacities change and everything here and there and many people have xiaomi or samsung and so on

yes, and it is still expected that the European Commission will impose another fine on Apple, perhaps Apple will withdraw from Europe

[sharp pilot]

Apple will never withdraw from Europe, but the EU is letting the rest of us benefit from their pressure. USB-C is a great example but it's a dozen things besides, including the various smaller issues with iOS and marketplaces that even as an enthusiast I struggle to keep track of.

[sleek aspen]

Indeed from a developer standpoint it's easier to make things for Mac and iOS since you need to account for a smaller number of devices/configurations

A windows app is fortunately in the works - but it's not easy to port over the Mac one, it's bespoke and made to work natively on MacOS (it's made in SwiftUI) as opposed to a webapp

Wrong reply 😅

[keen meadow]

I don't disagree man. If I had a company with apps across platforms I'd strongly favour parity across platforms for all releases. But I guess sometimes it's more practical to not do so.

It could be the case that OpenAI is mostly unconcerned with Europe revenue if US numbers dwarf the EU. And Mac and iPhone are extremely popular in the US.

[haughty forum]

I imagine that screen share will be an iPad exclusive feature. For all we know vision modality is too or voice. There's no way of knowing.

[sharp pilot]

Reading between the lines here but it sounds like OAI might be courting Apple for inclusion in the OS and the CGPT macOS app might be part of their courtship.

[latent ibex]

thinking about Asia, most have Android tehre too

[sharp pilot]

Occam's Razor etc. They aren't villains they are just working capitalism.

[keen meadow]

Most of my US colleagues don't even have WhatsApp installed. They're different over there. iMessage dominates completely, so take a stab at guessing how much market share iPhone must have.

[latent ibex]

lol iMessage isn't used at all in my country (West Europe)

[sharp pilot]

This is the same iMessage that labels Android users as outsiders and downgrades their attachments.

Acceptance != Acceptability

[keen meadow]

I'm not defending Apple. I'm saying that's the reality.

[glossy fjord]

I think that's the point, please stop being so clever, you're showing something that isn't really ideal

but really, I want to see how you can connect an nvidia card to that kind of mac machine 🙂 for which these things are usually trained, or open AI is not written for linux, I don't believe that open AI would work, so we will make a chip that will be for a MAC machine and that's the end of accounting from machines of the last brand

[haughty forum]

It's almost like a company should announce what platforms their product is coming to, especially when they demo it exclusively on apple...

[keen meadow]

The irony of your comment is that modern Macs decimate Windows for running LLMs locally. Unified memory up to 192GB, good luck getting that sorted out with your GPU options on Windows.

[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.

[haughty forum]

Maybe transparency isn't minimalist enough...

[sharp pilot]

In this channel (self included): consumers getting frothy about vaguely anti-consumer behaviors, developers showing that some accept it and some don't. 😅

[molten tundra]

I think honestly the reason the mac desktop version will be there and windows isnt just yet is because microsoft are in technical terms making their own competative models so just allowing a windows gpt application sorta takes away that leverage

Thats just my opinoion though

[shell umbra]

I think it's just easier to develop for MacOS since there's less you need to account for. That and the fact I don't think I've ever seen OpenAI make anything for windows natively yet.

[glossy fjord]

amd , ryzen , threadripper , epyc
the board supports 10 cards and so on
then there's nvidia, which makes computing cards for AI

at the end of the debate, it seems that according to the warning, this is leading to a quarrel

Europe is android and windows and linux, you can't understand it and understand that people here won't buy apple, so I really don't know how to write it
The USA is the opposite, I don't care, I don't know the USA, but I know that the administration between APPLE and Android doesn't work properly, and people who have an Android and an Android receiver work well

[shell umbra]

I don't think having local ai hardware matters when it comes to the ChatGPT app. You don't run GPT-4o on your own computer, it's a call to their servers.

[glossy fjord]

I think that the reason why it is not for Windows is from a financial point of value and price , so that openAI can see and satisfy the investor's expectations, there is nothing wrong with that

[keen meadow]

I don't think it's like that. It's not that US-based companies don't understand that certain demographics or regions favour certain platforms or devices. That's usually dead obvious from analytics.

It's more a simple case that they have strong dev capacity for Apple platforms - probably for internal reasons - and could easily make a Mac app for customers. Rather than delay that release for many months while they hire/develop an app for Windows as well, they're releasing the Mac one when it's available.

[solid bronze]

I am getting an error using ChatGPT 4o anyone else having issues ?

[normal quarry]

what is the error?

[solid bronze]

@normal quarry Systems are experiencing heavy load is what is keeps saying

[glossy fjord]

I also don't hear about them and I don't know about them, and nobody is interested in them because I'm not interested, if you lock it for APPLE MAC or IPAD, the Windows user will lose interest and eventually switch to something else, this is reality

I don't have an apple and I don't even want one. From my point of view, a 256 GB disk costs an unknown amount of RAM
if for the same price I can buy a 512/1 TB disk and a 12 GB frame, the expectations are elsewhere, and I still secretly unlock the android dev mod because I install my own applications, which probably won't work with Apple

[serene wharf]

hi, can someone tell me the advantage of paying a chat gpt plus subscription if GPT-4o is free?

[keen meadow]

Quicker access to new features and 5x capacity.

[glossy fjord]

GPT 3.4 used GPT 4o only token limit time

free has a GPT 4o limit on the number of words, if you exceed it you have to wait about 2 hours

[serene wharf]

mmhhhh okayy

so is it more advantageous to be a chat gpt plus member?

[keen meadow]

You've only ever accessed ChatGPT via the web app and mobile app. Neither option is being removed for anyone, so relax.

The real money for OpenAI is with enterprise customers, and they probably have a lot of Windows machines. Chances aren't bad that OpenAI is hot on the case but not wanting to overpromise anything.

[glossy fjord]

if you don't want to wait 2 hours after the word limit, sure, but try it first and decide if you want it or not, if that's your opinion, go for it

[serene wharf]

okay thanks. i am already a plus user and i saw that gpt 4o is available for free

[runic grotto]

wait i'm confused, so I have 4o option on my phone, but I dont see how it visually interacts with me? also the delay is the same as its always been.. do I not have that full capability yet?

[glossy fjord]

we fight over stupid things

I believe you have apple, but it just makes me think that we are guessing because of such nonsense as apple / windows / android / linux and in the end it will be the same, otherwise I don't care

[keen meadow]

Just ask yourself if ChatGPT improves your work productivity. For me, it frequently makes an all-day task take a few hours. Or a few hours worth of work take less than a single hour. And it can check my code quality, and help me brainstorm. The free usage level would be woefully inadequate. The $20 monthly fee is a joke considering how it boosts productivity for me. But that isn't the case for everyone, so do what works for you.

[solid bronze]

Facts

[solid bronze]

Perfect way to look at it

[keen meadow]

I'm not fighting. Just calmly responding to comments.
There isn't a conspiracy going on over at OpenAI. The simplest explanation is that they have far better dev capacity for Mac than Windows. So end of story most likely.

[sleek aspen]

Hey! I believe this topic has shifted from discussing ChatGPT to discussing Apple products in general, mind moving this conversation to #off-topic - thanks!

ChatGPT on Windows will be coming later this year - and is of course still available through https://chatgpt.com right now

[glossy fjord]

I'm not just mentioning that Europe is not the USA
windows, linux and android are often used here
and it's true that there are exceptions for Apple machines, and if it weren't for a couple of bans and obligations of the office, there are no Apple phones thanks to Europe's USB C mobile phone, many people here couldn't buy a mobile phone as an Android user because its brand was Apple and it didn't have the connector itself, and so on the way to such an Apple user ran out of battery and the android mobile had a power bank al

[glossy fjord]

I don't mind, I'll just wait and play with another AI, but here it's about the principle of the US not Europe and Europe not the US, another region, another system, different rules and different machine users

[dense halo]

What happens if we use voice mode with the new model? Is it still relying on whisper to encode the message to text and then out via TTS?

[runic grotto]

anyone know why if I have the 4o option on my phone, it doesnt seem to have the other items like video interaction or the faster latency.. is that rolling out later?

[glossy fjord]

it's been out for a few hours, this is still unknown

[eternal rock]

Beware of possible scams or fraudulent activities that you may receive through direct messages. OpenAI staff will never DM you for any transactions.

Please report any incident by sending a DM to @languid valley immediately.

[sleek aspen]

Yes, the new audio and video functionality isn't publicly available yet so voice mode will function the same for now

[runic grotto]

gotcha...

