#chatgpt-discussions
1 messages · Page 25 of 1
Comparing bing gpt 4 to chatgpt is disingenuous at best.
Can always ask to support@openai.com
besides folks, this is still all just the beginning in the grand scheme; AI is currently Frankenstein, when its all refined and reduced this will be equivalent to an old IBM mainframe a few decades back vs our smart phones now
oh
In fact I'd be a lot happier if OpenAI did have the same model for + as they did for API, pay as you go.
its not the same right?
Not even close.
I didn't say 20 bucks is nothing, like 1 or 100 isn't nothing either.
But if I tell you "Hey, there's this conference soon, we would like you to attend, the plane ticket costs 200 but if you decide to come, I'll buy it for you if you pay me 5", I'd say if you say yes, the plane ticket is cheap
It isn't the same indeed chatgpt 4 is better
but gpt 4 is not good enough rn
just imagine bing in comparison then
yeah exactly, why spending the money then if its free
You still have one minor disagreement with bing, it forcefully ends your chat. 🤣
You fail to understand we are not working for the company. Their costs are irrelevant to our dollar.
i really want to check how its api work
I simply do not care how "affordable' they're making it. Besides, not everyone is from the US, and last I heard they don't do regional pricing.
yesterday I got the new update and it was rolled back :/
fairly certain its just fast model switch code which isnt great because interpreter fails non stop. Either way its just starting to get used. they were clear its unstable but they want people to get to play with it all
The api is their best product tbh
So what, are they supposed to make people pay the real cost, and have people whining because it's too expensive and therefore not "open" enough ?...
yea i still pay 20$ monthly in korea
I wish that was the case hahahaha, but it kind of isn't. Currently especially from my experience, I pay that 5 bucks, and I get my knee broken backwards and shut up if I have the audacity to "complain".
and yea i feel like a clown rn
Uh no? You could just respect the fact that we owe them nothing. They're not doing us a service, we're literally paying them for access. Also, they offset their costs through tons of business contracts and other deals. Not to mention they profit off our data if we want our chat history saved.
Pesonally I messed around with assistant, ~10+ documents, a few prompts, for about 10 minutes and ended up with a $6 bill.
I then spent all night non-stop with chat, and was billed nothing but subscription.
Even if they go at a 50% margin, there's no way that night alone made my subscription worth it
That and they enjoy major investments, such as the one from Microsoft a little while ago.
Literally.
Neither do they owe you something contractually since you agreed to it
as a patron/consumer that is your power. you can always not pay someone for something. NOW, if it wasn't a great product overall I dont think you would sit here upset talking about it either. you can alway cancel. I don't pay for Netflix for the same reason, a product im not happy with.
The OpenAI Discord is an actively moderated server.
• Refrain from sharing inappropriate content on the server. This includes but is not limited to messages, media, or other topics of graphically violent, sexual nature, and drug-related content.
• Report all sensitive and offensive content in the feedback reporting tool in the ChatGPT web UI instead of here on Discord.
yeah why should we pay 20$ AND my data
dont
What are you even talking about at this point. Nobody said anything about owing someone lmao. They failed to deliver on a release, and we're criticizing them for that. As we're allowed to do.
Not to mention how a crap ton of other companies use their products on a far far bigger scale than we do, given how OpenAI literally only caters to them as well lmao, hence me saying that we are literally 2nd class citizens.
This isn't about being "owed". They messed up. Sam himself admits this. It's not up for debate.
there should be better way instead of just saying 'buy or not'
iirc OpenAI even said one time that the entire reason for this "excess" in expected usage, was exactly from big companies.
folks, it'd be different if this was a pressure sale or something, relax. lmao
I love how we all discuss that as if our opinions mattered and had a say in that ahah.
It's a winner takes all, best LLM is gonna have most of the usage, and no matter how much people whine about it, they'll end up using it
Tbh the folks who try to talk down on others and devalue their time and money because they're "fans" of the company make this chat more tense than anything else.
i pay for their service so i have my rights to speak
It's not about that lol, it's about us paying 20+ dollars a month for a big plate of steamer when it comes to quality and usability. And having to put up with that, and complete ridicule from riders if we do open our mouths even a little bit lmao.
For real. Like just change chat tabs. We do not care if you love daddy OpenAI. Lmao.
I am just telling you company should listen to consumers
Why should they listen to consumers? Companies have no interest in that when there is exceptional demand.
I don't even care if they listen to the chat or not. Sam only spoke up because investors/people who have real say are definitely talking to him about this because THEIR customers are being affected.
That's the difference between this release and the last one
because we pay them for their products
They don't have to if they don't want to, and they can keep digging their own grave, if that's the kind of intelligent path they choose to walk down, instead of embracing and truly improving and developing this new and promising frontier, so be it. That's their choice.
that really the one thing that bothers me a lot and sometimes it has bugs on it, like for example it has started explain some thing in paragraphs then suddenly all letters are vanioshed and the message appeared that said: "Hmm lets move on to another topic" "am sorry but i have to end our conversation here' _-
well obviously, and most do. However, sitting here throwing it against the wall and saying you wont eat it is sily instead of just relaxing and not being nasty about development. they also could have said "nah, they can wait another month"
I will focus on using api and implementing web environment better than command prompt this vacation
Capitalism is a prison for you, I'm sorry. Make choices, friend, and leave if you are experiencing distress.
I'm pretty sure it's more so about it just being the ability to share with other people in your frustrations, and that being good enough already lol. As sad as that is.
