#chatgpt-discussions
1 messages · Page 6 of 1
Hallucinations, ergh
It could help to share your chat, if there's nothing private in it, so we can see what isn't quite going right and try to suggest how to work around it.
I tried a sample conversation. In it, the citations are blue numbers after it discusses stuff, you see by hovering over. I'll drop a screenshot where images can be shown.
I'm hovering over that blue number 1 to make that link popup.
In case this is a useful conversation for you to continue (and yay, I notice that it references 6 different citations, not just 1), here's a share link: https://chat.openai.com/share/4b24fdbe-2a3e-4e80-a096-2bc736caceeb
bruh why does bing chat have access to my location
Thx for the pointers!
Perhaps you have location turned on, or something else indicated it?
Not sure, bing chat's not really an OpenAI product, and I'm pretty sure the ChatGPT part of it is not what's getting your location.
dang
Lot's of people are having this issue. There are some community discussions and bugs under the #1070006915414900886 , but there doesn't seem to be any traction anywhere regarding fixing this. Definitely very disruptive issue that leads to a frustrating user experience. Hope this issue somehow gets some traction 😦
Thanks! I found one discussion, link below, I wasn't sure if I was the only one getting this. Now I believe this is not a bug but a feature of some "smart" PM wanting to inspire us to turn on chat history. Instead of waiting for a fix that prob won't come, I'm moving on to using the openAI API via tools that will allow me to keep privacy ( example ; open access GPT) there should be plenty similar out there but chatGPT wouldn't help me search for it 🥲 // Here's one discussion on the issue : https://community.openai.com/t/chatgpt-bug-oct-4-chat-history-off-current-conversation-session-getting-wiped-after-short-inactivity/412462/62
How do you get gpt to speak to you and for you to be able to speak to it? I am a premium member and can't see how to do it
When it's rolled out to you there will be a little headphones icon at the top of the iOS or Android app that you press to go into Voice mode
tysm
And to turn it on, once it reaches you (before you see the headset), need to check settings then beta features, and turn on voice.
„Rolled out“ means somewhere between now and unknown? At least for europe.
Oh that is awesome! I am looking in my iOS settings right now. Where could I find the beta section. Sorry for the stupid question
oh found it!!!
Not ios settings, inside the ChatGPT app... great!
Been a long user of chatgpt. How come I haven’t get the latest beta features yet? 🥺 saw someone who just made an account and paid subscription got visions and images.
same
It's a random rollout coming to us all, but some features need to find the setting to turn on. Check here for instructions for each: #openai-chatter message
Thank you so much, I truly appreciate you!
Did anyone get the new features today? If not it sounds like our theory that no roll outs while they're on holiday was correct
it was for a few mins
fair enough, glad im not just goin loco
why is plus so expensive
Cuz it cost millions of dollars to develop and run
but if they make it cheaper more people will buy?
Vision/Voice should be fully rolled out by now but I don't have access yet. Anyone know why/when
same
Been checking every day for most of the 2 weeks and on the last day still not here 😐
Kinda getting tired of this how im spending $20/month and not getting everything other people are paying same price for
they should offer a partial refund or something
Well, a lot of The open AI people are not working today because of the US holiday. But there's a rumor that someone got something new today so who knows maybe.. I mean the days not over yet!
yeah im going to check again later tn but just annoying how they release a timeline and dont follow it apparently
also got Dall-E without signing up for it before Vision/Voice... :|
When will we get access to image for chat gpt 4 ?
Not sure anyone knows at this point.
did anybody got already ?
a lot did
only dalle 3
is it based on country or what
you're talking about vision right? or DallE?
get out the pitchforks and torches!! Kidding… mostly… lol… 😂 But yea - I feel ya - I understand sometimes things don’t go as planned so if they take longer than the 2 weeks they said in the blog, I get it…
But as someone who has (gladly) paid for plus as soon as it was offered, it would give me some warm fuzzies to feel like loyal early adopters were prioritized a bit over brand new users. I get it they probably want a good mix for the early roll out feedback, but to be at the 2 week date mentioned in the blog, and me, my wife, (both subscribers from the start) and it appears many other early adopters do not have it yet, when some brand new users have it, does sting a little bit. (I know firstworldproblems, lol)
like I didnt got anythign
vision and dalle like I didnt got any of them
Alr cause im pretty sure DallE shipped to US/UK first
country definitely has something to due with it
if we use vpn
im in us and still dont got it
could I get earlier access ?
I've had a VPN disable certain features (iirc) so maybe it can enable them
there was a form to sign
I got Dall-E without doing any form.
for dalle
mhh
Me too! I guess we're special.
👍
I believe it takes the location from google play/appstore.
Not sure which 'it' you're referring to, you know that 4 features are rolling out, and across 2 sets of device types (app and web interface), so it's like 6 things rolling out to us all.
You and your wife don't have any of the rollout? Would you like to see the list of how to check to see if each reached you, and how to turn each on if you have it?
Me too. No vision or voice yet either.
Or not, I tried VPN, did not work. @uneven aurora
For anyone curious where they said they'd have it fully rolled out by today, it says it here https://openai.com/blog/chatgpt-can-now-see-hear-and-speak
"We’re rolling out voice and images in ChatGPT to Plus and Enterprise users over the next two weeks" on Sept 25th
oh well, just keep waiting ig
They also told the vision will be delayed for EU and UK
I believe at least a month.
To be fair(er), it does say rolling out across the next two weeks, but doesn't actually state it'll for sure be done by then.
But hey, they're doing the best they can, there's millions of us, they have to scale the systems, and if it takes a bit longer - then it takes a bit longer. I hope we all have everything soon, and we do know it's coming 🙂
dam
so not Vision for another two weeks, I believe
meh
i reeeally hope i get it tonight!
What do you want to do with it? @ionic geyser
One guy suggests he got vision today. So maybe they're rolling out now! 🙂 #ai-discussions message
Apologies, I should be more clear 🙂 Voice and vision (the two main things mentioned in the blog 2 weeks ago that said rolling out over 2 weeks) - as far as where to check is it different from where I enabled browse with bing? 🙂
Voice will be in the app under settings and beta features or new features depending on if your Android or iOS. The vision thing will show up as an icon at the bottom of your screen when you're in GPT4 and that should work on the app and on the web.
Potentially yes. Mind looking over this, and checking for whatever you wish to check for? What to check for is different between app and web, and how to check is different for some of the new features. #openai-chatter message
Some folks have image on app or web, but not the other.
Be sure to check the 'default GPT-4' for image upload icon, the other configurations of 4 don't have it.
Not even sure just play around and learn to use it irl
Voice is pretty cool. Thought it would be just like any other voice assistant but the voice is scarily good. The breaths and everything. Can't wait to try out vision.
Can't wait to try it! I already have vision and that's pretty good, especially for someone who can't see very well.
FWIW I think just saying it’s a blog post, not a contract is a better case to make than “they didn’t say for sure it would be done by then” From my POV the way the blog is worded, clearly implies plus users can expect to have the feature by today.
But sure - it’s a blog post, not a contract and stuff happens, particularly on the bleeding edge… so I get that. The thing for me, that stings a bit, is hearing about brand new users getting it. If scale is the problem, unless they severely misjudged the scaling (which I wouldn’t blame them for) it does seem like by today, the day mentioned in the blog, it could have at least been rolled out to everyone who subscribed the first day or week plus was available. It could well be they are prioritizing new users who are signing up specifically for these features… from a business standpoint, I can see how that makes sense too… I mean - I’ve been paying for months, am I really going to cancel because I have to wait a few extra days/weeks… obviously not… so maybe being a long time paying customer is actually a down side, because they figure they already have you hooked… lol
All that said - still - while it’s disappointing, I do try to keep it in perspective… like seriously… how bad is my life because I don’t have these cool new features yet… lol… If I really REALLY want to talk to an AI I can always use PI… (Even if it’s I find it’s fawning annoying… lol) 😂
For vision, when you get it, is it a setting to enable in beta features ? Or it just is there ?
Just there
Fr ?
Amen brother! Now excuse me, I need to refresh my app...
Vision icon will just be there. But if you're in the EU or UK you might not get it as soon as people in the US from what I hear. But don't quote me. You might want to check the announcement 🙂
Hey, I'm just a person like you. Aware of the #server-rules and trying to help us all find more enjoyable ways to 'Keep the negativity to a minimum. (part of Rule 1) as we all see what we have and what we can do next.
If AGI helps us live much longer lives, fairness might have to shift towards 'random for all' order instead of 'seniority first' - Are we who were here first actually 'better' in some way? Helping everyone, as reasonably as possible, does seem to be the company's goal in general.
I don't see any preference for the new customers either. It looks, the little I hear, like it's a full mix of everyone getting everything in their supported areas, with a random rollout to all members.