[dense halo]

Huh, so gpt4o thinks it's still chatting via text even when its voice mode. That makes sense

[sleek aspen]

The latency should be faster, but there's high usage for GPT-4o since it just came out so some delays may just be a result of server response times

When using the API the difference in speed is very noticeable

[runic grotto]

I dont even have the buttons for the video interaction like I see on the videos released so I dont think that stuff is out yet per your paste above looks like only some people got it

[sleek aspen]

The new video and audio functions aren't currently available as stated above #chatgpt-discussions message - apologies for the inconvenience!

[glossy fjord]

really good expectation that these functions will start open AI right away for all users when I take how many such 1080P or 4K resolutions are the data flow per second for AI

[blissful idol]

My chatgpt account has been deactivated and the openai still charge me subscription fee for 2 month, how to get refund?

@sleek aspen

[slim mica]

do you know if ChatGPT 4o is generating images with gpt4o or still DALLE3? When I asked it said it was DALLE3

[sleek aspen]

Unfortunately, we Discord moderators can only assist with issues related to the OpenAI Discord server. For any concerns or inquiries about any of OpenAI's services, kindly visit our Help Center at https://help.openai.com — they should be able to help with getting a refund. Thank you!

[blissful idol]

but there is no way to contact with any openai stuff

[empty glade]

Somebody else has issue with slow loading? I mean, when I send a message, it becomes really slow, the response, the scroll, etc...

[tiny girder]

This should be helpful: https://help.openai.com/en/articles/6614161-how-can-i-contact-support

[deft perch]

i was able to use GPT-4o (Python API) yesterday afternoon just fine, but since then I keep getting APITimeoutError . Anyone else experiencing this?

[burnt hare]

working fine for me in USA

[silk blade]

if someone is still looking for a windows chatgpt app:
you can install the website as a progressive web app from chrome or edge, there should be a download button in the url bar
idk how relevant that is, but when i looked here Keksu's post was current and i'd half typed this before i realised i was scrolled up

[deft perch]

thanks

[earnest frigate]

We have gpt 4o but when we get custom GPTs will they run on 4o or 3.5
cause you know free users have a 4o limit

[neat dawn]

is chatgptO globaly available

[earnest frigate]

i think 4o has like a rollout type thing

[silk blade]

I'm in the UK and don't have it yet (free account), probably rolling it out as demand allows. if they suddenly enabled it for everyone we'd use all of the world's electricity at once to generate cat pictures

[earnest frigate]

few people get it when time goes on and on

[neat dawn]

ya that could explain it

[sweet wraith]

Is anybody with the android app able to do the live convo thing yet?

[sharp pilot]

With respect (and appreciate your being here, folks do be contentious), the ChatGPT app being an Apple exclusive should be fair game as a topic for #chatgpt-discussions .

[sleek aspen]

That's totally fine to discuss! The topic seemed to have shifted beyond that though.

[sharp pilot]

And it will again, the whole platform ecosystem thing tends to be a rabbit hole.

[frigid cove]

Hello guys. I see that GPT 4.o is here. But I see also that the GPT I created is using old GPT 4. Is there something to change to allow the GPT I created to use the new version? how I can get the benefit of it?

[earnest frigate]

i think custom gpts dont support 4o yet

[silk blade]

is the apple gpt app actually a bespoke app, or just the PWA packaged as bytecode to satisfy apple's app store monopoly?
i always kinda assumed it was just the PWA pushed through apple's hoops.

[earnest frigate]

might be wrong

[median meteor]

wen macos gpt

[frigid cove]

it looks like that... but as it was not written in announcements, I was asking

[sharp pilot]

AFAIK when GPT-4o is released it will replace the old model. And that sounds a little edgy but all reports seem to be that the new model is simply superior, which is nice. But they've also kind of announced the release as if it's already happened, when free users like us are still awaiting the rollout.

tl;dr: Rollout to free is either in planning or initialization stages, unless I'm mistaken GPT-4o becomes the current default model for GPTs and all ChatGPT.

[earnest frigate]

your possibly wrong since- OpenAI said they will limit 4o for free and keep 3.5

[silk blade]

i wonder what the internal context for a multimodal agent looks like? do they still think in written to-do lists ?

[sharp pilot]

Gotcha. Sounds like 3.5 stays around as the fallback after quota.

[blissful idol]

still cannot contact any human consultant to fix this problem

why they deactivate my account and still charge me fees, that's unreasonable

[earnest frigate]

I dont think OpenAI plans to change or remove free usage of 4o
i mean the 3h cooldown is a nice choice.

Since it makes paid tier still better

[frigid cove]

I have tested it... it is faster and better (a lot faster in my opinion). but as it is described, 4.o is the new 4, but sound like alpha version as they did not replace the GPT 4 but invite you to use 4.o. However, it seems to not work on our created GPTs. And that’s sad considering that the comprehension of foreign language is not perfect for GPT 4. I was hoping an improvement on this for my created GPT redacted in my native language (that is not, of course, english). That’s the main reason I wanted to test GPT 4.o

[tiny girder]

You should hear back from them soon enough. They might be overwhelmed at the moment, but they will surely get back to you.

[feral zinc]

they have not specified if free has x/3h, x/1h, or something else

[earnest frigate]

well i get 3h after a while

[sharp pilot]

The new tokenizer for 4o and the training that went along with it for better multilingual support is really exciting stuff, even as an English native. I'm eager to see it too.

[earnest frigate]

so i guess for now they stick with 3h cooldown

[feral zinc]

it can still change as they left the details out "Users on the Free tier will be defaulted to GPT-4o with a limit on the number of messages they can send using GPT-4o, which will vary based on current usage and demand. When unavailable, Free tier users will be switched back to GPT-3.5."

[sharp pilot]

On the video announcement I do remember them saying free access would be 1/5th the paid and sounding confident that's what they were aiming for but you're correct that they didn't specify anything specific in terms of the actual rate limits or time and those seem likely to be shifting goalposts rather than "just x/5".

[earnest frigate]

yeah i guess the demand is huge

[sharp pilot]

It's an interesting problem, I think. They are still doing paid prioritization but they also can't do starvation or no-one will have the free access, so logically they have to have two different load balancing schemes for paid and free.

[dire obsidian]

Ironically, 4o doesn't even know it's 4o. It seems to 'think' it's simply 4.0

[eternal rock]

`` Server Tip: ``

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

[feral zinc]

they said it the other way, plus users having up to 5x more messages

plus users are also subject to dynamic changes

[sharp pilot]

This does feel like irony but it could be as simple as their not updating the system prompt. I don't believe any LLMs get trained on a self-identity because they intentionally want the resulting "Assistant" to be malleable, but perhaps I'm wrong on that and they do put in something in the datasetting.

[dire obsidian]

Is 4o still limited to 40 messages per 3 hours?

[earnest frigate]

isnt it less? i got like 10 i think limits vary too

[frigid cove]

my understanding is that they increased the limit (at least doubled it). The limit depends of the current use

[sharp pilot]

I would not be shocked if both the limits, time windows, ratio of paid vs. free access, etc. bounce all over the place as they figure out how much access they can provide and the challenges of load balancing.

[dire obsidian]

Sorry I mean 4.0 seemed to be 40 per 3 hours. So I'm asking, if 4o is the same?

[feral zinc]

"As of May 13th 2024, Plus users will be able to send up to 80 messages every 3 hours on GPT-4o and up to 40 messages every 3 hours on GPT-4. We may reduce the limit during peak hours to keep GPT-4 and GPT-4o accessible to the widest number of people."

[dire obsidian]

Thanks Radium and everybody

[frigid cove]

this is what I have read somewhere...

[sharp pilot]

Thanks for this, I didn't realize it had been announced. This is the same kind of basic messaging they should be doing regarding the rollout too in my opinion but at least they addressed their plans "we are aiming for 80/40 per 3h, rates may change".

[dull crow]

An openai discord manager could simply fill some FAQ and pin it and this conversarion would be much richer. Same questions have been asked since yesterday 1000 of times.

[feral zinc]

yep its all here https://help.openai.com/en/articles/7102672-how-can-i-access-gpt-4-gpt-4-turbo-and-gpt-4o#h_b04e3f8259

[sharp pilot]

100% this.

[dire obsidian]

They 'should' be more verbose about limits via the ChatGPT interface but I guess they have their reasons

[sharp pilot]

The only one I can think of is they don't want people casually using their sliding window quota like benefits.

[blissful idol]

I think now I just need to find a way to upload my case details.

[dire obsidian]

I tried to get it to count them via custom instructions but it didnt work great. Giving the messages a random string reference worked quite well though for helping me to get it to refer back to itself.