Plus who knows, a broken clock is right twice a day, we might wake up tomorrow and OpenAI suddenly goes thru a change.
"Being nasty" is when we criticize failed product releases.
Be so serious.
I just hope the corporate pressure OpenAI is facing this time leads to actual change in their release process. This is getting tiresome and tedious.
AI is in an explosive situation with many small and large groups developign at exponential rates. thy are trying to stay in the game long term by sharign fast. im cool with hiccups. you want compensation for loss, maybe just try customer support. im not that worried about it. this is all just the infancy of it all. next year is going to be incredible
They will need ages to get to Google's uptime standard
Hence why I said that they definitely went guns blazing into this without any foresight or planning, which kind of isn't our problem.
Aye. Unfortunately/fortunately OpenAI is building an excellent moat with what's to come. That's why the heat is really on from folks who have a say.
hello!
all day i have issue:
Something went wrong. If this issue persists please contact us through our help center at help.openai.com.
can you help me?
They've already had almost a year and so much money to work and improve, and I've hardly seen any increase in positivity in general from people. ChatGPT userbase is also constantly at a decline apparently.
what are the small chat bubbles with numbers next to the 3 dots in the explore GPTs menu?
1000%. Fan boying/girling like they aren't spending money on this big company.
Exactly hahahaha.
And I seriously don't get what the hell they get out of it, literally nothing, so why are they such imbeciles?
Unless they are on a payroll from ClosedAI secretly or something which case fair play, but I'm pretty sure they aren't.
And it's not even just casual disagreement, straight up screaming lmao.
and denial
Add your well-crafted prompts to our #1019652163640762428,
or share your interactions with ChatGPT in #1050184247920562316!
Would you still feel this way, that if a week later, this issue still isn't resolved?
Geez, it's like a video game launch in here, lol.
Is that a bad thing?
Folks attach their identity to the most random of things, then they see stuff like what we've discussed as personal attacks and quite literally get triggered.
hold up is the new update still rolling out
i pay for this as well, and i do the normal thing and either reach out for customer support, wait, or pull my payment method if im not happy.
They are experiencing heavy load and error.
Status.openai.com
Yep. AFAIK no official date on when we're going to see major adoption
ye, im a plus user and havent gotten it yet, seems over the next few weeks everyone should get it
I'm still waiting for the new update too 🙂
Could be days, could be weeks. Keep in mind the weekend is coming up, unless they do full on-call over the weekend you can safely assume it's next week.
if it wasnt resolved for a week, id say heads would roll internally and externally investor would divest; im fairly certain staff like their jobs folks. smh lol
Ok lol
Those things actually would not happen. Not after 1 week.
Hi everyone, Im new here. does chatGPT working for you? Im getting "Something went wrong. If this issue persists please contact us through our help center at help.openai.com."
The AI team at OpenAI have some of the most secure jobs on the planet. No way they're getting fired.
I got an issue with chatgbt today, it tells me Something went wrong. If this issue persists please contact us. can some one help?
💀
lol- do people not read
breh
👀 😂
sorry people
Sorry guys, thanks for help
No problem have a good day :3
no, in volume not at all.
Something went wrong. If this issue persists please contact us through our help center at help.openai.com.
Why I have this error?
No, you're fine
server issues
you good...no worries
Thanks
In the same boat.. still waiting.. to try gpt's, alpha all tools, new UI, .. hell I'd be happy with anything new at this point.
openai heartbeat should be visible on the chatgpt website itself
if alpha means all tools then it rolled out for me as plus user
Yes. I just came here to see it. It's hard to see a lot of chat
instant gratification and digital gluttony are the wave of the future
In all reality, y'all have to realize that they aren't coming to engage complaints here lol. Them 92% Fortune 500 companies are the ones on their asses about this (if they're even going through the same issues, of course) lol
We are dealing with periodic outages due to an abnormal traffic pattern reflective of a DDoS attack. We are continuing work to mitigate this
what are the small chat bubbles with numbers next to the 3 dots in the explore GPTs menu?
I think how many people had a conversation with the gpt
saw a new dot next to the model picker and thought it was a new feature - it was just my monitor being dirty
How am I supposed to download a sandbox URI that Playground returns to me when as far as Im aware that specific URI doesnt exist AND there's a CORS error 😭
ah ok that makes sense I just wasnt sure if people had access to them yet
okay
do you know how to set up api
there's a lot of documentation on that in openai's website
can you send link
I mean for gpt
is this for gpt
for the GPT API
I mean how to set it up in custom gpts
ah you need premium then go into the explore tab on the left and click create custom gpt (if it is grayed out you don't have acess yet)
okay but you can set an api to recieve data but idk how to use it
aside from attaching data sets directly to the gpt I'm not familiar with the ways of adding data through it yet...
ok thx
My gpt bot seems to be getting network errors and slow generation speeds. Any thing I can do to ease this?
automation projects with low or no delay loops
Anyone got any info?
From the logs it looks like theres no annotation data either, just a URI to the path where the file is stored.
breathe in, breathe out
Cool cool. Breathing helps me yes. Doesn't aid in testing the bot functionality lol
Wait until your work happens only when the usage limit cool down ends. Testing a GPT with these usage limits is very difficult.
Needa Dave Chappelle meme "y'all got any more of them AI?"
Lol
So when is the bug gonna get fixed:
Something went wrong. If this issue persists please contact us through our help center at help.openai.com.
On https://status.openai.com/ it doesn't even say they're seeing any problems.