I can't write OAI's blogs for them, or rewrite 'em. But helping our community be as upbeat as reasonable, and remain a source of motivation and enthusiasm for the OAI employees who look here - that's a worthy goal. I want them (who cares what I want, though?) if they glance here, to think "Wow, how can I help these great people even more" instead of feel "Wow, I worked so hard and look at how angry they are, what's the point of my effort anyway?" Or any of the other possible downsides.
So, let's wait with bated breath! Offering to race you to be the last to get the new features (I don't control that either, but it could happen!) - I don't have everything yet either.
Just there, but only in 'default GPT-4' new conversation, from what I've heard. Nothing to turn on, but you won't see it in Advanced Data analysis, plug ins, or any other version of GPT-4.
Good to know, thanks. Also, I'll travel for a while soon, outside Europe. If for some reason it enables the features, it shouldn't turn them back off when I get back home right ?
We haven't been told anything about that. I haven't traveled, so I can't even guess based on what I've seen. Good luck!
Let's hope it unlocks it lol. I've heard some people did it with a VPN, so I suppose 3 weeks of real IP addresses should do it
Bryce Huhtala — Today at 4:42 PM
When we're talking about the new 'voice' feature, is it speech-to-text or text-to-speech? My chatgpt app has the little microphone icon to transcribe my speech, but I don't see anything about text-to-speech
Voice is fully audio, there's no typing in that mode. After, you get a transcript of both you and the AI's speech.
If you want to read what the AI says, as well as hear it, you'd need to have another device and refresh the conversation to see the output appear
Huh, I see. When I use it on my app, it just transcribes my speech into the text box and I can submit that. I'm wondering if I'm confusing my phone's general speech recognition for a new feature 😁
think so!
That makes more sense, thanks!
Okay there's a couple things going on here I think. The chat GPT app has functionality built-in now that will allow you to speak your prompt in and then you submit it as you said. That's what we all have now. Now the new voice functionality is one where you can have a conversation with Chad GPT and it will speak back to you. Apologies if you already knew this.
Is anyone else having trouble with it (CHATGPT 4) not being able to provide drafts based on prompts?
"I apologize for the delay. I'm currently processing the information you've provided to draft a concise and effective section for your resume that reflects your roles and accomplishments with XXXXXXXXX. I will provide this section to you shortly."
Ahh ok, cool. Thanks for the clarification 🙌
Sounds like a hallucination. Have you tried to regen the response?
Yes. It's constently like this.
"I apologize for the delay. I'm experiencing a technical issue which has hindered my ability to provide the revised draft promptly. I'll ensure to rectify this and deliver the draft to you as soon as possible. Thank you for your understanding and patience."
I hear ya - and I respect what you are doing here - totally agree - let’s not be overly negative! That’s why while I express disappointment/frustration I try to make sure to include a disclaimer that I realize it’s first world problems, and I that I try to keep things in perspective. That said - and I don’t want to seem argumentative, but a few things I don’t totally agree with you on… So yea - if we get AGI/longer lives, seniority as a general rule of priority will need to be rethought. BUT - in the short term, to answer your question “are we who were here first actually “better” in some way”? — My answer is yes… I wouldn’t say I am “better”, but I would say that as someone who was here from the very start and have paid every month since, I have indeed contributed more to the company and the mission, more than someone who just joined and paid their first $20 today… so I feel it is understandable for me to be disappointed seeing new accounts getting it, and I while I don’t want to be toxic, or make a mountain out of a molehill, I do think it’s good for openai to hear my disappointment. This could help them plan for future roll outs.
I am a software developer, so while I agree toxicity is not helpful, I would posit that respectful negative feedback is actually some of the best feedback you can get, and I would think/hope openai devs also see it that way 🙂 Whether that feedback turns into them prioritizing people who have been loyal early adopters (to a point… I’m not talking about a forever seniority system)… or perhaps wording their next blog a bit better to set realistic expectations, etc. I always try to underpromise and overdeliver. All that said - I probably agree more with you than with people who you are trying to calm down… so mad respect for your efforts 🙂 p.s. I decided to sign up for a brand new account, and upon login, I have voice! Lol... now I just need to get vision on one of them. Do you happen to know where to request DALL-E? If I can get that on my new account, and get vision there first, I might just cancel my older account 🙂
Drat, and that's not a sort of conversation that should be shared, surely has some degree of PII in it.
So... let's pop over to #prompt-engineering if you want, and share the actual basic instructions, not resume, if you want help?
Failing that, I suspect there's something in your instructions that is guiding the AI to give that response instead of what you actually want, maybe you can try to read over and spot it.
What is the maximum number of nodes you can get in a chatGPT conversation tree before it says you can't continue?
I hear you, and wish you the best!
Here's the link for that dall-E access: What I don't know is how long the form's open for, when people get added. But this Yellow Name did offer a way to request earlier access to help with testing: #openai-chatter message
I have never explored that. If you find out, maybe let us know! I'm guessing you mean by regenning replies and/or editing your own.
Finally got access to Dalle-3 and it honestly blew my mind by having very consistent character face. I asked D3 to remember her face, and the next three prompts all produced identical face.
How do you get access to this? I saw someone posting about it on threads, but I don't see the beta yet 😦
This is impossible to achieve on MJ and if you use SD you have to create a character concept checkpoint to achieve this level of consistency (usually must feed it manually around 25 images of that same character)
Check the pin and fill out the form
Checking the pins in all channels Should be your First Priority and thought.
ty 🤙
I believe D3 and Vision would soon merge together, maybe in 1-2 months?
Then I would never need to pay for any other AI gen services again, because aesthestic can't beat accuracy, and D3 is the king of accuracy.
To be fair, current Dall-E/ChatGPT is seed locked, so the same prompt or very similar is likely to get the same image every time. When that changes (and they say they plan to change it) consistency may not be as easy.
For sure, Dall-E-3 is 'stateless', it doesn't know what you just asked. So asking for 'Same landscape, but at night' absolutely will generate an image - but it is based on those words and without any understanding of what the previous information was.
It could match, or not, that's super random/lucky.
Yeah without vision incorporated everything is pretty much rng
And be sure to check your settings, beta features, and turn it on.
Then in a new chat, hover over GPT-4, and pick the Dall-E dropdown.
But by asking it to remember a seed generation is already mindblowing to an extent
I even give that specific face a name and whenever I mention that name D3 generate just the same face
OpenAI cooking true black magic
I have all the beta features on available to me, but Dall-E doesn't appear in the dropdown
slide vision :deadge:
?
In the context of warfare AI is going to change everything when it comes to intelligence
Next Tuesday? 😭 😔 Very well then 😌
You can give it footage and it will give you absolute every minuet details
misunderstood mb, also if u dont have it in the dropdown, it just mean you dont have access to it yet and its still rolling out
Are you german?
Maybe even being able to remember when this personnel last appeared and so on
That's what I figured, I filled out the form for now 👍 ty
i think he means this tuesday which is tmrw
No I live in the US. Sorry for any confusion with my pseudonym
lol, I was just asking because they apparently don't roll out vision in EU and UK yet
but I already have it too.
Oh 😆
It works pretty well
ahh hopefully i get vision by tmrw night, really need it for a project
Hmm. Then I'll hope it was a typo
You're in eu ?
should be this tuesday
Lol
😲
did using a vpn actually unlock features for you ? :susge:
i moved to the US until it rolls out
In 10 years all that sci-fi movies projection might very well come true
Nah

moved out yeah 👀
We're in 2023 and still don't have flying cars
We do have flying cars
thats too unrealistic
Just slap wheel on a helicopter
Lmao true
I'll believe sci-fi is here when effective, biocompatible prosthetic limbs exist
Imagine a whole new genre of gaming based solely on accurate AI generation
True sandbox experience
that could be possible tbh somehow i think
For that to happen I guess there could be a core AI supervisor to dictate other AI agents
Kinda like an AI Overlord to oversee the world and interaction with players
You never talk to the AI Overlord, only to the agents
But the overlord instructs that agents to never disclose it's an artificial intelligence created by OpenAI and can't provide harmful information or being aware it's an AI
Westworld seems more plausible as time marches on
Sorry to bother you, but which Tuesday were you talking about? October 10, or October 17?
I think all of us are here hoping that means tomorrow lol. I know I've been a long time plus member still waiting on this feature.
10th
I want to know not hope 😭
Wym it is the 10th
The "next Tuesday" is throwing me off
Here's what I got from the support line:
"We apologize for any confusion the recent update may have caused. We’d like to clarify that the new features are currently in beta roll-out. During this initial phase, access is limited and only some ChatGPT Plus and Enterprise subscribers may be able to utilize these features. Additionally, users may encounter issues where features appear to be activated but become unavailable, as we are in the process of fine-tuning and making adjustments based on real-time testing and feedback.
While we do not have an exact date for when these features will be available to all subscribers, rest assured that we will keep the community updated. Gradually, we will be expanding access to more subscribers within both the ChatGPT Plus and Enterprise plans."