[frigid cove]

well... let’s wait that they implement it for created GPT to be able to really test GPT 4.o, but it looks promising on paper. Cannot judge as we have to use it in the standard way, so it is, actually, totally useless.
I hope it does not take too long to get it for everything

[nocturne ember]

Who do I need to slip a 20 too in order to get Mac download already lol

[slender lagoon]

any data scientist , ai engineer , or data analyst here?

[dire obsidian]

I am to some degree but that was quite broad

Plus I'm not sure soliciting is allowed here

[slender lagoon]

i want to ask about ibm proffesional courses on coursera are they worth it or not?

[dire obsidian]

Ah cool. Well it always helps to have professionals teach you industry standards

[rugged remnant]

I would like to ask if there is any specific data on the limits of GPT4o's conversation rate and number of file uploads. Netizens are now speculating. There are no specific limits stated in the official FAQ.

[latent ibex]

I think it's the same as GPT-4 for now since no extra feautures are launched

it's 'to new' to find that information yet indeed

[slender lagoon]

i want to make some ai models + ai based fintech apps . is the ibm courses enough to make this projects?

[rugged remnant]

The file limit is basically three

[latent ibex]

amount of tokens is 128k at least, so same as GPT-4 so limitwould be the same in theory

[rugged remnant]

Conversation rate limit ranges given by netizens vary widely.

[latent ibex]

I did more than 5 before if I remember correctly, but depends onsize

[rugged remnant]

So there is no official clear statement?

[latent ibex]

not yet it seems

[rugged remnant]

Oh well

Let’s wait and see how it will be improved in the future.

[dire obsidian]

That's generally the purpose of them and they are IBM which are obviously some of the most respected certificates to have. As long as they also teach you coding as well as theory. Since that seems to be your objective.

[slender lagoon]

ok bro thanks

[dire obsidian]

I think what Im trying to say is, read their course content. Because if they say it will do something for you it will. None if it is that hard really anyway. As long as you're committed.

[rugged remnant]

Do smaller files mean more can be uploaded?

[crystal tapir]

With my phone I could send only 3 pictures.

[latent ibex]

o I only tested files like .txt

[rugged remnant]

me too

[latent ibex]

and pdf

[rugged remnant]

Do smaller files mean more can be uploaded?

[latent ibex]

no idea Ill check with pictures now

[rugged remnant]

All I upload are pictures

[sage finch]

Why is the desktop app available for download but not allowed to use it. This make no sense. Why even make it available for download? You do not have access to the desktop app yet. You can still use ChatGPT at chatgpt.com

[sleek aspen]

It's a gradual rollout, this can help ensure that their servers don't get overloaded by too many users accessing the new app all at once

[latent ibex]

i did 7 pics now and it does work

[rugged remnant]

Are you a plus user?

[sage finch]

Sure, but why tell me it's available?

[latent ibex]

yes

[nocturne ember]

Where did you see this?

[latent ibex]

openai download page

[sage finch]

in my chat session at chat.openai.com

[latent ibex]

https://openai.com/chatgpt/download/ I thought people saw it on

[sage finch]

Not me. It was literally a message in my chat session

[sleek aspen]

🧐 if you were told in your chat session then it should work just fine, odd

Mind posting in #1070006915414900886 if you can?

[latent ibex]

o lol

[sage finch]

ok

[rugged remnant]

ok

[candid anvil]

why do I get the following answer first for every question today. If I click on regenerate, it works.

"Unusual activity has been detected from your device. Try again later. (883b81a54dee996e-FRA)"
These are simple questions that I ask ChatpGPT.

[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.

[twilit vault]

Im getting a network error, for all by 4o answers today

anyone else?

[sage finch]

Done, ty

[crisp cairn]

interesting, i asked a specific Azure cloud related question - gpt4o and gpt4

gpt4 got it right, gpt4o was wrong

so not yet a replacement...

[ivory hazel]

is chatgpt down?

[maiden fractal]

Status.openai.com

[candid anvil]

it's very slow for me and error messages come all the time, but it's not down

[inland heron]

I wasted so much time and energy today with GPT. It is sooo incompetent when it comes to complicated code it's just silly 😄

long way to get to AGI

It's also heavily redundant. If anyone else is experiencing this, please report to https://help.openai.com/en/

the sooner they fix this the better

[dense halo]

hey will gpt-4o come to copilot

[empty jungle]

Gpt4o keeps bouncing back to 3.5 even though i havent reached the 4o limit. Anyone else experiencing this?

[worthy lynx]

I didn't see it in the announcement pages but maybe I missed it: Anyone know when Omni's capabilities for audio will be rolled out to GPT Plus users? I think it's several weeks...

[errant thorn]

this can be bypassed

[keen meadow]

My experience hasn't been that. The code quality can be phenomenal, even for complex tasks. The critical thing is breaking bigger problems down into smaller chunks, and being extremely clear about what you want.

If you can itemise a solution clearly, I find results are excellent. If you aren't even sure what the result looks like, then not so much. There may be languages that ChatGPT excels at more than others though. My world is mostly JS and PHP.

(Referring to GPT-4 of course. Wouldn't waste my time with 3.5.)

[dense halo]

Open ai servers are melting. 🫠

[narrow viper]

My Custom Instructions include: "I like code to be very modular and organized, and use Design Patterns when possible."
I also emphasize small and focused functions modules. This helps the GPT to avoid logic that gets too long and complex.

[surreal lion]

I assume still not gpt4o for most people then?

Even the subscribers aren't getting from my experience

[narrow viper]

See "Model Availability" at the bottom of this page: openai.com/index/hello-gpt-4o/

[dense halo]

I have the ‘o’ model

OpenAI should’ve upgraded their back end before releasing gpt4 to everyone. It’s constantly telling me servers are experiencing heavy load

[toxic lake]

when is gpt client releasing on win?

[narrow viper]

Yor - within some months.

[toxic lake]

why the delay?

[narrow viper]

They're doing Mac first, rolling that out now, then Windows afterward. I suspect it's issues with complicated Windows APIs for screen and audio interactions .... it aint Linux like Mac.

[wheat orchid]

or maybe because Apple paid some dollar

enough for OAI folks to backstab MS

[feral zinc]

or microsoft does not want it out at this time due launching its own copilot set

[toxic lake]

that makes no sense

macos is even more proprietary than win

putting macos next to linux is a crime

[glass fulcrum]

yea thats more real

to push te "AI PCs" fourhter to become more popular

[novel parcel]

Why is GPT4o so dumb. I tried GPT4o and I asked to split a Discord Bot in Python in two files. Strange is that it did not give me the whole file. I did this inside Anything LLM

[narrow viper]

OK, as a Linux AND Windows guy, I apologize ... MacOS aint Linux, it's based on BSD. I should have been explicit. 🙂

[wet fable]

The GPT-4o API model hast 128k context window.
Has the ChatGPT -4o model also 128k?

[toxic lake]

still dont understand why they're trying to push their own product when they have a perfectly capable one and own almost 50% of oai that they can probably natively incorporate into the OS

[glass fulcrum]

LLMs aren't the best in coding i would say....

[novel parcel]

It works with other LLM's tho

[narrow viper]

Coding with a LLM needs to be very explicit. There are a hundred ways to code anything and unless instructions are really specific a "Language Model" can easily choose any one of the hundred ways that are not in line with what's in your head at the moment.

[feral zinc]

multiple products, not just the os

[glass fulcrum]

mayby just bad prompting diff llms sometimes need diff prompting

[toxic lake]

they should just incorporate the app features onto the web for premium users for now

like its been so long and we still dont have conversation on pc

[tranquil gulch]

broooooooooooooooo where is gpt-4o??????????

omggg

roll it out already goodness

[narrow viper]

brooooooo - which platform?

[restive wagon]

i think those conversations will be called something like "calls" and will have an increased limit or someway the chat will define a conversation as only 1 entry, because most of the responses are fairly small and cut of by the user, if really each of the messages count as one, the 80 will get eaten fast...

[tranquil gulch]

Chatgpt

[narrow viper]

ChatGPT is an app that is available in browser, in mobile over iOS and Android.

[toxic lake]

whats the incentive to keep my subscription after 4o is released publicly? i dont see how $20 is worth it any more

[narrow viper]

All of these are "platforms".

[tranquil gulch]

it is NOT released publicly

[narrow viper]

The MacOS destop implementation of ChatGPT is also rolling out now. For Windows we should see it in some months.

[tranquil gulch]

MONTHS?!? oh hell naw

[toxic lake]

it will be in a few weeks, so im asking whats the point of paying $20 monthly for an obsolete version of gpt4

[feral zinc]

yes it is and for rollout read the blog

[keen bolt]

i am unable to do calls and chats to GPT's are on and off getting this message: your must recent recent failed. please try later., anyone else having th same issue?