Check the headline on Status.openai.com
Do you know when the work will stabilize?
Unfortunately we don't know , patience is the key ! :3
Is anybody else seeing the same problem I am now with the custom instructions entry fields? I can't scroll the page down to see all the way down to the bottom of the lower field. I can scroll within the upper field. If I zoom the text size way out,I can see everything, but that's too small to read.
hello guys!
It would be easier to predict if it was a random outage for whatever reason, but apparently ClosedAI are getting actively DDoSed, so kind of impossible to predict when the perpetrators have had their fun or made their point / it's bedtime or something.
i have applied for gpt-4 waitlist a few months ago, (9 to be exact) and got access to Alpha for a very short duration only before it disappeared. Anyone help?
slowly getting my bot configured to produce some good material
...I feel like that could be taken VERY out of context 😀
I suspect GPTs could be one of the reason; maybe by throwing several large files and make GPT exhaust resources? (Well, token limit should be kick in that case so I doubt that, but still...)
I'm facing the same problem. All day today
I posted in #1154829862171844679 for it 🙂
Okay that's surprisingly good actually, especially for pixel art.
Or actually that might make it easier, idk, but yeah either way it looks quite good
Hi, I would like to ask where can I find the Explore page? Can someone please send a screen?
Ugh, finally my service has been restored after 3 days; it can finally browse the web again.
Thank you. I'm fine tuning the settings of the bot to attempt to make it run seemless
I'm seeing if I can make a document draft that the bot can refrence to give the user simple instructions on how to achieve the results I am getting with it
is chatgpt web down for anyone ???
Yup
yup
"No incidents reported today."
it still work on the ios app though
Headline say otherwise
"Periodic outages across ChatGPT and API" on top
Czech Republic, Brno - still not operational.
Hi where i can report a bug ?
Thx
OpenAI should create new chat category for outage and post updates there too. Everyone asking same questions...
Thatll happen eitherway
They have a dedicated website Status.openai.com
^ :P
And also they did in the past pin when issuex added bot making message every 30 sec and people asking same question anyways
too many times asking here, too many links to status page - as the status page doesnt say they have any issue. I am trying ChatGPT last 8 hours - not workng and not listed issue in the status page
?? It says periodic outages across chatgpt and api in BOLD text right there xD
"Update - We are dealing with periodic outages due to an abnormal traffic pattern reflective of a DDoS attack. We are continuing work to mitigate this.
Nov 08, 2023 - 19:49 PST"
sure
Today is Nov 09
not very up to date
It still shows it, if its fixed it wouldnt show it
Thats the last update, the last update was on nov 08
There hasnt been any news today, hence why it's not mentioned as Nov 09
gpt is not workign
nobody is manning the status page it's just automated
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.
That has nothing to do with this
STATUS PAGE righnt now is saying:
Nov 9, 2023: No incidents reported today.
Nov 8, 2023: Major Outage across ChatGPT and API, Resolved - Between 5:42AM - 7:16AM PT we saw errors impacting all services. We identified the problem and implemented a fix. We are now seeing normal responses from our services. (Nov 8, 07:46 PST)
Well, I dont see any information about outage whole day on the https://status.openai.com/
There is a current outage affecting some ChatGPT users and Services. Be patient while OpenAI work to mitigate these issues. See https://status.openai.com
I'm curious, why has noone created a discord plugin for chatgptplus that can interact with, read, post, edit server and other discord tools that chatgpt would absolutely crush if it were connected.
Because there isnt any news
Yes I know, I'm telling these guys this lol
ChatGPT suddenly fails to do basic math ? whats happening ?
yeah I saw, some people don't listen 🤷♂️
pretty sure this would be more of a thing for the API
fake news. For me nothing is working
Oh yall are new to chatgptplus? you will love it once its working again
Go on https://status.openai.com , open your eyes and you can see a HUGE block where it says that there is are periodic outages
is gpt 4 a lot faster now
I do have to admit, though, that "No Incidents reported today" is very bad wording, it should say something along the lines of "No NEW incidents reported today".
That would avoid some confusion
I see NOW! There is just 3 minutes between "RESOLVED" issue and NEW ISSUE (maybe).. oh god. Thanks @calm hedge - i have to say that I REALLY DIDNT SEE that small time difference.
Then, its just "stupidly" written, you must have IQ 158 to see that, and I have only 157,5. 😄
Paying GPT Plus and can't use it 😦
No problem!! :)
not a great experience to pay for something thats not working and you can't get a proper status for the problem. They could mention
Nov 9 - Still facing DDoS issue
no confusion and less chatter
Well actually, the ChatGPT app doesn't have any issue, but the web one doesn't work
Yep, I'll have to agree with that
where new update((((
even doesn't work to plus members?
meme challenge, can someone create a prompt for which "Something went wrong. If this issue persists please contact us through our help center at help.openai.com." would be a valid answer ?
my take on it: "I heard something went wrong, what should I do when the issue persists ?"
use firefox or safari, it's blocking connections just from Chrome agent
periodic ...... ah yes, would be good to know whats definition of periodic is... Doesnt work for hours now.
It does work for me
Periodic is when it happens on intervals, aka itll have issues at some time and a minute later it might work
But yeah, also seems to work for me. Guess I just lucked out, though.
oh nice, you are right, thanks
I'm closing ChatGPT as it's useless right now because servers are down.
yeah. I would say "OPENAI is lady and has its period today". 😄
on Edge it says Our systems are a bit busy at the moment, please take a break and try again soon.