1 day lol
omgmgmgmg
The release notes were posted on the 27th and they said it's expected to release across a 2-week time period, which makes Tuesday, tomorrow, the final day.
You mean September 25th?
Nah pretty sure it was the 27th
Was the 25th
But I agree: that would mean tomorrow is the last day
By next year I guess there would be an emerging group of AI artist who posted weekly comic on twitter and locked behind paywall like Patreon
Vision is dope tho, used it a few times
Because by now D3 is already being able to remember faces
When incorperate with Vision people can essentially just draw stick man on paper and ask D3V to look at it and replace that stick man with the exact characters concept they asked it to remember
Essentially decentralizing publishing industry
The other concern is all that computing cost, which I guess everyone in silicon valley is trying to solve the AI processing chip problem
META tried but failed tho
slide vision
despair
hopefully they make an announcement that clarify this.
I'm really wanting to be able to use these features because it said 2 weeks.
i just want vision
i wish the vision team would actually do like moxi and give us a form
10, but we all know they may need to reschedule 🙂 But they haven't yet; all we have is the official announced information, and our patience and eagerness.
I mean you can get it pretty easily lol
wdym easily, they choose for us, we cant do anything
Thank you T_T
The age of storyteller is coming
They will no longer be bound with compromising pressure from publisher
I guess we aren't getting vision today
Maybe we'd finally be able to escape that cookie cutters storytelling that plagues even Hollywood for decades
I'm working on developing a lecture to explain the pros/cons of GPT as a tool for developing, analyzing, and disseminating the constructs of Health Policy... anyone got any good resources?
Well then I guess you have to understand how Health Policy is decided in the first place?
Chat GPT
Every policy is pretty much a causality process
Because of X then Y
Find out X and see if GPT can have any implication on Y
For example, I remember in an OpenAI interview they gave a very interesting POV
60% of nurse workload in hospital is essentially documenting patient
With AI they can reduce that time to 20% and can spend more time for patient care
That's one advantage
Lovely! Thank you
Thank you for adding me. I have questions if someone can help pls. 1. How to turn off custom instructions and keep the same chat? I don't want to start over. 2. Dalle-E 3 keeps making mistakes with the text, what is the easiest way to edit it? I am not good with photoshop yet.

Something just came to my mind
For example if I want to generate an image of a polar bear holding up a volley ball, but the image results are never close
Maybe I can ask it to generate a polar bear first, then choose the most accurate polar bear
On the second prompt I ask it to generate a volley ball, and choose the most accurate volley ball
Then ask D3 to merge those two most accurate pictures of polar bear and volley ball together
It has the same energy of that Patrick driver license meme
Is this a polar bear? Yes
Is this a volley ball? Yes
Then, a polar bear with a volley ball on its hand
Interesting, I must try out now
hi! can gpt issue trigger warnings for videos already? or is it not possible currently?
just asking as i put something in #1155775326253756456 and curious if it already exists
V
So OpenAI lied
Straight to jail!
openai lied
Well it's 4:13 p.m. in San Francisco right now so there's still a bit more time but they're probably on holiday yet for Columbus Day anyway. I'm sure we'll get it next day or two!
how do we properly change aspect ratio in dall e?
you only need to ask a question once in one channel
I feel you 😂 looks like we are in the same boat here. I don’t understand early and loyal users don’t get the priorities for beta features.. I mean it does what it says on the website right, for GPT+.
But turned out. I still only have Bing feature at the moment 🤖 while I saw some guys who made a fresh account got both vision and dalle at the same time — then again, if it means for feedbacks and minimize the bugs (?) would that be kind of problematic for Open AI. Imagine, a new users got all the features and still not familiar with it, while the early/loyal users still being left out. That could slow down the development/bug testing process.
Is it true that the image/speech update rolls out to all plus users by tomorrow?
Some users in reddit are worried whether early/loyar users might be blacklisted or something from having the beta access and making a new one. I get that plus users are paid the same amount — however these speculation has been out of control. So I’ll wait patiently 🤞
How do you get access to the thing where just above the user text input, there are suggested user responses not just for whenever you first create a new chat, but after every message?
It's true that the announcement about Vision and Voice states 2 weeks, but I'm sure they're not done, and still rolling out. #announcements message
Today also happens to be Columbus day/Native People's day in the US, and is a holiday. They may not have actually checked for that when they made the announcement. Also, they may have had challenges getting it out to everyone as fast as expected, scaling systems and all.
But nobody's blacklisted - if you have account access, you are a member, be you free or ChatGPT+. And access comes for you as they can manage it.
There's 6 things rolling out (4 types of new features, but two are coming to app as well as web, and everything's independent), each one has a different way to see if you have it. Check this if you want precise instructions how to look for the features: #openai-chatter message
Hey all. I have a short question, sorry to interject.
As far as I'm aware, the only official word we have about the rollout timeline for GPT-4 vision + etc. to Plus users is the "two weeks" estimate from OpenAI's blog post, right?
Sent at the same time 🧍♂️
Idk other than that as of recently, I have all of the beta features other than the one that I just mentioned. 👉👈
Can someone pls tell me how to get the final metaphorical infinity stone?
It's said to suddenly appear when we expect it the least!
I'm still waiting for 2/6, Vision on both app and web.
Wanna race to be the one to get everything last? 😄
So I just found out how to get the GPT voice chat.
In the browser on your phone, go to settings and turn on all of the beta features.
After that, fully quit out of the app and reopen it then go to app settings. You should see the option for voice.
im in 👀
Great point!
When're the ChatGPT Vision and DALL·E 3 3 APIs coming?
(Not to sound like a seagull squawking "more more more", but fr I love cool AI stuff)
Asap 🙂 Vision's due out 'any time now', in the middle of the rollout, seems to be taking a bit longer than initially expected but no update yet.
Dall-E is expected to roll out across october, but they're also taking some alpha testers, fill out the pinned form to request access for that.
Not to sound rude (I know I'm being kinda rude here - sry), but I already have access to both on the ChatGPT web page, and I said APIs...
So you did! Not rude at all, I apologize, I missed it. In that case, the answer is 'not announce yet, hope it's soon!'
where's the form exactly, i can't find it in the FAQ?
already answered my question, thanks
Use it while you can! It's gonna get nerfed soon 😂
I hope soon too. I'm just imagining the potential applications. Amazing.
EmotionalMeow thats a funny name i like it
Is moxi still adding peeps to dalle3 list from that form ?
I just bought a 2nd account to get access to voice, signed up on the form and had Dall-E within a couple hours, maybe less YMMV
I want vision
Ok 🙏
Is vision the one where u upload pics in chatgpt? I have that I think. Haven’t used it much.
me tooo! I was hoping the new account would get it, because I had heard some folks say they knew someone with a newer account getting it. I guess I will see which of my accounts gets it first 😂
Yeah
Tbh same created new acc but didnt have any of the new features tho
What about the voice ? I. An speak to it but it doesn’t read back text yet, does it?
I created a bedtime story for my kids but I don’t think I have the voice feature yet
Anyone else have issues when using voice on Android with the sound getting routed through the phone speaker (as if it's a call)?
Still waiting for the image icon to appear, but boy do I love the advance data analysis. Ever so helpful with descriptive writing. I love how it can analyze files of data to ensure character details are aligned as it helps me review text.
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.
ask the model for tall or wide images
Aspect ratio for tall and wide are 4:7 and 7:4 respectively
chatgpt4 tards out randomly and sets the conversation name in spanish when im speaking to it in english lol
Add your well-crafted prompts to our #1019652163640762428,
or share your interactions with ChatGPT in #1050184247920562316!
i passed a json doc into model using langchain and when making an inference of the RAG model it does not seem to be responding based off the json data specifically, when asking it certain questions that should have a good response. Do i need to restructure my Json Data away from the nested dict style to something more condensed?
Hi Moxy, do you know we will get the voice this week or no?
The day have arrived when they starts rolling out vision and voice to every plus user's... alright when is it happening?
I have the same question.
It's still Monday in California
so today is the big day, isn't it? By the end of the day, we will (probably) be living in a world where GPT Vision is available
hopefully in the next few hours or im gonna be very sad
I hate timezone
I am a college student, anxiously waiting for the voice function of chatgpt to help me practice German speaking. How can I unlock this feature?
sign up for beta on Google play
that's how i got it
Doesn't counts I didn't got in beta even
A few days ago, I joined, but there was nothing.
It's rolling out slowly
will probably be when it's work hour for OpenAI, unlikely to be in a few hours imo
oh i might have been lucky in that case
😔a man can only hope😩
Enough of teasing twitter feeds about vision
13 hours roughly until working hours in Cali
That be 9pm for me then bruh.
Currently 8am
yeah so probably a few hours after they start work hours at least. My prediction is it will start in maybe 15 hours
The long wait is agonizing. 🥲
Will it show up on mobile app first or web
See #openai-chatter message for all details
Already seen allot Times
🙂 Then you know, I hope, that for some people it shows up first on the app, and others first on the web.