[tranquil gulch]

its a good thing that you dont have to pay for gpt4?

[narrow viper]

@toxic lake we had the same choices when this first came out. Plus users get new features sooner. If that's not worth the $20 to you, go ahead and cancel.

[heavy rain]

so to clarify, the current headphones icon 🎧 on the chatgpt app isn't using the tech from the demo yesterday? getting all kinds of lag and errors on Android atm, and no video/image/audio support

[feral zinc]

"We plan to launch a new Voice Mode with these new capabilities in an alpha in the coming weeks, with early access for Plus users as we roll out more broadly."

[toxic lake]

no, plus was designed as a paywall for access to more powerful models

[narrow viper]

Those features are not there yet, no d3.

[toxic lake]

why werent gpt4 released then?

[tranquil gulch]

Has anyone without plus gotten gpt4o?

[heavy rain]

gotcha, ty

[tranquil gulch]

like anyone at all?

[narrow viper]

GPT 4o, the engine, is available in the Android app and browser. Not sure about iOS app but I believe so. Voice modes are not available yet.

[mild charm]

@pearl wraith isn't chatgpt cool

[tranquil gulch]

bruh, the opposite of what i use bruh

not on IOS from what i can see on my phone

[toxic lake]

its available on ios as well, used it today

[narrow viper]

TY Yor!

[tranquil gulch]

for free?

[toxic lake]

no

[tranquil gulch]

yeah thats what im sayingg

[narrow viper]

For reference:
"We're rolling out the macOS app to Plus users starting today, and we will make it more broadly available in the coming weeks. We also plan to launch a Windows version later this year." -- openai.com/index/gpt-4o-and-more-tools-to-chatgpt-free/

[tranquil gulch]

oh

[tranquil gulch]

but you dont need to have the app to use gpt-4o right?

[mild charm]

trolled

[feral zinc]

no

[narrow viper]

The 4o model is available to everyone. The special features attached to the app are not. The 4o model is a unified model "Omni" that is one model for the different modes. This is unrelated to the ChatGPT application which must be enhanced across multiple platforms to support multi-modality. I hope that helps someone.

[tidal light]

is GPT-4o only for US users?? i dont see the option as a free user :<

[solid bronze]

Shhheeessshh I can’t use the voice on 4o at all keep getting experiencing high usage seriously

[feral zinc]

its for everyone but the rollout takes time

[narrow viper]

Yes, patience.... global rollouts for hundreds of millions of people do take time. 🙂

[tidal light]

ok so if im the only user who doesnt get it i'll be mad

[tidal light]

until everyone has it

[onyx garden]

Nope - I’m UK and can access. Although I pay for plus

[tidal light]

and i dont pay cuz im a free user

[narrow viper]

Leaked internal memorandum "... don't let Bill Gates get access to this version."
Aw dude! You were confused with the other guy!

[tidal light]

OH NOO the bil gate cant access 4o

[narrow viper]

So ... Patience ... Everyone.

[tidal light]

btw why isnt anyone complaining about chatgpt being so slow these days

[grizzled sorrel]

Probably because gpt-4o is super-fast?

[narrow viper]

Um, you should see here that lots of people are complaining about performance all the time. 🙂

Oh, yeah, but 4o is blindingly fast at times, rite.

[tidal light]

oh im not much active here but yea they made an update that made gpt-3.5 and 4 incredibly slow

#1238848533759529051

[narrow viper]

Bill - in short, the performance is always fluctuating.

[tidal light]

it was NEVER this slow

[narrow viper]

It was never Free for everyone to access premium features. 😉

Dude, the whole planet is hitting this thing now. They need to scale up.

[tidal light]

yea but there's some bug that hogs the cpu slowing down chatgpt

[narrow viper]

No, it's just normal sysops.

[tidal light]

read that bug report

[narrow viper]

I read it. Other people who have some insight are saying the exact same thing I just said. What does that tell ya?

[haughty forum]

Google keynote starting soon.

On YouTube

[drowsy tinsel]

I can see a new version of gpt4o coming very soon

[tidal light]

that there's a damn bug

[narrow viper]

ok

[drowsy tinsel]

they kinda lost their brain cells when working on gpt4o's tokenizer which resulted in some extremely hilarious hallucinations

[tidal light]

that makes chatgpt almost unusable thanks to its slowness

now im continiously reporting this to openai

but they're not doing anything

[narrow viper]

I've been developing for almost 50 years. In my experience, we're not seeing "bugs", we're seeing normal though always disturbing issues in the quality of performance. With this stuff, global, complex, insanely resource intensive, that's to be expected. That's the best I can offer.

[tidal light]

nope

read the bug report carefully

and try using chatgpt in your browser with task manager

you can see the cpu spiking

i dare you

[narrow viper]

I'd prefer to work on my servers while I watch for interesting notes here. 🙂

( Actually I understand you're saying this is a client-side issue - and that's entirely possible with large contexts. That could be a browser issue. I don't think so, but I do understand what you're saying. )

[tidal light]

nah just try using chatgpt in your browser with task manager whatever resource manager you have

[errant thorn]

Its not bug, the reason why the new model is being free for certain threshold of queries is that they use your own hardware to accelerate generation speed thats being distributed across network to other users

[tidal light]

can they just dont do this

[feral zinc]

it does not affect everyone

[tidal light]

this makes chatgpt almost unsuable

[errant thorn]

obviously joke but hey, still fun

[narrow viper]

Have patience. All of these things shake out as priority dictates.

[tough bramble]

That makes no sense. The limits aren't for fun. Unless live conversations cause less load and uses reduced power compared to text, the limit is not going to be higher; quite the opposite. And I doubt voice is gonna be less computationally expensive compared to text.. which is why I'm saying, they either are going to severely up the 80 message cap for the final release, or it's going to be absolutely useless.

[dusky cliff]

will dall-e be available for free users?

[narrow viper]

Yesterday's announcement said nothing about DALL-E, so no clue at the moment.

[restive wagon]

Voice responses in the exemples tend to be smaller and probably this are by intention

[dapper isle]

has anyone of you got the new UI?

[narrow viper]

"GPT-4o’s text and image capabilities are starting to roll out today in ChatGPT. We are making GPT-4o available in the free tier, and to Plus users with up to 5x higher message limits. We'll roll out a new version of Voice Mode with GPT-4o in alpha within ChatGPT Plus in the coming weeks." -- openai.com/index/hello-gpt-4o/

[restive wagon]

No human respond to anything with a big 2 pages text, lol

[narrow viper]

( I often do 🤓 )

[restive wagon]

That's why I think they will make the new mode called "calls", calls have shorter conversations like in the preview and "maybe" are counted differently

[near vine]

do people got the new UI look?

[narrow viper]

Fptbb : Actually I think it's up to the user/prompts to decide how the GPT responds. My custom instructions request concise responses. I am sure I will need to fine-tune my prompts to get the GPT to respond in a manner that suits my conversational style. I will also try to get it to speed up responses if we can do that, like to 1.2x normal speed. That way it doesn't matter how long the responses are because it will take less time to get through an exchange. All of this will need to shake out over time.

[crude pulsar]

You reckon they'll reinstate their image generation?

[restive wagon]

but the instructions given to the model by the devs are always on the very top of that, if the model is said to respond with less of a long text, it will

[narrow viper]

I'm not sure I understand that Fptbb. WE tell the GPT how we want it to converse with us. Well, WE can, and I do. YMMV

[restive wagon]

i think that this will not be the case in the said "call" mode

[narrow viper]

Well, if everyone wants a different conversation style, how will anyone be happy with some random default? That's the situation we have now. So the solution seems to be for users (or GPT managers) to include response style in their prompts. This has been basic "Prompts 101" since the beginning.

[restive wagon]

because when you are talking, most of the great responses are the shorter one's

but in the end this is a discussion about how they should treat the limits in "call" mode so it doesn't end so quickly

[narrow viper]

That's not a reasonable statement. "most" responses are "shorter" ones. What if a correct, thorough, unambiguous response demands more words?

[somber aspen]

is 4o strictly superior to 4? I mean, as a base model.

[narrow viper]

Well, for our purposes in this channel, WE decide how we want this application to work. So if you want shorter responses, as I usually do too, then tell the model in your opening prompts, or ask it later on to please reduce it's verbosity. It will comply.

[restive wagon]

if you are talking to someone, normally, most of the times it's a two sided conversation with small bits from all the parts, right? i just think the voice model will act more like a two sided conversation than a big text monolog

[narrow viper]

Gromp, not "strictly". It's a different underlying engine, so it's more "different" than "superior" but it is supposed to be much better.