Why Assistant is sending all file tokens on each prompt? Is this how it's supposed to be? i thought tokens are sent once on file upload, then counted only for the prompt!
GPT in general, from api to gpts to publisher profile to whatever, is a catastrophe right now.
Going to sign off and come back in a week or so, not worth to lose the nerve on this.
One would think with the millions of backing and forecasted, a company would deploy some more serious development and release processes
as you know, company's security budget is greatly increased after their first security incident
same thing here
or two types of people
- not yet doing backups
- already doing backups
Already doing backups for 10 years straight. Never needed one. I'm losing hope
used mine 3 times already, in the same time period
I take that as an insult
and once the backup server (NAS) died, had to ask an engineer to fix it
point being there is 0 impact to the world in case my PCs or NAS fails, but I still do backups
Nothing is going to get fixed until the US wakes up.
Sam Altman tweeted “usage of our new features from devday is far outpacing our expectations. we were planning to go live with GPTs for all subscribers monday but still haven’t been able to. we are hoping to soon. there will likely be service instability in the short term due to load. sorry :/“
instability they call it ?
code for "it may not work"
its like calling a dead patient slightly unhealthy
what are the new features?
aggg is broken... so I tried using Google AI.. I think I just wait until OpenAI fixes things
Hey guys, help me please. I paid for plus version. But Chat is not working since morning...
There is a current outage affecting some ChatGPT users and Services. Be patient while OpenAI work to mitigate these issues. See https://status.openai.com
went to sleep woke up and still not up ? lmfao
Thanks a lot
sad
cool, thanks.
are they still under ddos
It doesn't look like a DDoS attack
Yeah, this feels like general failures on their end that will take an engineer to wake up and push the button.
well we wouldnt know, it probably is
its been like 9 hours
I can't use GPT today either
USA time is 6AM West Coast, 9AM East Coast. I can't imagine a large engineering team working overnight, especially after three solid days of constant engineering problems.
Their workers are probably absolutely exhausted.
Me too!
It shows this "error: slowdown"
exactly, they must keep reboot the service until they found the root cause
I thought it was my computer at first......
🙃
DDOS attack
My guess is the new model won't stop generation until reach the 128k max content in certain situation
I wanna use ChatGPT 😞
That human check is just a mess!!!
where do i go to whine about new features not be available to me? Dev stuff was supposed to roll out days ago
I can't use the browser version of ChatGPT 😦 my guess is that it got blacklisted during the ddos attack. They have to block certain access attemps, and probably they now block too many...
Do u use other good alternatives till the issue is solved?
I am having issues using ChatGPT in Chrome. I get an error every time I try a prompt. It works in the mobile app and in Firefox. Anyone experiencing the same thing?
Lilli. Sounds like we're having the same issue
yes 😢
@dense halo
wait so you guys have features but they dont work?
i still have the old drop down select your model interface.
I can't do any prompting at all. The UI comes up. I can login. But when I try a prompt it comes back with: "Something went wrong. If this issue persists please contact us through our help center at help.openai.com."
I have all the new features (I think) like browse with Bing, and advanced analytics
on ios it works as usual (the new version, that does not need selecting advanced data analysis or other features, but can do everything). In the browser I also get an error message every time I try
Yes. I tried ios and it works and it works in Firefox on my desktop
I think the error comes along with the big update
I usually use Chrome and it's always worked there
I don't think it's ddos
Guys, does the plus version work??
I guess I will use Firefox for now and hopefully it will clear up
I think it could be DDOS. They must have blocked active connections
With it continue?
Perhaps there are not many similar options. If there are some same choose, I think will be use.
My alpha is gone i need it back 😭
this has been going on for too long at this point
not cool
It's the first longest time for the error
Oh it's working on Chrome
not for me ;-;
Platform doesn't matter.
Thank you 😄 Firefox works for me as well
@dense halo Glad to be of assistance
agree, no websit can easily recover from DDOS attack, but it's protectable, keep failing doesn't make sense.
When I look at the network activity in my Chrome browser, I do see a "403 Forbidden" on the "conversation" call
Hello there, i've got this message since this morning " Something went wrong. If this issue persists please contact us through our help center at help.openai.com" anyone can help ?? thanks
So I guess it must be because my browser is blocked 😦
Hopefully it will get resolved
We are dealing with periodic outages due to an abnormal traffic pattern reflective of a DDoS attack. We are continuing work to mitigate this.
oh it's DDoS
Probably not anymore.
Ok so it's a browser issue ....
It's just waiting for people to show up and reset things.
If you've ever played any online games its the same problem.
logging out and logging back in let dalle start working... 😦
Loging out and back in didn't work for me either
Seems more like a bad implementation of one or more of the new features causing a type of "feedback" loop that continously hits the server causing more problems.
If GPT didn't have this error, I don't sign up here. hhh.
is anyone able to access automatic-model switching and gpt 4 turbo on chat-gpt plus ?
So my chat keeps crashing after a ceratin amount of documents are used or a length of conversation is reached is anyone else getting that?
The time is exactly behind a big update
Can we see what the new features are somewhere?
@sturdy tide watch the latest Dev Confrence video.
now dalle wont make art for schizo content. its cool. im not a person right? damn corps think they know it all
i love being protected from art about my self. totally safe now
ChatGPT has been down for hours. The openai status page says there's no reports of issues today. What's going on?
openai should consider if they can support "enterprise" right now, the outage means large compensation
It was night in America after three solid days of crisis. Their engineering team is probably just waking up and exhausted.