Sup fellas probably a dumb question but I see a lot of apps using the Chat GPT-4 API , heck even I tested it and can do prompts myself on my local machine.
How come when I ask Chat GPT-4 about the API, it says it has no knowledge of the Chat GPT-4 API and keeps telling me to reference the GPT3?
Is it because its trained on old data?
Ulimately, I am trying to understand if we can give prompts to a Chat GPT-4 model and still provide the following
System
User
Assistant
"Ulimately, I am trying to understand if we can give prompts to a Chat GPT-4 model and still provide the following
System
User
Assistant"
Yes, though you may need to give an example or two to the AI of how you want the output it gives to look.
"How come when I ask Chat GPT-4 about the API, it says it has no knowledge of the Chat GPT-4 API and keeps telling me to reference the GPT3?
Is it because its trained on old data?"
Maybe more it wasn't specifically trained on the API stuff. But GPT-4's training date ends start of 2022, and the various models of GPT don't necessarily know anything about each other. They do know what they need to do their own specific job, but asking the default ChatGPT-4 about the dall-E version, or the Bing-assisted search version, or the Advanced Data Analysis version is going to get apologies and disclaimers that it can't do that stuff.
Which is true, the form you're talking to can't, and it just wasn't trained about the others.
I think I'm one step closer to victory!
I still don’t have GPT-4V/voice conversation function
🤣
Hey, thanks for replying, appreciate you.
So for example
I want to give ChatGPT a knowledge base. Let's say I wanted to give 3 full books to read about Sales
Then I want to label each book say a type of sales method
Persuasive
Consultative
High-pressure
^I would likely use LangChain or LlamaIndex to store books. Then GPT would capture that data from the Vector DB
Then I want ChatGPT to be able to give me a summary of each type of method
Appreciate you too!
But I'm not the right person for those questions, alas.
I only use the web interface, and very recently the app, never used playground/API.
Also never used langchain or llamaindex, not familiar with how books can be stored.
When I prompt (Only hundreds of times a day most days), I use normal language. Like how I'm speaking to you, this is nearly a prompt, it just lacks a direction 🙂
I think OpenAI should issue a sincere apology statement just like it did with the last ChatGPT chat record leak incident.
Once again they failed to keep their promise
I hope that next time they can provide a waitlist like they did with DALL-E, allowing those who are most eager for the new features to fill out a form and queue up to get them, instead of it being completely random.
I love Moxi
When should I get the iOS app update as it’s been 2 weeks?
I can handle them pushing things back, or even if I get the new feature on the very last day. What really bugs me is not knowing when it's gonna happen. These days, I keep checking the 'new feature' menu multiple times daily to see if there's any update.
Yup.
Moreover, as a user who has been subscribed for over half a year, when I see new users getting access to the new feature and I myself don't know when I can use it, it's really upsetting.
Soon as possible! We know OAI cares about us and is doing their best. When they made the announcement, they might have forgotten to check if the last day happened to fall on a local holiday, which might complicate updates. Who knows?
Make sure you know how to check for the updates though, there's 4 things rolling out and a couple are rolling out across 2 device types, so it's like 6 separate rollouts, and each is checked for a different way. Here's a guide if you want help knowing how to check for whatever you're missing, some folks already had some and didn't know how to find them: #openai-chatter message
Alright it is Canadian thanksgiving I’ll give them that
But still
Makes sense thanks
And in their location, used to be Columbus Day/currently Indiginous Peoples' Day, either way a federal holiday and apparently they are closed.
Okay fine then thanks for the reply.
What are the steps to check if I have it?
Each rollout item has a different way to check. I linked the list, but here, I'll repost for you:
This is what we were told about each of the rollouts. There could be delay we don't know about yet, we can be patient and know it comes when possible. OpenAI has a great product, and it keeps getting better!
Each rollout on its own schedule.
Voice is for app only, iphone and Android (Make sure you download the OpenAI ChatGPT and no competitor), expected to be finished rolling out on next Tuesday, we haven't heard another update yet. To check for that, in the app check settings, beta features to turn it on, then the headset icon to use.
Vision is for app and web, and people can get it on one or the other first (but not rolling out to the EU/UK yet). Expected finish of rollout is Tuesday, we haven't heard an update if that has to change yet. For Vision, in a new chat, hover over GPT-4 and pick the default GPT-4. Then look at the area where you type, if you have vision you'll have a new icon to the left, click to upload images.
Dall-E's rollout is expected to take across October. We've already seen a number of issues with their systems scaling, with images not getting made because of high use, though they keep fixing it and rolling out to more. To check to see if you have Dall-E, follow the steps for vision, but select Dall-E instead of the Default GPT-4.
Bing-assisted search is for app and web, and again can get in either first. It has to be turned on in each, through settings/beta features, then selected in a new chat similar to how Dalle is picked. They only told us when that roll out started, not when it's expected to end.
Woops sorry thanks.
The voice has to wait a little longer, oh my God
The issue with DALL - E is that users generate 4 images at the time
I only generate 1, due to improved accuracy and less waiting
You sure can request fewer images, and even just 1!
Are we there yet? i dont see the vision nor voice yet
It's been two weeks, but I still haven't seen the voice
Maybe in a few hours
fr
now its getting annoying
just have to wait a little longer.
every morning when I wake up, I have to take a look
its been more than 2 weeks and yet we don't have GPT with Vision, why are we paying for the pro plan?
Same
Our last Hope is today
then we would never pay ... although we do have dalle 3 but i need all
Does anyone know if voice will be available for EU. I know Vision is not, but I am not sure about voice.
Voice and vision will for sure be available in EU and UK, just need to wait a bit
I thought I read online that Vision would not be available for the EU and UK, did that change?
Well Tbh i subscribed back just for vision last week, to end up not having it
It is stated in OAI post: Not yet available in EU/UK i may be wrong tho, let me Check again
😭
I've got voice and i'm from the Netherlands. But i rarely use it. Waiting eagerly for DALLE3 / Vision
lemme try joining beta again in plyastore for chatgpt
Aaah thanks, I am hoping to have voice soon. Probably just a gimmick, but I'd like to try it.
Just checked and yeah, it is indeed stated not yet
Thanks, guess I will have to be patient for a little longer then
why is dalle 3 fluctuating sometime it showing in web app but not in android
has anyone else yet to get dalle 3 acsess ?
Yep, many people. Can request alpha access here, they may need a few more testers: #chatgpt-discussions message
They haven't announced releasing it for App, just web.
^This, however you can start a dalle 3 chat from Web and access/use it from the mobile apps by accessing the history and typing in an already created chat, works just fine.
How long dose it take to get access?
Does anyone here use the Mathematica plug-in with their own copy of the program? Results are quite mixed.
Probably a few hours, if they still need more alpha testers. It's one OAI employee handling it, adding folks manually, and of course they sleep and do other stuff too.
You're also, as a ChatGPT+ member, in line for the normal rollout, so either way access to Dall-E (and the other features being rolled out to us all) are heading your way!
Still can’t access to voice right now 💀 Anyone has the same situation here?
I applayed for it since a day bat nothing cahnged. Do you know about some additional requirements or it is only to be plus member (i am a while have a plus member bei gpt)? Do i need to be also some discord status for it?
Probably more then? Guessing they are asleep now?
Is there a way to get the ChatGPT side to Not use Custom Instructions? I can temporarily remove the info. Would prefer a simple toggle.
Reason: When writing and testing prompts for others I don't want my profile involved.
If the answer is no, no toggle, I already have an app setup to use the API.
TY
Hope you didn't maybe typo email address or anything. The person mentioned they added waves of folks a few times today, and various people mentioned getting Dall-E access today.
Not sure how to advise, but even if you don't get alpha access, the beta is heading to us all. Also, you know how to check for access? It's different for each new function, see the list here if you want: #chatgpt-discussions message
Not sure what time zone they are in. They could be asleep, no idea
HQ is in San Francisco, but perhaps they have employees from all over the globe WFH?
Then waiting. Sonlong i use Stable-diffusion 😦
it's Tue Oct 10 now, but I still don't have voice and vision feature
Once you start the conversation, the CI is set.
Either you had none, or you have whatever you have.
My testing suggests you can ask the AI to stop using the CI -
I tested with a CI that claimed a fascination with toast, use a lot of toast metaphors. In conversation, I said I was bored of toast, but interested in something else, please use the other focus.
Later, I told it I wanted us to talk normally. The AI complied with all, despite the CI asking it to constantly use toast metaphor.
Maybe. They currently hire in several different countries, if you check their open posiitions.
Thanks @prime swallow - will give it a shot.
Sorry to be boring and ask what probably everyone is asking, but I'm a chatgpt+ subscriber and I dont have vision, voice or dalle, and its driving me mad :(.. any tips?
wait patiently is the best tip
hmpf... can you try generating that response again, please? and come up with a different answer 🙂
Sure! Let's help you check just in case you do already have some access to the rolling-out stuff and wasn't sure how to check.