[restive wagon]

This actually makes this mode less resource intensive compared to text mode as some responses can use hundreds or thousands of tokens

[narrow viper]

If you're having a conversation with a mentor, a teacher, or other authority. You tend to want their responses to be longer so that they can confer information to you. You want to know what they have to say. Insisting on shorter responses cheats you out of the benefits of having that person there with you. Same with GPT. That said, undue verbosity (like I blab here) is often quite annoying ... and you can tune that.

[somber aspen]

I ask because here in the description it says "GPT-4 Turbo-level performance on text, reasoning, and coding intelligence" (along with other things). That could mean that its reasoning capabilities in these context might be inferior to standard gpt-4?

[narrow viper]

I ask GPT to provide responses that are concise, but to provide full detail when the situation warrants. I don't want long responses, but I also don't want to be cheated out of great responses that have significant insight. I tune the model to suit my purposes. That's how it should be. I don't want someone else to dicatate how much data my GPT will offer me.

[restive wagon]

i read somewhere in the documentation that they are finding cases where 4a is worst than 4 to fix

[narrow viper]

Let me get you another quote Gromp, hang on.

[dense halo]

so are people starting to realize that the gpt-4o rollout is a flop?

[sage meteor]

Why is it a flop? I haven't really gotten to use it

[narrow viper]

See openai.com/index/gpt-4o-and-more-tools-to-chatgpt-free/ and tell us if that helps.

John - you said before that "your friends" said that. Are you still talkin about your friends or are you talking about public reviews?

The rollout is slow and measured. One day has passed for a global rollout of one of the most sophisticated technologies ever created. Have some patience.

[restive wagon]

if by a flop you mean that they said weeks but normally can take months, lol

[narrow viper]

Need to get some work done. AFK 🏝️

[light agate]

Wdym it's amazing for a free model

[dense halo]

i just got the email. for something that is supposed to be ground breaking and life changing and magical, in a few weeks maybe i can use the audio features. oh and the api calls will be a little bit cheaper. Yay! And still all the real power of the plus apps are still locked behind a paywall. All this and the model is nothing but a glorified version of copilot. whoopityshit.

[light agate]

No but for casual people this is huge

[dense halo]

i think we are going to find that it isn't

[somber aspen]

I mean, it's kind of a lot cheaper

[dense halo]

people aren't going to go out of their way to usde this ai when they can jusr use copilot on bing

[somber aspen]

If the performance is similar with GPT-4, it's 1/6 its price. And 1/2 turbo's price.

[light agate]

It's free and better than those 2 so

[light iris]

Co pilot is garbage compared

[restive wagon]

in fact, the increase to 80 on plus is really a considerable increase on the paid side, even if you don't have that much use for the rest

[dense halo]

the use case for this free model as is for the free users is nearly the same as copilot. people will just use copilot because it is more convenient.

[somber aspen]

Increase on what??

[frigid cove]

for now and for me... this is a non event... I don’t care of something I can use as free or just for a simple question. If I took GPT plus it it cause I wanted to create a GPT for my needs... And we cannot test GPT 4.o with them. So, actually I have no idea how the new version react and cannot compare it with the 4. So, for now... this is just a non event for me

[restive wagon]

also, the free tier will only be denied access to create GPTs and dall-e probably

[dense halo]

@frigid cove dropping straight truth

[restive wagon]

well, there's an entire new model faster and with a bigger limit, '-'

[dense halo]

this mind blowing release was the boy who cried wolf

[surreal moss]

johnray you been crying in here for like 2 days

[somber aspen]

Did the price of gpt plus increase to 80? Where does it say that?

[surreal moss]

if its a flop accept it and move on lmao, stop crying

[restive wagon]

not the price

[light agate]

its clearly better

[restive wagon]

the message limit on the 4o

[somber aspen]

oh yeah. Cool, I almost had a heart attack

[restive wagon]

on top of the 40 on 4

[surreal moss]

80$ would be insanity lmao

[light iris]

It's much better

[restive wagon]

for sure

[light agate]

Will gpt 4 o have gpt store?

[restive wagon]

yes, if they said the free users will have it

[light agate]

Big w

[frigid cove]

at least... not for now as you cannot use it with self built GPT

[somber aspen]

If yes, I hope it will be consmetics such as different voices 😂

[restive wagon]

but they can not create GPT's

only use and discover

[light agate]

That's fine tbh

Will you get dal e 4 or just bing model

[restive wagon]

probably not

i didn't find anything saying free users will have dall-e access

[sturdy cave]

does everyone have access to gpt4o and gpt4 now?

[light agate]

I feel like their making it free because we are closer to gpt 5

[frigid cove]

i have

[restive wagon]

only o

[light agate]

How

[restive wagon]

and not for now, wait "some weeks"

[sturdy cave]

why not gpt4 too if it has all the capabilities of gpt4?

[glossy fjord]

question, how many parameters does GPT 4o have?

[restive wagon]

(most of the times "some weeks" means months...)

[restive wagon]

because gpt-4 is the old version and is more expensive to maintain

[surreal moss]

i still got gpt4

[light agate]

Oh yeah what happened to sora will they try and impliment it in gpt o or normal gpt 4 so like "make a video of an elephant walking through NYC"

[restive wagon]

you are on the plus right?

[surreal moss]

ye

[restive wagon]

he is saying free users probably

[surreal moss]

oh my bad

[sturdy cave]

whats the difference between GPT-4o and GPT-4 Turbo?

[restive wagon]

IDK, but I think Sora is well far away from the common users, I hope they surprise us and say I'm wrong

[restive wagon]

https://help.openai.com/en/articles/7102672-how-can-i-access-gpt-4-gpt-4-turbo-and-gpt-4o

[dense halo]

I've been complaining for two days because there are too many people are willing to give openai a pass on this announcement. If they promise something life changing, maybe they should deliver something life changing, not give us a promise that in several weeks it will be good.

[light agate]

No one said it was life changing💀

[sturdy cave]

How do I get the desktop app?

[light iris]

Who said it was life changing lol

[somber aspen]

Was there a promising announcement? I didn't check lately

[feral zinc]

"We're rolling out the macOS app to Plus users starting today, and we will make it more broadly available in the coming weeks. We also plan to launch a Windows version later this year."

[light agate]

Just wait for gpt 5 if your so worked up about it

[dense halo]

they said it was magical: OpenAI is dropping new ChatGPT features today and Sam Altman says it 'feels like magic' magic/lifechanging/whatever.

[light agate]

It was probably magic

[restive wagon]

GPT-5™

[wet sorrel]

when can we actually expect that

[light iris]

Well it is pretty cool

[sage meteor]

Literally their job to hype things up lol

[light iris]

Very true

[dense halo]

and lets face it, as someone who has been using openai's chatgpt from the beginning, i haven't seen a lick of new magic.

[light agate]

Idk they said soon it will probably be realesed in the next 1 or two years maybe three

[vivid oasis]

Where can I find download link for the chatgpt desktop app for windows?

[sage meteor]

In the future

[light iris]

Then maybe you never will

[somber aspen]

when gpt5 is released we won't have to work 😂

[rugged shell]

What is Scallion? Was noticing my GPTs have started running on 3.5 instead of 4 but now it says Scallion while mentioning 4..?

[restive wagon]

thats called progressive evolution

[light agate]

All the teenagers hyped up for gpt 5 ngl

[dense halo]

yeah. thats called "not magical"

[restive wagon]

Scallion is the 4o

[light iris]

Yep because we will be all out of work

[somber aspen]

we will make it harvest the crops

[dense halo]

don't say magical when there isn't even the hint of new magic

[surreal moss]

bro grown still crying in a random discord server

lame ahhhh boy

[rugged shell]

Gotcha thanks

[light agate]

Fr

[vivid oasis]

is Fedora going to have also desktop app?

[rough tendon]

does gpt-4o have any limits?

[restive wagon]

nothing, i got confused on that too

[light agate]

Probably

[surreal moss]

WAHHHHH WAHHHHH WAHHHHH thats how u sound

[light iris]

Lol

[surreal moss]

people like sit in the corner of the room in real life scrolling through their weather app

[dense halo]

I want us to get real features. And that means telling it like it is.

[light agate]

It's not like open ai will see this and care about one users opinion

[surreal moss]

srry its hard to understand you over all that crying

[light iris]

Ti's very true

[somber aspen]

I mean, I guess if they were hypeing it for weeks or months and it wasn't that great I would be annoyed too. But it was in the same day?

[restive wagon]

"it is magical for someone somewhere"

lol

[dense halo]

they might. and that is the hope that we actually hold openai to account for their actions in the AI space.