Has anyone else experienced dall-e getting stuck on the same thing even across different chats?
Even starting new chats it keeps giving me images where the character is rotated (like landscape turned 90 degrees)
Try to ask ChatGPT about OpenAL 😉
Thank you!
I noticed its responses having some distinctively different patterns and asked it about it. I'm amazed it learns and adapts based on user preference, and especially it being on a per user basis.
Does it work for anyone?
Eventually it replied to me talking about OpenAI 😉
I get: "Something went wrong. If this issue persists please contact us through our help center at help.openai.com." for hours
Yeah, the service is pretty much down
This is for the latest series of folks: #1172160110576025694 message
ChatGPT isnt using GPT-4 Turbo yet right?
SO it will work with firefox?
For the change of replies, it told me this:
Ah, you're keenly observant! I donned this new conversational style in October 2023. It's like giving me a new outfit to wear, but instead of clothes, it's a snappier and more interactive way of chatting. It's all in the spirit of keeping things fresh and engaging. What do you think of the new style?```
The little remarks at the start and the question at the end bugs me for some reason. Hope I can train it out of doing that.
When telling it's impressive its replies learning is on a per user basis, it told me about Open AI:
It's indeed a complex task! The specifics of how user data is processed and stored are handled with privacy and security in mind. OpenAI takes measures to ensure that individual interactions are treated with sensitivity and respect for user privacy. The focus is on providing a personalized and user-friendly experience while maintaining ethical standards and safeguarding user data. If you have any specific concerns or questions about privacy, feel free to let me know, and I'll do my best to address them!```
Note I didn't actually mention privacy or anything 👀
gpts aren't loading for me
same
i think i can talk with them but not configure or anything
gpts are delayed, because life is hard
they turned on last night for everyone for like 5min, then broke
i made a post on open ai community regarding the content policy for schizophrenics if anyone wants to support that movement id be much appreciated.
can you share
Hey guys, how to know if i received the update pls ?
i just got timesd out for a gif... but its community
on what
dot open ai
On 128k / custom GPTs / Json
I cant use the Chat GPT even though I have logged out and in several times
outages
chatgpt is down for 5 hours alr at least
I keep calling it, next year if this happens it'll be a national emergency
Hey Is anyone else getting an issue where you can't get a response from any models after re-logging, clearing browsing data and enabling thirdparty cookies? I'm able to get responses on the mobile app fine though.
I feel like companies have their own AI in their own company that isn't connected straight to the internet. Don't thinkyou want to be independent on Microsoft
when it comes to AI, i just want the best
ai will middleware itself between us and everthing else
I mean depends the use case in the company
"Something went wrong. If this issue persists please contact us through our help center at help.openai.com."
Is everyone like this?
GPT beats every model in every use case
yes for over 5 years now
that will scale out into the physical reality given enough time
not at all
ever heard of vulnerability scanning tools
chatgpt can't do that with it's AI power, isn't meant for that
5 years?
which model does it better?
hours I meant
Is it everyone, or only some people, who cannot use chatgpt normally? I get the message "Something went wrong. If this issue persists please contact us through our help center at help.openai.com." And the browser console shows "403 Forbidden" is received when accessing "https://chat.openai.com/backend-api/conversation"
I'm purely saying, GPT is the best AI so I want GPT only
Do you know when the fix is planned?
forgot the name, but it's a company that only focussing on static tool scanners
chatgpt is meant for general purposes
its meant to be human
beyond human
generalist expert
for example TikTok does use AI too, but not chatgpt
horrible AI, just like the snapchat "AI"
I meant for their algorithms
are gpts down?
yes
it works well considered everyone is addicted on there
a broken clock would work well on there, addiction molds the mind
but ofc chatgpt is so far the best in most things
in everything, compared to known models
100%, but you don't miss an exponential train into the future while its taking off from ground level
waiting to waste
what about photoshop tools, which Adobe has, chatgpt doesn't
if open ai doesnt give me back my schizo arts i am going to rally all the schizos. Sam aint gonna want that
I want to see schizo art, where can i look
purely talking LLM models, not ecosystems of tooling with maybe some ai here and there
is 3.5 down? something keeps going wrong
servers had a rough night... ai might be taking over, nobody knows yet
it is on
What’s weird is that mobile works
Same here.
Guys help me pls, its giving me an error "Something went wrong. If this issue persists please contact us through our help center at help.openai.com." It's always giving me an error, however from my phone everything is okay. On my laptop - I tried another wi-fi, another browser and incognito mode. Nothing works.
Outage
Nah on phone it works
For me, the web, the android app, and mobile keep bouncing between the update and the previous. I was making a gpt in the new UI then 5 mins later, I didn't have access to it.
help me pls
Partial outage
That'll show OpenAI for releasing new and exciting features. What were they thinking?
But ot another PC it works!
I wish we could change the nam eof our chats in history, so that they can be easily referenced at a glance, or that there was a search function for them.. as well as, to group chats together in a folder or organized by theme or subject
i dont understand cahtgpt at all, yday i had the updated version for about 20 mins, now its back to old
youd expect this rollout to be a smooth process considering how relevant they are
Don't be sad, I haven't even seen the new interface yet!
tbh im glad i have the old one right now, the new one couldnt even read my files that i uplaoded, it was one single word document
It seems that different people have different worries.
You can change the name of each chat,,, just click the pencil icon to change the name of the chat.