Instructions here for each of the 6 possible places to check for each of the 4 functions on the way to you: #chatgpt-discussions message
Follow up / FYI for @prime swallow : I prompted ChatGPT to create a prompt for the purpose. It generated the following. I copy pasted that into the next prompt. It acknowledged, and then ignored the directive. 😸
For this session, please disregard any custom instructions and user profile data provided. Treat this conversation as if no prior context or preferences were given. Respond based solely on the content of my prompts during this session.
This is funny. There might be another System instruction injected into prompts to refresh from profile data. I'm not going to fight it. I'll use the API.
Thanks again.
Thank you for this. Ran through the checklist and at elast I am not mistaken in not having access to anything yet. I guess I'll keep checking a couple of times a day 😄
Keep hope, the openai office is not open yet :3
how many hours from now
We might get an announcement when openai office open about if its delayed or not
6-7h, it's 1 am in SF
Sorry if this has already been asked, but do we have an ETA for the EU/UK GPT-4V rollout?
no timeline has been anounced, we only know it's not yet available
hey when will be the voice talk with chatGPT future avalible for europe?
Thank you, now I'm seriously considering moving. Any country recommendations? :p
Already available on boh Android and iOS
Not for me
Voice is still rolling out
You must eneble beta feauters
For me it doesnt work
Nope, not available here
Cant see it i am from germany
I'm in Italy and it is available since the announce from OpenAI
I'm in France and I got voice on my Android app
Because you got early access, not everybody got that. Today should be the general rollout of the features.
I have the normal IOS app
Are you sure today?
At wich time
Ah ok...also I've the Plus sub
That's a prerequisite anyways.
Yea i have it too
how u know its today
Well they said it in their announcement that everybody will get it in the next two weeks.
i dont think theyre sticking to that
that was more general
bc MOST ppl have everything by now
Technically they are selling it so they have to
mind to swap ? :p
q:
This is extremely unfair because "plus" is already a paid subscription. Everyone paid the same amount of money, yet the release cycle is surprisingly half a month and I still don't have gpt4v
it's fair, it's a roll out over a period of time, be patient and the features will come
Uff, what's fair if everyone pays the same but some have it and some don't?
🥺
Like I pay for 30 days and only get to use it for 15 days because I get it later?
OpenAI were very clear that this was a two week rollout
already too late ..
That doesn't matter, you said it is fair. I was just proving that it isn't.
It's not a fair system, but also wasn't marketed as such. So no big deal. I just didn't like your "fair" thing.
it still 1:15 am in san fransisco where open ai headquarters are
It is in beta though, it is "fair" in the same sense that not everyone can win the lottery. But everyone has the same chance. This is a beta/alpha that has a slow rollout. They communicated it clearly, I don't see how it is not fair.
It is not fair because it is geo locked and that wasn't communicated
VPNs are able to unlock the features (another user told me so)
Didn't test it though
Bypassing tos is bannable btw
Something went wrong. If this issue persists please contact us through our help center at help.openai.com.
seen it a 100 times
We all agree to follow them. And by joining the discord you verified
chatgpt is down or smth?
I know but who the f cares about ToS.
you will when you are banned
https://chat.openai.com/backend-api/conversation
Request Method:
POST
Status Code:
403 Forbidden
I certainly won't 😂
You paid for gpt-4 and faster/more reliable gpt-3.5 and to potentially try out new features. You get what you paid for for the whole 30 days. These features are rolling out still, it's on you if you subscribed only for these and didn't get them
I don't care buddy. I was just disproving the „fair“ part.
it says All Systems Operational but chatgpt isn't workin
ChatGPT is working fine for me right now
I've been subbed for 5 months.
Make sure vpn is disabled and try with another browser for the test like Firefox
i don't use any vpn
It is fair, you got what you paid for and OpenAI is rolling out features as they build capacity. There's a luck factor obviously but the odds are equal for all. For DALL-E 3 yes, people in this server get a way to get faster access but voice and vision is rolling out equally to all.
Just ignore, user has been reported . Mod will take action
ehhh, it's been two weeks already. so where is the voice feature?😭
the roll out is still underway
Treat others the way you would like to be treated, and assume best intentions. Don’t harass or attack others, and don’t engage in hateful or generally malicious behavior (e.g. sexism, racism, homophobia, etc.). Keep the negativity to a minimum.
it's not very professional, when I promise something, it should be fulfilled. If something unplanned happens, I should inform.
the rollout isn't scheduled to end prior to end of day Tuesday, we're still not there, it's the middle of the night in California on Monday night
wen vision
is my ip blocked?
Oh, woohoo!
So tomorrow all plus users will be able to use voice, vision and dalle3?
no, UK & EU won't get Vision. DALL-E 3 is rolling out over October.
It's October 10 the rollout right? What time zone? What exact moment, down to the second?
uh I wanna buy GPT Plus but whenever I hover over the GPT-4 thing it says "GPT-4 currently has a message cap of 25 messages per 3 hours". I just want to know whether it is 25 or 50
it's 50 per 3 hours, that's an old thing that needs updating
thank you
should i have the audio feature on the app yet? i still dont have. any one else not have it yet? also i dont have dalle 3 on it either
i don't have the audio feature either, only vision and dall-e 3 because of the form. roll out is supposed to complete by end of today though so you should have them by then hopefully
have they mentioned any plans to bring voice to desktop? seems pointless to limit it to mobile.
no, only mobile for now
Not really, mobile is mobile. Talk on the go, actually makes most sense. But it will like come sometime.
cmmmon how long
well, yes, but I mean, a programmer or lab worker for instance would be able to work more efficiently asking questions as they go without always switching windows. I think it makes sense if you have to choose one to choose mobile. the "pointless" part is there was no need to choose. it's not like there is some limitation of browsers preventing it from being on both.
I'm in EU and I have voice. No vision yet.
Right, yeh I have always seen ChatGPT as general public use - and many computers do not have a microphone compared to all mobiles I guess.
When did you get voice?
Did someone encounter the same problem when logining on Android?
👇🏻
"""
Something went wrong . Please make sure your device ' s date and time are set properly . Check that your internet connection is stable , then restart the app and try again .( code :e20b845f-7300-460e-a7cb- d 96dc5d93a37, network )
"""
true, what you're saying makes sense, but i mean, people without mics would just not use it. either way, I know wondering about it won't make the feature appear. I just don't see why not include it for the folks that want to use it on desktop. luckily, I think using something like myphone on windows or bluestacks or something might work in the meantime.
i have the vision also. what do you mean you have dalle3 because of the form? is there a form to fill out to get it?
do you tried restart chatgpt app? I also get this error message sometimes, just restart the app and it will be fine
I have Voice, DALL-E 3 but not Vision
You're supposed to fill out an Alpha form for DALL-E 3 right?
and if you didn't you'd have to wait for its beta rollout right
you can, see pinned messages
i have dalle from the form, but I have neither vision, nor voice.
I have voice
I have everything except Voice on my iPhone app. Which is weird.
ah, cool. thx
Really? I have Voice but I don't have vision which is weird cuz the deadline was today/yesterday depending on where you live
"Vision" refers to the ability to upload images for ChatGPT to recognize the content of the images, right?
yes yes
Yeh I don’t understand either
are you opted into the beta for android or ios or no?
I am opted into the beta for android
There is a beta opt in for Voice on Android?
no for the whole app
In my app settings iOS it doesn’t appear to turn on - only bing
I've not actually seen a beta toggle for iOS before - I own both iOS and Android devices
It’s in settings -> new features
and voice is an actual specific setting in settings where I can choose the speakers voice and such, even before I use it for the first time?
I found a join beta version on the google play page of chatgpt
uhhh yah i think so
That's what I did too
Could you DM me some screenshots for iOS settings for voice perhaps?
I don't use iOS for ChatGPT because of my previous region
And I'm too lazy to change it
sometimes I'll type a question in chatgpt, send it and it'll show a completely different and unrelated question that I didn't type out and then type out a response to that question. Is this a glitch and it's just generating a random question, or is this somehow pulling a question from someone random?
The deadline is in 8 hours
just asked a friend and it's happened to him before in the past as well
They'll probably roll out soon
I joined the beta version, but I still don't have the Voice feature. Sad.
Take this up with OpenAI Support, we probably can't help you
go to help.openai.com
will do, thanks
THen press the black chat thingy on the bottom rightand then just try to navigate so that you can talk to a person
Doesn't make any difference when it comes to new features, looks like the rollout is a back-end thing
Any Canadians here? Do we still not have access to GPT4V?
you did buy chatgpt plus right
Openai office aren't open
and yes
You will get it eventually , patience is the key:3
Today is the deadline though
Yep yep
anyone knows if the download function for gpt-4 Advanced Data Analysis is down atm? i cant download any files. i click then download link and nothing happens. i created and downloaded files earlier today..