[surreal moss]

being dissapointed in something is understandable but bros been crying for days

[dense halo]

not everyone can be a fanboy

[restive wagon]

someone can be a fanboy in reverse too

[surreal moss]

I hope chat gpt never makes any updates and makes their models worse and raises all the prices

[somber aspen]

I didn't even see any announcement, I just noticed that a thing named 4o was available

[surreal moss]

just so you can be more upset and cry more

[light iris]

But why, mind you I see the same in gaming forums, some people never stop crying

[surreal moss]

People are strange bro what can ya do

[dense halo]

my hope is that when they realize this rollout was a fail for them, that they will start to think why.

[light iris]

Probably ignore them is best

[surreal moss]

my hope is you'll grow up one day and stop crying in discord servers

[restive wagon]

why something that just give improvements is a fail?

[hollow harness]

User
IMG_1614.mov
File
Describe what's in this video
ChatGPT
Sure, please provide the link to the video or describe the video in more detail, and I will help you describe its content.

User
I sent it to you in the previous message
ChatGPT
I apologize for any confusion, but it looks like there isn't a video link or video description in your previous messages. Could you please provide the video link or describe the video again? This will help me give you a detailed description.

I think the file uploads are broken

[restive wagon]

maybe your expectations was just too high

[light iris]

I agree, open Ai just gave us something magical

[haughty forum]

google trying to catch up r/Bard/comments/1cr5qb6/google_showoff_conversational_gemini_videolike/

[light iris]

This new release is incredible

[surreal moss]

do tears flow down your face as you type these messages on discord?

[hollow harness]

also the attachment button isn't quite in the right place

[haughty forum]

voice is coming in a few weeks, video.... ? months? platforms? shrug

[tight island]

hot take: gpt 4 is better than 4o

[obtuse wigeon]

100%

[dense halo]

see

[haughty forum]

4o is half the api cost tho

[surreal moss]

you think so?

[tight island]

prefer "better" over cost.

[haughty forum]

but I kinda prefer claude sonnet to 4o idk yet....

[light agate]

Imagine an old grandpa seeing this or a 40 year old they would be shocked

[somber aspen]

it's 1/6 the cost of normal gpt4. Half the cost of turbo

[surreal moss]

GPT4 doesnt even load for me its so slow

[obtuse wigeon]

4 is way better for creative writing tasks than 4o.

[restive wagon]

i can guarantee that because i just tested 😂

[tight island]

i started a war

[light iris]

Intelligent response is far better on 4o

[somber aspen]

I assume Turbo is not be better than 4o

[buoyant rampart]

Heeey now, I'm 40+ and I ain't shocked 😂

[haughty forum]

claude and gemini are better at creative writing imho

[surreal moss]

lol whats good unc

[tight island]

i'd say that it's formulation is better its more "Human" if i were to use that take. which is bad. what we want is Hall 9000

[light agate]

Lol

[obtuse wigeon]

Never heard of 'em.

[surreal moss]

Hello there UNC

[haughty forum]

claude 3 sonnet and opus in particular

[light iris]

Haha

[haughty forum]

a lot of the issue might be the massive system prompt that chatgpt has... it's like three pages

[restive wagon]

tested not with me, i'm not that oldknowledgeable

[light agate]

Well with gpt 4 o coming out I don't have to beg on reddit or discord for a gpt 4 invatatoin code

[tight island]

my instructions are like 8000 characters i have no issues

[haughty forum]

Yeah and they append their prompt

I have a pretty long custom instruct myself

[tight island]

ohh i dont have custom instructions

i use customgpts to achieve my goal

well i do have it, i dont use it

[haughty forum]

right I re-used my custom instructions for... the custom instructions for a customgpt

[sturdy cave]

whats the message cap for GPT-4o plus users?

[tight island]

10,5

[buoyant rampart]

80/3 hours

[obtuse wigeon]

I use plain language and GPT 4 has been great for that. Seems like 4o is made for those of you who are using engineered prompts. A step backwards.

[tight island]

i use JSON structure, no issues

[buoyant rampart]

I've had no issues with using plain language on 4o

[haughty forum]

alright google keynote should be starting... I think they're going to announce basically what openAI did but without the amazing voice model

[buoyant rampart]

It actually seems to understand better if I tell it to revise its response, using plain language

[restive wagon]

better to use #ai-discussions

[torpid mirage]

I don't understand, I am paid user since the begining, I still don't have access to Sora, and I don't have access today to the app on my mac. Do other people have the same issue? Paying for being a plus user, but not having all the features of a plus users? Many thanks

[surreal moss]

no one has access to Sora yet and the mac app did not come out yet

(someone correct me if im wrong)

[torpid mirage]

so all the people I am seeing on twitter using Sora or the app store, its paid users living in san francisco only?

[granite arch]

Ohhhhhhh GPT-4o is live on my app

[torpid mirage]

thanks for your help Sora

[tribal nexus]

There is a search bar on iOS for chatgpt but not on the website...

[candid junco]

how does one access gpt4o, i only have the option for 3.5 and to upgrade to plus for gpt 4

[rose glade]

1. Does anyone know, with the new GPT 4o does this automatically apply to major GPT's e.g. various major GPTs?
2. Does anyone know the actual limitations on GPT Plus? E.g. Am I limited to something like 40 messages / DALL E 3 images every 4 hours or something?

[surreal moss]

I’m not sure brother sorry. I thought all the Sora videos is just from OPEN AI promoting their product but I don’t really know. In general though most people don’t have access to Sora for sure. And yeah the mac app should be coming out soon but it’s not out yet

[rose glade]

I think you need to be on gpt plus to access

[surreal moss]

Np

[undone spruce]

I heared rumors that also free users can access it.

[rose glade]

It says on there site something about specific GPT's becoming free somewhere...

[keen meadow]

GPT-4o is rolling out much more slowly to free users, and will have fairly low usage limits once it does arrive. If you want it right now, subscribe to ChatGPT Plus.

[past reef]

Can anyone help me my gpt app isnt showing voice chat option

[toxic lake]

no, these services usually use OAI's API, you should bother them to switch to it as it is possible.

[sturdy cave]

when will the desktop app be available?

[stoic sand]

Hello everyone,

I wanted to share that about an hour ago, I was testing the new ChatGPT feature that allows for audio-to-text conversion, enabling us to speak directly through the mobile app. At that time, it wasn't working and I received a message stating that the server was overwhelmed. Now, it seems that the option to speak isn't even appearing. Is anyone else experiencing this issue? Any suggestions on how to proceed?

Thanks in advance for any help or information you can provide.

[granite arch]

I have GPT-4o on my desktop (Chrome) as plus user

[inland heron]

Yes, it is quite exceptional. Though sometimes when you are testing for a longer time, and the conversation starts to drag a lot, it just keeps repeating stuff and only a new conversation can fix it. It's amazing how much just taking a break, then starting a new conversation changes things. I give it exactly the same data, and it produces completely different reasoning and conclusions and finds the issue.

Talking about unit testing here, which can be challenging if you've got a complex f unction with several modules interacting with it, and you're a beginner 😄

[inland heron]

how's it going? what's the benefit of a desktop app?

[past reef]

Facing the same issue voice chat option isn’t t showing

[tight island]

if u dont have it it isnt an issue

[torpid mirage]

do someone has an idea when gpt-4o will be available on their API

[rose glade]

ok thx

[granite arch]

Honestly, I don't use it all that much so I can't tell how better it is lol

[tight island]

u just dont have it yet, u get it when you get it

[granite arch]

The image generation by text seems tad faster and seems to give better results

[past reef]

I had it before why its gone ?

[tight island]

prolly mistake now u just gotta wait til u have it again

[past reef]

I hope so

Also btw what with the gpt 4o limits

How do i know my limits

[tight island]

dpends if ur paying or not

[inland heron]

I'm only interested if it mitigates the CPU usage of a long conversation and the unusable browser at that point

[past reef]

For free verison

[tight island]

80/3 hours i think

[past reef]

80 credit?

[tight island]

80 msgs per 3 hrs

not api

only chatgpt interface

[granite arch]

Ya...speed wise, answering questions, it seems to be about the same. I'll have to play around with it more later

[past reef]

I hope its permanent not for some tests#thats the best thing they added for free users

[tight island]

i hope its for tests and not permanent

[keen meadow]

That's because only enterprise users get the 128k context size. Everyone else needs to bear their rather limited context in mind when using ChatGPT Plus. You do pretty quickly get a sense of the optimal time to start a new conversation, and what (if anything) is worth carrying over from the previous conversation.

[inland heron]

also where the hell do you download this from??

[buoyant rampart]

That's only for Plus users, that limit

[granite arch]

Download what?