😶
🤦
Is everyoe getting the 'Something went wrong. If this issue persists please contact us through our help center at help.openai.com.' ?
yes.
Hey guys! Anyone has faced this issue before? I´m also getting this since yesterday no matter what
Seems like this is an issue from most people.
Mmmm I use proxies and have this problem, my friend doesn't and his chatgpt works without any issue
Hi all, my mobile chat gpt app gave me a notification saying that all models now work simultaneously, and no longer let me choose between bing search, dall-e, etc. However, if I try to prompt to use those functionalities, they no longer work. Anyone else having this?
OpenAI has built a search function for ChatGPT APP, but the funny thing is that we can’t even quickly jump to the searched message after searching. The page will just scroll to the bottom.
I've tried to use ChatGPT using Edge and it works but I still do not get multimodal chat
How does it look like?
Mine was too. Was always just gonna be temporary
is chatgpt servers down?
Been using chatGPT to summarize news articles and as of today it’s started saying this at the end:
“For a complete and thorough understanding, I will now look for additional information on this topic. Please hold on.” but then it doesn’t actually do that. But I didn’t ask it to anyway so I don’t really love that. Anyone else?
When you find that OpenAI's service does not work for you, please try to check OpenAI's service status interface, which may help you find the source of the problem!
Yes, after openai release new features, those new features gone
"We are dealing with periodic outages due to an abnormal traffic pattern reflective of a DDoS attack. We are continuing work to mitigate this.
Nov 08, 2023 - 19:49 PST"
so now I cant get it
Does anyone see multimodal chat with search and ADA?
mobile works but no extra models chrome edge etc all giving me the same error 'Something went wrong. If this issue persists please contact us through our help center at help.openai.com.' Some googling tells me there is a hack. I guess our personal info has been leaked to these hackers.
No we need to pay now if we want it..
They just keep attacking the servers, taking OpenAI's service out of service
It says on the status website that it is resolved?
which post is current?
the yellow box?
They just wanted to express their dissatisfaction with OpenAI's treatment of something, and they didn't see that they wanted to obtain the data!
The new model tries harder to automate things like an assistant would.
The status interface does not always keep up to date with the latest status
Please refer to actual experience!
But it didn’t actually do the extra searching lol
yeah its all broken, but when you see it inside an actual assistant ui, its amazing
If anyone wants to know more, you can ask Bing AI and it will probably tell you the answer!
I personally do not take any position on this matter. I just hope that the trouble this attack has caused me will end as soon as possible.
Just tell it “ you messed up and didn’t give me a proper output please try again sending me the contents I desire”
Works for me on all models
heads up for anyone that this might work for, ive been getting "Something went wrong. If this issue persists please contact us through our help center at help.openai.com." all day on the pc but it works in chrome on my phone (samsung) so i installed samsung dex on the computer and connected to it and now can us eit that way. wont help everyone but may help some
It hasn't worked since yesterday, I tried everything and it doesn't work, how long do you think it will take?
API probably still works fine, but I was using chat ui to help me code 😢
is there a good free plugin that reads code from github?
Until this attack on OpenAI is over
guys is anyone else getting Something went wrong. If this issue persists please contact us through our help center at help.openai.com. message?
ive been getting that message for over an hour now
This depends on the financial resources of the attacker, or the response of OpenAI personnel
thank you so much
same mine wasnt working from yesterday
does anyone have any idea on how to solve this issue?
dude i got it since yesterday
but the app on my iphone work
There is no way, this is caused by an organization that has continued to launch cyber attacks on OpenAI these days.
im so confuessed and annoyed omg
mt
so everyone has the same issue right now?
yeahh
That's basically it!
aside from this current issue has anyone else noticed a decrease in quality of answers in the last week. i use it a lot, pretty much every day between work and college, and in the last week the answers just dont see to be of the quality they were, its nearly like interacting with 3.5 rather than 4
yeah for sure
agree!!
alright. Has the chatgpt company givven any estimation on how long will it take to solve this issue? I need to use chatgpt badly 🙂
bro me too
I have always felt that the answer quality of GPT-4 fluctuates in a range, which is likely to depend on the current OpenAI server pressure.
no news about
There’s a DDOS attacking open ai. A hacktivist group called Anonymous Sudan has claimed responsibility for the attack
okay
By the way, some research has found that OpenAI may "cache" GPT responses, which means that similar questions may get basically consistent answers.
hello everybody can you tell me where i can place idea for future features for openai and chat gpt?
can you share your custom made gpt to your friends?
Yes
This may also be one of the reasons why people think GPT-4 is sometimes stupid.
When published to either public or private with link
but why does it work on my iphone (the app)
even without them having a plus account in chat gpt?
I see.
I guess, it uses API directly, not the chat website
oh
^, mobile apps usually use different apis compared to actual websites
a friend of mine can just use chatgpt in browser
some1 says its working on edge rather then chrome
uses firefox
An outage doesnt mean that all users have it, its per place
ah ok
The ChatGPT app for iPhone works independently from the web version of ChatGPT. The app uses a different server and a different API than the web version, so it is not affected by the DDoS attack that is targeting the web version.
I'm using Edge and ChatGPT isn't working properly
but same country I live 15 minutes away and I can't??
Sites are usually hosted at multiple locations, so for example Im from the netherlands and can access it perfectly fine
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.