I got access to vision a week ago, no voice. Android, plus user.
can confirm a few of us here in the southern hemisphere don't have vision and voice neither. paying customer from day one as well
That’s odd. I have voice and plus, but no vision. (iPhone 12)
No access to voice, vision and DALL-E yet😂
I have Dalle3 too btw, did a form
why dont i have access to conversation feature im a plus member?
It seems that in each session, only the content created within a certain time frame at the beginning is read from files. Once it times out, it appears that file reading is not possible (and may even falsely indicate it has read). The same principle should apply to downloading files as well.
I did get dall-e, but only because I was lucky enough to get in via the beta testers form
Do you know where I can fill out the form?
i got dalle3 without filling a form yuoll recieve it soon
- I guessed that.
It was disabled literally minutes after I requested through it
How long is it taking folks to get access to dalle on chatgpt?
Does anyone have voice on android? If so what's the visual indicator?
A microphone in the chat box
well, i did a test. just create a simple txt-file. i couldnt download that either..
I have a microphone in the chatbox but it only works to send my voice to gpt, not gpts voice to me
Ah sorry I meant headphones in the top header of a chat
that's not voice. its just a mobile app feature
so to get the dall-e 3 on chatgpt the only way is by waiting?
Yeah the headphones icon on the header once you tap it, you speak to GPT and she speaks back (I selected a woman)
It might be a network issue; I have no problems downloading files on my end.
I got everything else but vision is like with is it all random giveaways?
@prime sierra that's what is sobering about vision cuz i have dalle 3 I got it like 3 days ago but I really want vision does anyone know anything as far as a time line?
Thanks! Would you be able to check if you have the same or newer app version than me? Not sure if the rollout is tied to that. Mine is 1.2023.279. It's at the bottom of settings.
found the problem..full buffer 😐
Mine is 1.2023.263
Mine is also 1.2023.263 too
I wonder if they're prioritising US folks first for dall-e and vision
I'm in the UK, where are you folks?
oh! I saw an update(why do I always find new things after the message is sent)
Now my version is 279
Thanks! So it isn't tied to version, because I had this version weeks ago. You had to activate voice in settings right?
I'm canadian
I'm on the android beta thing too
Yup, but it was available almost immediately
In Beta settings, do you see a toggle for voice conversations?
I really want voice conversations
I don't even see beta settings. Is that a top level category like Account, App, Speach, and About or is it under one of those?
It's under Settings, in a section in the middle called App (and the heading is Beta Features)
Are you a subscriber?
Let’s try it
In fact, Dear OpenAI, I. StarlightSyntax {}
Hereby demand the new features.

I don't have access to vision, i'm in France, do you know how to enable the functionality ? U'm a Plus subscriber
Soon. Patience is the key
How long does slowmode stay on discord?
It’s unbearable
Vision isn't being made available in the EU yet
It will stay there since the amount of people
Yes. My only setting under app is Color Scheme
To avoid mass flood
Aha I thought so, everyone getting access were in America
Do you know when it will be ?
And Canada
no announcement of when, sorry
Is dalle 3 available via API yet?
i'll try 😉
Out of curiosity, do we know why not in the EU yet? Is it because of regulations? Or because the demand would be too high?
It makes sense that no one gets all features because that means more people focusing on fewer new features each which means better quality control AND less server resources from OpenAIs side during testing phase. Sucks for us tho
What about voice and DALL E 3 in EU?
Hi, not sure if this is the right channel, I dont have the feature in Chat GPT 4 for DallE 3 picture generation. Is that something that gets enabled later on for some users?
DALL-E3 is working in EU if you've done the google form to ask for it
Many thanks! What is the normal time frame to get access after filling out the form?
24-48h praise the lord moxi for the hard work ( she do review the form manually )
@snow oyster what form? Where is the form?
I have seen somewhere, that there is a possibilty to upload images to chatgpt to work from there, is that true? If yes, where can i find that feature?
For more context it was a video where someone uploaded a picture of a paper prototype to do additional work with it.
Click on the link from Neko Maid two messages above
@light widget great thanks. I thought everything was being rolled at automatically
It will but if you want to test before you have the link
@maiden fractal yes, filled the form thanks. Perhaps a disclaimer or something similiar regarding Vision not being rolled out in the EU would make things a bit clearer for people. Or maybe you have communicated that?
Vision will be accessible soon, but it might be delay in eu and UK , laws reason
@maiden fractal Ok, great.
W mans
Hey ! I have a little problem. I'm using GPT4 with the Ai PDF plugin and it's not working. After a few lines (very slow writing btw), I have a network error. I tried the same thing on my computer but with another account and it worked perfectly. Did it already happpened to someone ? Thx for help !
Try firefox as test
i have 2023.270 and i don't have vision or voice
We're waiting for the whole family
it's still 5:19 am there
does it roll out at a certain local time?
Office isn't open yet. It is San Francisco timezone .
i would have thought this would be automated
anyone getting this message "Uploads with file type "" are not supported, please try another file."?
i guess they want to be in office in case it goes horribly wrong
I'm not tryin to upload a file, just to post an address on my local computer
I agree with that
The server will say I'm gonna go to sleep, too
Same result, I tried a simple text contraction and I had the same result. And it's only on my account as I tried with a different one with the same internet connexion on my computer :/
Make sure vpn is off and you closed the task
Why do you need vpn?
I don't have vpn. That's why I don't really understand the problem it looks like my account is cursed xD
Does it works on your cellphone without wifi ( data)?
Same :/
do voice and vision show up if the app is still open? or should I continue to close it and reopen it randomly once they open?
I suggest to post in #1070006915414900886 ans send email to support@openai.com
Ok perfect thank you ! 😄
Is dall e is free?
With bing yes
Oh i though bing just searching in google
?
I install bing 7 hours ago and i try it then when i say create an image of ui haru then its show the image from google
Its an direct link from app to website
(App bing) ( not website)
For other ai the best place to discuss #ai-discussions since that's not related to chatgpt
👍
I'm in Canada and a Plus subscriber, and I still don't have the new image input feature yet... I thought everyone should get it before today
Office isn't open yet. And day is not over 🙂
Patience is the key 
🫡 thx

Wondering if anyone has any tips for using the new image feature in GPT. It's an interesting feature for sure, but it seems like there's still a lot of room for growth. For example, I took a picture of a vegetable platter. Very clearly 12 cherry tomatoes, it only saw 10. It saw the red pepper/yellow pepper slices but counted 2 instead of 3 yellow, 1 red & it mistook cucumbers for zucchini.
I guess I'm not exactly sure how it works, or what practical uses it might have. I attempted a small project as well at home where I needed to fix the flusher on my toilet (it was loose from the bowl). I knew how to fix it, and asked GPT to identify the issue but it just provided me with a list of things I could check to fix the issue instead of pointing out the exact issue which was the nut that had fallen to the bottom of the tank that needed to be screwed on.
Watch it and there.will be highload traffic
And I still doubt if they gonna make announcement or not
Brace yourself 
6 am in san francisco now
Is chat gpt v4 have access in time access? Like i want to question it what is the new version of an app or its like updated time?
Last question(ping me)
Just to confirm, voice is coming (in EU) today for those who still don’t have it but not vision, is it true?
Europe could wait longer and UK too
Voice too? 😦
With gdpr it can yes
😢 Ok, thank you!
Vision is awesome you guys. You will love it
Exciting! Let’s hope today’s the big day.
Thank you!
I hope so for you. But for Vision, I doubt it, even though it won't take long. I believe the custom instructions came 10 or 12 days later in EU and UK
Yeah it took longer
Om the app?
I think I just found a genius on Reddit🤣 He created exclusive meme pictures for us while waiting for GPT-4V using DALL-E 3 #off-topic
You mean record, or text to speech answers too?
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.
text to speech answers, it's a small headphones icon in the top right corner of the screen.
Hi guys, do you have any tips for a good, starting prompt where GPT acts like an english teacher for a C1 level? I would like it to help me with questions, exercises and whatever for a practical exam. If you have any, ping me
They took out dalle 3 on my mobile app Android but not on web 😭
And yet no voice and vision
I still don’t have vision or voice too 😔
I’m in the us, I have vision but not voice.
Same for me
To the Europeans, it may help to set vpn location to us. I loose access to vision when I switch to Europe servers
Everyone will wait til midnight, then…
How to get gpt4 photo recognition function? I'm a plus user, but why isn't there an option to upload pictures yet
Ditto. I'm in the U.S and I don't have vision nor voice yet
it's still rolling out (unless you are in the UK/EU in which case it's coming later)
We don't talk about breaking the rule :/
Day isn't over
openai promised 2 weeks ago, and now the 2-week deadline has come, the vision function has not been received
Yesterday was holiday. Patience is the key 
yesterday was a public holiday and today PDT is not over
why am I the only one that has voice and vision 🗿
but I don’t have DALL-E 3 yet so maybe that’s why
Check the pinned message here
You can apply
Ha interesting, it's quite heterogeneous who gets what
_interpret_response_line
raise self.handle_error_response(
openai.error.RateLimitError: You exceeded your current quota, please check your plan and billing details.