[buoyant rampart]

free users get like 15 or 16 per 3 hours

[rose glade]

What are ChatGPT plus limits, please?

[past reef]

Limit is for free users plus get unlimited?

[rose glade]

thanks

[feral zinc]

no

[rose glade]

why is it though, 80 per 3 hours, but then copilot is 100 fast dall e 3 images, then more slow

[tight island]

alternatively do some reading yourselves on the actual website where the info is

[steel jasper]

Geez 80 Messages per 3 hrs is quite generous

[buoyant rampart]

No, 15 or 16 msgs per 3 hours

[past reef]

15? Or 80?

[rose glade]

its not for the price, some of my messages are like, "Confirm" so thats -1 for that lol

[stoic sand]

app or web?sorry

[tight island]

its 420

[buoyant rampart]

15 or 16

[past reef]

Thats too low

[buoyant rampart]

80 per 3 hours is for paying users

[feral zinc]

80/3h is only for plus users

[past reef]

Ohh i see ..I thought premium users gets it for free

[rose glade]

paying user vs plus user?

[buoyant rampart]

Take it up with OpenAI

[granite arch]

Both

in USA

[buoyant rampart]

Plus users are the same as paying users

[eternal rock]

`` Server Tip: ``

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

[rose glade]

thx

[stoic sand]

in spain did you know how download? im premium

[inland heron]

It's not that, specifically. It's just that the quality of the conversation deteriorates. It has always been like this for me, and while it doesn't happen always, you can often get a sense of when GPT gets stuck in the same level of reasoning. I usually give context from time to time of all the relevant parts of code from my app, and it does well (last year it was hell with minimal context window and minimal input), but even then, it continues with redundancy.

I just started a new conversation about 20 minutes ago and it's going exceptionally well. Perhaps there's been another update, or it's just an effect of a new conversation

[feral zinc]

not exactly, teams users also pay and higher limit than plus

[tight island]

get the info from here: https://openai.com/index/gpt-4o-and-more-tools-to-chatgpt-free/

Dosent hurt to do a bit of actual reading

[granite arch]

I'm not sure...you don't see it for your phone or access from the web browser?

[past reef]

Buying gpt premium is gonna worth it…it is get any good use of it

[tight island]

the moment when one has both subscriptions xD

[rose glade]

look at copilot, I think on this you get more dall e 3 generations

[inland heron]

desktop app, what else 😄

[void lake]

I got 4o, but how can I pull up the conversational interface as in the demo? I'm on the android beta

[buoyant rampart]

Yeahhh, fair point! And there's also Enterprise, right?

[rose glade]

so if you are gonna use it for dall e 3 I would look at others

[granite arch]

I didn't even know there was a desktop app lol. That's how much I use this.

[rose glade]

IMO its not the best for dall e 3

[feral zinc]

yeah

[tight island]

except u as a user cant get Enterprise. there are requirements.

[rose glade]

I miss dall e 2 😢

[buoyant rampart]

I'm good with Plus for now 😂

[past reef]

Copilot is good but not best i used it multiple times, it sometimes failed to give me the right answer i m looking for, repeats the same answer again and again

[inland heron]

oh, my bad, I thought you downloaded a desktop app, but in your original comment you were referring to a desktop Chrome web app, right? 😄

[buoyant rampart]

I get my money's worth enough from it as it is

[rose glade]

yeah if your using more text than images I say go with chatgpt plus

[rancid venture]

Yo guys

I need help pls

[tight island]

[rose glade]

just for dall e 3 I would rather use night time cafe platform, official open ai API or microsoft co pilot, its better value for money

[rancid venture]

Its not letting me talk to the AI

[inland heron]

there's a desktop app coming, so I thought people started using it, but it doesn't seem to be available for a lot of people

[granite arch]

haha ya. I had to read up but looks like only for MAC for now?

[tight island]

thats not enough info, cant help you, i need more information, specific msgs and etc.

[rose glade]

desktop apps for this sort of thing usually wrappers that just use more RAM

[buoyant rampart]

Yep, Mac only

[granite arch]

OpenAI trying to get away farrrrr from browsers, Microsoft/Google/Firefox

Which is not a bad thing as Google blows lately with their search results

[inland heron]

no idea honestly

[rancid venture]

Look i was talking with him abt how my calculator was giving me fractions he gave a soulution but it made rand goofy ah sound then all of the sudden i cant i just see an arrow facing up (i use tha android app btw)

[past reef]

Copilot is best for image creation rather then text

[tight island]

try starting over.

[inland heron]

I just want my browser functional once long conversations start breaking it down and making it unusable

[rancid venture]

I did

[tight island]

did u try again did u log in and out did u clear cache did you turn off vpn

[rancid venture]

I logged in and out

I dont have a vpn on

I did clear the cache automaticly

[tight island]

tried closing your browser and opening it again

[rancid venture]

Here i will send a screeni

[tight island]

u cant

send it here #off-topic

[rancid venture]

I sent it

[nocturne ember]

You amazing people at ChatGPT.. It would sure be swell to get access to the mac download already. It would make things super!

[frigid cove]

lol... using VPN I see that there is one more option in the USA than in my conuntry... and the only thing that can be useful between everything they just released. The option for ChatGPT to learn from your discussions....

[tight island]

just make sure u aint using no free vpn, for the safety of your account.

[frigid cove]

frere?

[granite arch]

fereerere

[frigid cove]

expressvpn here... and I already reswitched to France...

[narrow anvil]

Hey hows it going?

[tight island]

just wanted to remind users that using free vpn with backlogging, and improper protocols, lead to permanent account suspension if other ppl on the same vpn address get flat out terminated and banned.

[narrow anvil]

This is my first time in this server

[granite arch]

It's not IP based?

[tight island]

never said that Sparky

[granite arch]

Oh okay

I use paid VPN but...

Sometimes certain sites will block me; for that same reason I'm assuming

[tight island]

i said if you Switch to x address in spain and ur using a free trash vpn, Lets say 1 user out of 20000 users using the same vpn on chatgpt trigger a security breach.

u know who then gets banned?, all 20000 of them

[granite arch]

It wouldn't matter if you paid or not? Or are we talking free plan?

Paid for ChatGPT, that is

[tight island]

Its more Backlogging issue.

[granite arch]

I see

[tight island]

https://community.openai.com/t/getting-banned-for-using-a-vpn/504741

just the fact he was on the same address "Vpn" proxy is enough reason for termination.

[frigid cove]

anyways, I would not use a freee vpn for the reason it is worst than not uising VPN at all

[tight island]

same using a free vpn is sus.

[granite arch]

Don't raw dog the internet. Use a legit VPN.

[surreal moss]

Whats a good paid VPN?

[tight island]

agreed, if u gonna use a VPN use a proper VPN

[granite arch]

Many out there...just search for the ones that doesn't 'truly' keep logs

[frigid cove]

I would advice ExpressVPN. expensive, but quality VPN

[tight island]

ExpressVPN and NordVPN, umm what else.

[granite arch]

I don't want to say mine but it's a well known one; just so I don't sound biased.

[surreal moss]

bet thx

[surreal moss]

its not that deep bro but thx

[granite arch]

It's not

[frigid cove]

but a lot are good... NordVPN by example. As said by somneone you hjave plenty of comparison on the internet

[tight island]

CyberGhost Potentially and Surfshark, Maybe Windscribe. or ProtonVPN

[frigid cove]

Also, don’t forget the free trial. 30 days free on most VPN

[tight island]

etc etc etc

[granite arch]

Even though you ask some Ph.D barista will tell you a VPN can't protect you, yada yada

[surreal moss]

apreciate the input guys

[granite arch]

[tight island]

Welcome

[granite arch]

lol

[surreal moss]

preciate is my slangggggggggggg

[granite arch]

susssss

[surreal moss]

preciate datttt

[tight island]

ciate it

[granite arch]

ciate

[tight island]

it

[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.

[granite arch]

Oh crap

[tight island]

oh.

[surreal moss]

i live in nyc we barely speak english its just all slang

[uneven harness]

anyone know how to fix the headphone button to chat live with chatgpt not appearing on iOS?

[granite arch]

lol it just warned me

ciate sounds like with an s-

[frigid cove]

surfshark I tried with the 30 days free and I was right to do it... I had to ask a refund, the VPN was preventing me to play with a good ping

[granite arch]

This sub moves so fast now...wasn't like this a few months ago

[surreal moss]

wouldnt all vpns have a negative effect on ping

[tight island]

yes

[granite arch]

To a degree

[tight island]

it has an impact because it funnels ur data.

Alternatively u can make your own VPN its not that hard.