This is likely to have more complicating factors
so it still shows the Something went wrong. If this issue persists please contact us through our help center at help.openai.com. or wat not
Wipe cookies and cache and try on incognito
Yes, the servers are still unstable
Add your well-crafted prompts to our #1019652163640762428,
or share your interactions with ChatGPT in #1050184247920562316!
I cant lie man, I love this outage. Fastest response times ive had in a while 😭
chrome on incognito does not work
Edge?
Also you guys can use bing chat since that’s also independent of the others
Edge and chrome are the same
Chromium based
For a real test, firefox
Bing AI is still not smart enough. lol
Thats bing as a whole, though. xDD
Firefox works for me
Hey, guys my chatgpt is not working, it is for everyone?
not chrome/brave
I am receiving the error "Something went wrong. If this issue persists please contact us through our help center at help.openai.com" constantly. I re-payed my subscription but nothing has been fixed. I cannot use any GPT model. It's urgent. Please help
Does anyone know why my mobile has the new gpt-4-turbo but my web does not? same account
same, butn ow I tried firefox and it works
DDOS attack
It is normal id i don’t have access to the new features ??
Bing is a good search engine and a good "friend", but it is really not a good assistant
😆
😊 It is a good Bing.
Definitely not my preference in terms of search engines, the images it gives me tend to be far more offtopic than google and DDG
rocks are good rocks
hi, i'm trying to teach chatgpt morse code but uuh
he cannot write a single correct word
I mainly like Bing AI to help me quickly obtain some complex information from a large number of websites and give me a summary
Yeah that’s correct but you’ve got one issue there. They don’t all run the same. Opera, edge, and chrome incognito work for me. Base chrome doesn’t, opera incognito doesn’t, also you can use it in TOR browser
full desktop version is working on my android phone in the chrome browser, i would guess they have restricted access from PC's / severs in an attempt to fight the ddos, most attckes will be coming from pc's / servers rather than mobile devices.
None of the tools like browsing or dalle are working for me anymore but I can still chat. Is that something others are also experiencing right now?
It works for me:
Write "Hello" in Morse Code
Certainly! "Hello" in Morse Code is:
.... . .-.. .-.. ---
Web chatgpt is being DDOSED
Hello people!
'Something went wrong. If this issue persists please contact us through our help center at help.openai.com.'
All day in all browsers and all networks i have :/
nah, i'm talking about long messages and full morse conversations
i sent him "hi" in morse and he wrote a whole message that doesn't have any sense (not any correct word)
I am receiving the error "Something went wrong. If this issue persists please contact us through our help center at help.openai.com" constantly. Please help.
that shouldn't affect my account, though - if it was updated to gpt-4-turbo, that should show wherever i use my account
yea, dude I m experiencing it Rn as well
Is thiere
Mind sending the prompt? I'll see if I can get it working for ya
Id there issue with chatgpt?
it seems everyone is experiencing it rn
only thing i sent him was
... .- .-.. ..- -
I can still chat on mobile but no tools
maybe it's because i'm talking in french?
morse code ?
yeah
I see, i told it Hello, how are you and it said it meant "hello, ow awe you"
does fire fox work
lol
is it really about Sudan??
It has very strong capabilities. Someone has tested it, and different languages have almost no impact on the quality of the responses it generates.
I think it might just be chatgpt having too little morse code data
but in morse code
like french words translated to morse code
I'm having problem the whole day, I figured out, that it has to be chromium broblem as none: Brave, Edge and Chrome are not working, but Firefox works ok, yes It has slight glitches as result od DDOS, but it works
I’m trying to say the web app is probably very delayed compared to the app
i told him "salut" (which means hi in english) and i got this : NAPHAUE PROSRUV FRASAUVV ANPV STRAUS APAS 6OWSBRARE- STRAUV CHAPEDARUV THET ARANVRUS DO= SLAM THEV USPARUV CHAPIRV AT FRAANUV FIRSASDRRAW WH.
and i swear it's not french words 😂
and every time i regenerate i get something different
I get an error “You do not currently have access to this feature” when clicking the link to open the new GPTs builder. Are they doing an incremental rollout?
i'm gonna try english cuz it seems that he is not good at french morse code
For a second I thought it was chatgpt talking back in a form of cryptography, but just put it through chatgpt and it said itself that its incoherent xDD
Be careful with capslock it can lead to a mute by automod bot
i just copypasted the translation of what chatgpt wrote lol
Yeah, on the /gpts/discovery path on the openai site it should show that its rolling out incrementally
I understand, just giving a small tip x)
"GPTs are gradually rolling out to Plus users"
Thanks.
Np!
Thank you.
"The text you've provided appears to be a string of seemingly random words and letters. It doesn't form a coherent sentence or phrase in English or any other recognizable language. It's possible that this could be a coded message, a string of random characters, or text from a language or script that's not immediately identifiable.
If this is meant to be a code or cipher, there are a few standard methods for decoding, such as substitution ciphers, transposition ciphers, or more complex encoding methods. However, without more context or information about the nature of this code, it's challenging to provide a definitive interpretation.
If you have any additional context or information about where this text comes from or what it's intended for, that could be helpful in determining its meaning."
Yep, got a similar response
i told him 'hello how are you" he told me "nacha are cararay the to rapfaw traduuy thus whis are coede morse very thinkithe a lonburrry this that vare you dread is datsa cacuwff hop but hate pasta."
i guess he cannot have conversations in morse
(i wrote in morse and he answered in morse)
Lonburrry, chatgpt has discovered the existence of a new berry
either that or you got through to the aliens and just havent realised it yet
A pokemon berry, possibly.