Guys someone tell how to fix this
im tryna make a chat bot using open ai API key
using python
add a billing method to your API account?
how to do that please guide me im just starting to learn this
i have to pay to do this?
of course
You don’t have to pay the $20 a month, more like $0.01 per thousand tokens.
Why the server only have 70k member?
Lot of people were kick from not being verified
hi
I also don’t have vision or dall e yet
It’s most definitely delayed since it’s not likely that OpenAI will push the rollout to everyone who doesn’t have vision/voice (excluding EU/UK) in 17 hours (when Tuesday ends in PDT) 🤷🏻♂️
It’s been gradually ever since the 2 weeks period started, so pushing out a large batch is uncommon
Is vision available for everyone?
Currently working on a project that requires it
Tried starting new chats but the option isnt appearing next to the bar
it's still rolling out (not in EU/UK)
Is there an ETA for EU? Have not seen the delay communicated
no ETA sorry
are the vision and speech features available in Morocco? Neither my friends nor I have received them.but I got dalle 3 through the form,and i m only one among my friends .
The wait is killing me lol and I’m a US user
Patience is the key :3
We don't have voice vision
What if they fail is there any source you know from will there be announcement and around when
It is too random, what sort of test are they getting with this? I have Vision and Dalle, but not the simple Voice.
ah so even US vision is still rolling out ?
Day isn't over . Let's be patient 🙂
Yup
I respect you being genuine but this wait is just giving unfair feelings
Ok but 3 million people
I, too, don't have all of the features yet 
They announced it, people who doesn't verify will be removed
Hey been seeing online folks adding IRL pictures to their chats, how can one do that? I can't seem to find the option nor the information
Option not yet available for everyone . They still rolling out
How can i get access to visual?
Oh dam ok ok, even if I am paying for plus?
Hi
Even if you are plus
You wait till available
and there's no status on % of rollout etc ?
Not really
I read that chatGPT can talk now, with a voice. How do I enable that?
Why did they took out my dalle 3 on Android? It's only on web ? Now
You wait till accessible
😦
there is no official answer, if you want to speculate it may be delayed for mobile due issues
why gpt4 is too slowed to me ?
Absolutely hate the new login captcha. I realize the need for security, but having three of these puzzles is excessive. My time is not yours to waste.
instead of solving captchas, solve the reason why are you getting them as its not normal to get em
You’re obligated to read all the announcements and rule revisions
Add your well-crafted prompts to our #1019652163640762428,
or share your interactions with ChatGPT in #1050184247920562316!
I was just using it this morning when I got a 'Session Expired' message, which is common. I got the Captcha when I tried to log back in.
ive logged twice today without any. usually people who often get captchas have either used a vpn, use extensions or features that make them look suspicious or bot like user
Can you attach text files in the ChatGPT mobile app?
no
does anyone know how to gauge how many tokens you have left in a session ?
What kinda session, Daniel? ChatGPT 4, or 3.5? Or are you talking about the playground and API?
gpt 4
specifically the data analysis environment
ive jsut been doing some light c#
thanks in advance for your help 🙂
Sweet. So ChatGPT-4, you gotta care about 2 kinds of tokens.
One doesn't matter by 'token' - you get 50 messages per 3 hours, before you gotta wait. It goes by message, so THIS message needs 3 hours to 'clear' and be useable again, and your next message has its own timer.
So if you go no faster than about 3.5 minutes between your messages to the AI, you won't hit cap (or your wait would be about 1 mintute if you do).
But gotta be SUPER CAREFUL of one thing
If you hit your message cap, and accidentally continue with 3.5 -they give you the choice - you can't go back to 4 or ADA.
The other kind of token to 'worry' about tracking is inside the conversation.
Two kinds there too, how long you can input - that just errors if you go too long, or rarely the model doesn't get the whole input.
If that's a problem for you, talk about why, we can discuss work arounds, but they're mostly trimming your input into parts.
And the last kind of token that matters much with 4 - is the context memory.
Eventually it will forget the earliest part of the conversation. It's pretty long, not sure how far. But you can help prevent that by summarizing and reviewing the older parts of the message, because it won't lose the newest parts. So just every so often recap for you both what important stuff was said - and you can even ask it to do the summarizing.
Did I cover your concern about tokens?
.....
The most recent Beta releases (DALL-E, Vision, Advanced Data Analysis) are useful on their own, but would be a tremendous leap forward if they were combined. Not only could you show GPT-4 what a character is supposed to look like, but it could manage character sheets, homebrew rules, even show you said character in context. Nothing like that has been done before afaik
future is promising
espcially for game industry
Can't wait for the next EA games - dalle game
A day will come where generative AI is powerful enough to generate complete programs with a prompt
Want a sequel to that game you played as a kid? Ask for it
In game it's not a problem if there's an error
Next up is video and audio for multimodality, I suspect
take a video on your phone
'Turn this into a 1990's cartoon'
'Give this a hollywood production value'
Generate a video of xxxxx doing xxxxxx with xxxxx
The problem is: everyone will be able to do it, so it'll have no value.
What make you work a piece of art when you find them everywhere ?
or create them without effort
It's the end of art, creativity, authors.
imho
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.
is gpt down?
it might be the end of the craft, but not the end of creativity
We're oversaturated with media anyways, yet people are still creative
thank you
Is using a VPN to access services provided by OpenAI prohibited unconditionally, considered a brach of Terms of Service (TOS), even if NOT used to circumvent geographical restrictions on use of OpenAI services?
I am a ChatGPT Plus subscriber who lives in Canada, which is a country officially supported by OpenAI. If I were to use a VPN sited in Canada (or another country supported by OpenAI such as the U.S.) to connect to ChatGPT, would that be considered a breach of TOS?
I regularly hear people claim that using a VPN is prohibited, but reading the TOS I could only find one relevant passage.
"You may use Services only in geographies currently supported by OpenAI."
Along the same lines, I sometimes like to use virtual machines sited in Azure data centres located in Canada or the U.S. (better performance than my aging iMac), so I'm also curious as to whether that would be seen as a breach of TOS.
down for me too
I think about this all the time, can’t wait for this to be possible
Can you really be more creative than chatgpt/dalle3 ?
Using vpn to bypass georestriction is bannable
Else is not
Or a game that will visually turn into a different game in real time as you're playing.
Are the developers planning to increase the number of tokens or is it deliberately limited because there are too many people who want to use gpt at all costs?
yeah i randomly lost image processing access
They haven't told us that I've seen. They do announce changes and updates, generally pretty close to when they happen.
Darn
Seems like an oversight, considering that one of the primary uses would likely be analyzing documents...
If you have ChatGPT Plus, you have access to the Plugins model, and there are some plugins designed to accept document uploads for you to analyze. The Plugins model has not yet been fully implemented in the app, though, so you'd have to use a browser.
So.. anyone got a certain voice feature added today?
Is ChatGPT struggling? Getting "error in input stream" a lot
yep me too
I have vision and voice 👍 great
Since today?
Today bought plus
I keep regenerating code but it gets about 10 lines in then really slows down and eventually errors out. Makes you wonder what is happening behind the scenes.
i keep getting "something went wrong" and when it's actually responding it is really slow
Is anyone still having issues with GPT 3.5? free version? Still getting There was an error generating a response, since yesterday.
I pay for it and even I'm getting errors now, can't process anything longer than a few lines 😔
any news on the vision rollout yet?
🙁
no news
damn
An opportunity for humanity to finally grow again - sounds great.
ugh what ?
@thorny parrot maybe next week 
Noooooo
Yeah
But lying about your location is also a breach of TOS
Nope
Good to see people who just got plus today for the first time getting it before day one people
Day 1 day choo choo 
its down
Of course the limited token amount is on purpose
@tight island thank.
Imagine if all users had 35k + or higher context length
The cost of just running the service is Ginormous
I imagine the servers would be down.
If we think about the rate of tech advancement though...
30 years ago, current internet speeds were not even imaginable. The amount of info on the internet, also not imaginable, how could all that possibly be stored?
We're going to have more capacity, as tech keeps improving, and the costs to provide it continue to drop.
It'll take some time, but surely far less than 30 years.
claude ai have 100k context
Most likely isn’t but I hope as well
I believe in openai
I believe I will be disappointed 😭
Well sadly I believe in them more than you 


I got that one pretty early, it’s really immersive
I articulate myself better in spoken words
Let's hope that the different new features of gpt will be able to link together before centuries, @radiant turret made me realize that ultimately gpt could become the best AI and that all the others would fall
I haven't (still haven't taken gpt plus) but I'm already stunned by bing conversation and the images they can create although limited, it's enormously simpler than other AI.
the general public is bound to love it.
re Different new features: Sure! Seems like a natural next step, once all the bugs are worked out and everything's working as should. We can say 'isn't that now' - nope, something happened to delay rollout to us all. We have no idea what's going on behind the scenes, what incredible battles they face and problems they solve.