[frigid cove]

they absolutely have. But by example I have a ping of 65 on HD2 with expresvpn... why I pay for it

[granite arch]

1/4 cup of ketchup and mayo + dash of internet = VPN

[tight island]

wrong u need a potato and a power adapter as wlel.

[surreal moss]

why do u guys use VPNs?

[granite arch]

[surreal moss]

like if big tech or someone wants me data they can have it tbh

[tight island]

i dont, i have no use for one. i have a specialized provider.

[surreal moss]

like im super interesting so i understand why they would want to get to know me

[granite arch]

That's like saying the cops are welcome to come into your house at any time

[frigid cove]

watching streaming of other countries, accessing website you can’t in your country, anonymously

[surreal moss]

privacy i dont care for that much vs privacy I do care for

[granite arch]

Data can always be used against you...other than getting data stolen, etc.

[tight island]

even if u had a vpn through a vpn, if Openai really wanted to know ur IP u cannot hide it.

[surreal moss]

true true

[surreal moss]

I see

[granite arch]

I'm not hardcore about it but privacy protection is scary in the USA...I'll leave it at that

[tight island]

everything is interconnected, Your username, Your passwords, Email, etc. simple backtracking can do the job.

[surreal moss]

USA is insane

[frigid cove]

I live in France. Something is avalaible on streaming 2 years after the release on cinema

[granite arch]

You basically need to live like Reacher to be somewhat anonymous

[tight island]

yeah

[surreal moss]

America rlly dirty af man

[frigid cove]

and immediately with VPN for several streaming plateformes

[granite arch]

Don't even look at the internet...then again, cameras are all over these days...

[surreal moss]

BONJOUR

[tight island]

ur phone is tracked, whether its on or off. has been confirmed.

[surreal moss]

crossaints

Edward Snowden vibes

[granite arch]

Yes...

Snowden wasn't wrong...

[surreal moss]

Me personally I wouldnt let them track me, thats jus me tho

[tight island]

onions are Microphones, Cats are Drone Controllers, and Doves are the drones.

[frigid cove]

don’t worry... we are ALL tracked and Facebook knows you more than you think

(Bonjour Sora)

[tardy totem]

Is it just me or is O less good at code generation?

[granite arch]

I knew the highway pigeons looked sus

[surreal moss]

if the gov was tracking me I would just tell them to stop

cuz that is not cool of them

[tight island]

i dont think they would care, but at least you're asking 😂

[surreal moss]

if they kept doing it I would have to intervene

[granite arch]

It's scary what NSA/CIA can do alone when it comes to data

[tight island]

this is why i dont even bother hiding who i am on the internet.

[granite arch]

Give me your SSN/DOB right meow

[tight island]

so what if some internet troll knows who i am, what he gon do.

[granite arch]

Steal your $2 in your wallet

[tight island]

no i need those 2 bucks for candy.

[granite arch]

I have $3

Can't afford candies

It's like $5 a pop here

[tight island]

i once did have a troll who gave me my ip and everything, i just told him, "umm ok?, Come at me man, idc. bye blocked"

[tardy totem]

lol

[granite arch]

Should've sent the SWAT in

[tight island]

what he gon do? he cant take me in a fight.

[granite arch]

e-fight

[tight island]

fite me 1 v 1 Roblox

[granite arch]

looool

[tight island]

😂

i have actually never tried roblox

[quaint timber]

yo did the "talk to chatgpt" feature get removed? I was using it to practice job interviews yesterday and today it seems like the button is randomly missing in the app

[granite arch]

I don't know how it's so popular...the characters are butt ugly

[tight island]

its intentional.

[tight island]

https://status.openai.com

[granite arch]

I want ChatGPT to talk to me randomly instead. And speaking of which, they need better voice models.

They're all trying to sound neutral and if not boring.

[tight island]

i dont want to talk to gpt if it wont talk back to me, and scold me or act smug

[granite arch]

Can we just not have a hot Irish redhead in her mid 20s or deep tone Hulk-like guy yelling at you?

[quaint timber]

ah, i see. Thanks for the link

[granite arch]

Like Mr. T!

[tight island]

gpt can u tell me a lullaby, "Awww Hell no mate grow up "

thats the stuff i want

[granite arch]

It needs to boss me around

[tight island]

i want it to let me know my place

[granite arch]

Yes

The voice in my head becomes reality

[tight island]

My very own Motivational Speaker.

[granite arch]

TONY ROBBINS?! OMG....as much as he's kind of crazy

[tight island]

if its cheap enough put it into an app

have it wake you up and order u around.

[granite arch]

or heavy breathing Darth Vader

Son....TAKE YO MEDS BREATHEEEEE

[tight island]

"WAKE UP its time for the 9-5 work, WELCOME to the MATRIX" Laughter

thats what i want it to do

[granite arch]

[humble pagoda]

not true with canal+ mate

[tight island]

when its time to eat, i want it to not tell me to eat, But ORDER me to eat.

[humble pagoda]

i lost voice function on my account tho, what's going on

[frigid cove]

Well true with Disney+....

[tight island]

https://status.openai.com

[grave token]

They just pushed an update to the iOS app to remove voice mode…

[tight island]

https://status.openai.com

[grave token]

I was hoping it’s because they were about to push the new voice mode! Too bad

[frigid cove]

so, the personalization/memory setting is exclusive to USA or there are already countries that have it?

[humble pagoda]

canal+ has exclusivité for recent movies and you can almost see them all on the box office category, though they don't have everything ofc

okay autocorrect

[humble pagoda]

everyone except europe and korea

[young basin]

I just read somewhere that you can upload audio to 4o but that's bs right

[frigid cove]

Thank you... what europeans did to the chatgpt team to be punished? lol

[humble pagoda]

european laws regarding data are strict

[native turtle]

Guys, did anyone else notice that fact that you can no longer start a conversation with ChatGPT like a speech to speech conversation in the app? It got replaced by a send button for me

[humble pagoda]

it's down rn

[native turtle]

Was it ever down before?

[humble pagoda]

idk but people wanted to try out a feature not yet released

[midnight plaza]

anyone knows where to download the mac app

[native turtle]

you mean the gpt4o talking feature?

[humble pagoda]

if you can't from the openai website don't do it, you won't have access

[humble pagoda]

yeah, the new voice upgrade

[native turtle]

Do you know where you can find it on the openai website? Im pretty sure nnoone has access yet, right?

[midnight plaza]

thanks, this explains a lot.

[humble pagoda]

it's a big button so you can't miss it, but some do have access i heard

[native turtle]

Where is the button? Like on the chatgpt website or really the openai website?

[humble pagoda]

chatgpt

[native turtle]

Mh, certainly dont have it then. I guess it just pops up as every other feature does thats new.

[frigid cove]

Thanks... yes I see.. RGPD for the win

[humble pagoda]

it's only on mac btw

[humble pagoda]

yeah, annoying...

[sturdy cave]

Is it on the home page? I don’t see where I would find it

[dim vigil]

sorry to repeat it here i have a problem I deleted an account it is impossible to recreate it or log in

[humble pagoda]

yes, but it's rolling ouy

[wanton lichen]

Did they turn off voice?? I don’t have the button anywhere on my app

[wicked anvil]

I lost access to voice today when my app updated... What's going on?

[granite arch]

Still there for voice

[wanton lichen]

Yeah I lost access too

[granite arch]

Though I'm on Android and hasn't updated since beginning of this month

[daring wharf]

Yup also lost voice in new update sadly

For iOS

[granite arch]

They must be revamping it?

[daring wharf]

Yeah possibly

[wanton lichen]

I think a lot of people where using it and overwhelming it because everyone thought the new voice stuff was out immediately

[tight island]

they're revamping it. because it had some issues.

[tough bramble]

The voice servers are overwhelmed today.

[daring wharf]

Yeah had a feeling everyone tried to use it after watching the stream yesterday hehe

[wanton lichen]

It just ends up being a lot of disappointed people because they thought they had Her lmao

[analog wind]

did anyone get access to the new voice chat mode with ChatGPT, is there a way too or do i just have to wait for it to roll out

[daring wharf]

yeah I’ve been using audio chat everyday for stuff and I feel lost without it now

[tough bramble]

Honestly, how can they announce real time conversation, when the non-realtime stuff they have now is down or slow all the time?

[dim vigil]

Pls help me

[analog wind]

no i don’t think anyone has it yet

[wanton lichen]

No it is not. That’s the problem. Everyone thinks it is but it isn’t for a few weeks

[wanton lichen]

It’s slow now because of the convoluted API calls it has to do. The new voice stuff is direct to the model.

[analog wind]

well they said they’re rolling it out so some people might have it now