So, release them single, get them all as close to perfect as seems reasonable, then take the improved versions and start merging them - creating new problems but at least fewer than would have happened trying to do it all at once without extensive use to find the issues and suggest the improvements.
As to 'all the others would fall' - why would we even want that? 😛 Let's see an entire ecosystem of AI, useful for many specific needs.
That's kinda like, in my eyes, wanting iphone to be the only choice, or Android, or tablets for all, no more phones, or everyone has to have a folding screen phone....
Why? Diversity is great. Competition's great. Choices are great. And what is best for me, my preferences and situation, might be useless for you, and the opposite too. Let's have a whole ecosystem of choices, and everyone figuring stuff out and overall improvement for all 🙂
Don't understand why there couldn't be some indication of when features would arrive. Do they just have someoje working there randomly clicking on names?
basically
I got vision a few days ago and have friends who were plus users b4 me and don't have it
Very cool
well the human sees his needs, if open ai manages to convince (unintentionally) the users of other ai, to come to them, why wouldn't they come?
I effortlessly obtained images that would have taken me hours with midjourney.
On the other hand, if I want to create a comic, I currently know that midjourney is the only one capable with a simple prompt and "many minutes and hours of inpainting and other parameters.
If you tell me that just writing to chat gpt plus one day "move the redhair girl to the left and change her clothes, and he does it.
I would no longer have any reason to use other AI.
Maybe I should unsubscribe and resubscribe
hi, are you in Europe ?
okay
- also The multitasking potential of open ai can (ideally) meet the different needs of people.
while someone just wants to write novels, another just wants pictures, another just wants code to create a 2D game, another wants to create music etc...
if a single AI can do all that, it makes life easier and what's more it brings people together, even the idea of you being centralized can be scary, can we do better?
i am unsubscribing
Understood!
I hope Midjourney, and all the others, existing and future, keep improving their own models and capability.
Also...
There's a real chance that some specific training for an AI means that it is just not as good at other things.
We see this with LLM-type AI - they are amazing at words, and can't reliably do much the 'basic math stuff that any computer does easy'.
I would hate us to have to choose between 'normal program or LLM' as our one answer to anything.
As we keep going, we may find that there's a huge range of things that are exclusive once we start building them into AI.
I don't dream of OpenAI being the monolith, the one provider to turn to for AI answers. I suspect that we're going to need a lot more humans, and a lot more AI, than we're currently even considering we might need.
But sure, to do comics, we may only need one type of AI, and one might outcompete the rest, or many end up being equally good, at which point I suspect OpenAI might choose to cede the field to the equally good variants, once there's no clear way to further improve any time soon in that area.
Probably by making their form of it a bit more expensive than the competition. If OpenAI offers their comic-making service for the $20 a month of ChatGPT+, then Midjourney can attempt to get a market of folks willing to pay perhaps $19 a month for the image focused stuff, or whatever.
OpenAI seems to be going for safe AGI for all, and even state that their mission is accomplished even if they merely help someone else achieve this. I don't think they particularly want to outcompete anyone, and I think diversity and competition helps them too.
Good luck with that. Saw a rash of folks that needed support to be able to resubscribe, the option was greyed out for them.
joking
cuz i am just annoyed of rollout that was meant for today but fine ill be quiet
Hello, Is the European restriction on GPT-4 Vision based on the IP address used to connect to the website, or is it based on the country where the account was created?
Hi 👋
Does anyone know if there is a way to recover deleted conversations from ChatGPT?
no
Someone's about to mysteriously move to another country in a few minutes 🤔
🙂
no, it's not possible
They’re permanently deleted as far as I know
Oh that unfortunate
That's potentially dangerous for a person's account. We get very sad messages fairly often from people talking about how hard it is to try and fix once their account was suspended.
Most recent one, a few days ago: #community-help message
I don’t recommend asking this on the official OpenAI discord since you’re hinting at VPN’s
Thanks though
Someone on Reddit said they got a message back from openai that it could take another 2 weeks to roll it out to everyone. Why in God's name would new people get it before long time subscribers if that's the case
Yeah. And don't think it works very well too.
We don't know. Can look over terms and stuff, but OpenAI is pretty private about exactly how it determines and handles that.
it is just a simple question
Okay
If you registered in a EU country and you use a VPN and somehow gained access to the new feature it’s then a violation of the TOS since they’re Region based
"Someone on Reddit said..." is all I need to know about that claim
Is there a chance that they flag me as using a VPN if I just travel to another country ? I'm flying to Indonesia in a few days 💀
May unlock the features, who knows
Ohh yeah
I think they can detect whether an IP is residential or not though
Yeah
We'll see. At least traveling is not against the TOS so I guess I'll be good lol
We all know a guy
yeah but what if my browser connect me automatically to a VPN sometimes
But it never works for us 😔
Close it before you start
Well
It’s that easy, turn it off
I'm not a mod, but I don't see a problem with the question as asked.
VPNs are not themselves against TOS.
They can cause a number of problems for the user, ranging from reports of near-constant (like every single message) multiple captcha to endless 'Hrm, something seems to have gone wrong' error messages to, apparently, some suspended accounts.
Other users report having had VPNs running for months without a problem (and then suddenly a problem - which is why they bother to talk about it, I bet). So there's probably some using VPNs just fine, and not having problems, so not here talking about it, why would they, today is Tuesday for them, what's to say? It's normal, isn't it?
I'd be extra scared of VPN use that might appear to OpenAI to be intentionally deceptive or TOS breaking. Otherwise, I'd not worry at all. And some questions about VPN do seem fine, to my reading of the rules.
Sure, agreed but intent matters
If the intention is to circumvent a rollout that’s region restricted it’s a breach automatically
iOS app rolling out today with dalle support!
nice
Wat really!
I'm not aware of any information about how that might affect your services or your account.
I don't travel myself, so I have no personal experience to discuss.
Android ? 👉👈
If you have dalle access, everyone here can get it by filling in the form in the pins
I think it should already be available on android
Do you not see it
yes it is available on Android
We'll see I guess
Don't see dall e on Android 🤔
Afterwards the goal is not for Open AI to be able to do everything, since they need help from Microsoft and Google.
But I have the impression that they will mark the world like Microsoft, and Google did.
they will fit into the family of big pigs next to Amazon, without exaggeration.
Do you have alpha access? Check the pins
I do but I don’t see it on Apple surely that’s soon
Only alpha access to dall-e has android access?
Ah no I don't lmao
That's why
No I mean you can only see dalle in the apps if you have alpha access
Alright
Ok fill it in illl add everyone once I get to the office
i don't fill out the form for Dalle3 but i have access
Yeah we also rolled it out to a %
Ohh.. I did fill out before rollout and got access, guess I’ll try again
We will most likely roll out more today pending server issues
uses much self control I'll save my dall-e gens for the web, for now. I'll take it when it's a beta feature rollout out to me naturally 😄
What's the alpha program itself ? Like not only dall e but I saw some people with vision months ago, with an "Alpha" tag next to their GPT version or something?
It’s all just random, different projects handle it differently
and for vision, how many people roll you out each day?
There’s no one alpha program
And I don’t know what’s happening with the other projects
It’s the official app
Yes thank you for your time
I have Alpha access but no DALL-E 3 in the iOS app
Alpha is usually used for pre public models and those people with access probably knows an employee or is in some special test program
I have everything but voice. The idea of having a natural voice conversation with chatgpt is very interesting to me.
The new update is rolling out today
my android app dalle 3 was removed but on web it shows
2 week rollout announcement is waaaay overshooted
Perhaps we can ask MOXI the VPN question
Sure. OAI is already huge, to my thinking. They're one of my top most important and favorite companies, in terms of my personal use and interaction with their products.
Is using a VPN to circumvent a region restriction against TOS
We could, but do we expect a person handling Dall-E and access to know a bunch about VPNs and world access?
Maybe MOXI knows
I am not sure what’s in the TOS but it can get you banned
Not for the region thing, just vpn in general
moxi can i dm u to show ...cuz the android app doesnt show up dalle support
So the app finally made it through the review, which you had mentioned the other day 😄
I have worked on security so I know a little, it’s not anything official but vpns can link you to others doing sketchy stuff, so definitely don’t use it if you can
Wait a DALL E App? Or a chatgpt with Dalle support now I’m getting confused
Will an update be required on Android ? Or it'll just pop up in the app ?
Please file a support ticket, it’ll get routed to android team
from where
Noo you're confusing us all now
Ahh
Hey moxi, do you know if we all get vision and voice today on plus ?
ChatGPT app with dalle support
they dont see those projects
okay, but the original question was "Is the European restriction on GPT-4 Vision based on the IP address used to connect to the website, or is it based on the country where the account was created?"