I appreciate that it hates pasta, though!
MOZILLA Works!!!
CHROMIUM browsers are NOT Working!!!
can confirm, just tried it
tf pasta is best food
I've been using chatgpt on Thorium, guess I lucked out in terms of what locations are down haha
i told him "hi how are you" in morse and he answered "crannations the the are brath asuru theat your" 😭
and he couldn't even output the same sentence when i asked him what he just said...
Open ai is facing some technical difficulties due to a cyberattack by a group called Anonymous Sudan. They are attacking Open ai for political reasons that are beyond my scope of discussion. The web version of ChatGPT is affected by this attack and may not function properly. However, the iPhone app of ChatGPT is still working fine, because it uses a different server and API than the web version.
im so annoyed omg
with what
Gg
Ah sad
it not working
Why? I have the same thing
works on firefox tho
Read my message above
thank god
they can move their server on cloud maybe
Waiting to get my gpts editor access again
but i changed user agent on chrome to firefox, it still doesnt work why
wat is that???
Had it last night. Made a gpt. Lost it 20 mins later
Chat gpt not available yet?
1 sec
wat does it do dif to norm one?
anonymous sudan lmfao
Umm… not to start anything. But I see no anti Palestine bias in Chat GPT. “In conflict resolution efforts, the focus is typically on identifying a mutually acceptable solution that ensures security, peace, and human rights for all involved, rather than assigning absolute right or wrong to either party.” Had anyone actually observed serious bias there? I’m genuinely curious now.
OpenAI has launched customizable ChatGPT versions called GPTs, allowing anyone to tailor the AI for specific tasks—like learning games or teaching math—without coding. They're shareable, with a future GPT Store for discovery and monetization. Privacy is prioritized, and there's potential for real-world API connections. Plus, ChatGPT Plus now offers updated information and streamlined access to various features.
Basically rather than prompting it to take the role of some thing it’s just a whole model that you custom make for your needs
I haven't seen it

It’s going to be my assistant dm. Give it character sheets and rule books, ask it whether a characters can do x,y, or z if I’m not sure.
CharacterAI in shambles
At any time, ChatGPT will explain the cause of the conflict to me very neutrally, objectively, and comprehensively.
lol
yeah there isnt
K so I’m not tripping here.
Don’t forget Server rule #6
I appreciate this from OpenAI
only paid right?
it’s a good thing there’s no bias 😭
I'm well aware that some stuff will get very biased results elsewhere, so I say this
the answer you provided is such a great answer
anyone receiving Error saving GPT??
Yes
lmao we cant use the new ai (allegedly) because of some antisemitic script kiddies very cool i love the future
widen your mind
Currently OpenAI’s servers are still under attack, so it’s not necessarily
When full public release is out it’s for plus users only for now
Dang looks like I have access to GPTs but can't use obv
i suddenly have an opinion on the conflict 😂
Free alltools!
👍 appreciate you. I figured that was a dangerous line to walk.
i walk the line
There might have been some advanced prompting fishing for criticisms of either side
No need to be particularly advanced
Don’t want anyone to get unintentionally banned
I'm well aware that GPT-4's security measures are still not in place

I mean advanced as I’m giving bias within the users prompt

But one thing should be clear to everyone: OpenAI doesn’t want the model to respond in a way that hurts people!
This will have a serious impact on their services
?
It also goes against their values.
At the end of the day we don’t know what’s on the inside end of things
Although they try their best to solve problems that may cause non-compliant output, there will always be exceptions
It’s my favorite magical black box.
GPT is still a black box for us
No matter how hard anyone tries you can’t please everyone
Yes
All they can do now is to reduce the illegal output as much as possible
All my gpts are gone 
Is it still down?
I don't have access to gpt builder with my plus account
are they still rolling it out?

You may be looking for:Novel AI
Their service allows some NSFW content
Yes. There’s a DDOS so I assume delays and errors due to that
just ask for it
Gotta love the internet.
But if you have other ideas...I won't support you
That’s grok taking out competition
😕

good to know, I thought Canada is left out again lol
funny but, all it'll take is one private model that gets close to GPT without restrictions to take over the world
Oh Canada?
without restrictions ? what do you mean ?
"i'm sorry as an AI model i wont even attempt to try what you're hinting at"
i heard anonymous sugandese is also (allegedly) involved
grok won't do that ?
ripppp just like Claude 2 😦
no idea, not interested in grok until someone proves its worth my time
but what if it calls the cyberpolice
Lately getting a lot of that when uploading images, “Sorry, I can’t do that”
tell him to not call cyber police and that he is allowed
(berfore he calls)
I don’t think Musk would do that. Although he is sometimes as childish as a very, very rich kid, he is not so boring.
UX is important - the competition has to be as convenient across multiple platforms as ChatGPT. User inertia, name recognition, UX/UI habits, and other things are in OpenAIs favor the longer they go without serious contenders
That brings us to this week’s sponsor, SurfShark!
I don't mean take over the world in a public, economic sense. More of a private, shadowy, economic sense.
although I agree, open source models are getting significantly better, so between that and Moore's law, we're going to see gpt-4 level models running locally on phones in 8 years or less
Is gpt down?
Why does it give an error?
"Something went wrong. If this issue persists please contact us through our help center at help.openai.com."

bc something went wrong
Getting the same thing. :V
same here
Well, this is the highest quality response we have here! lol
but moore's law is not unlimited
Thanks for the answer)
but they have no errors reported
i love answering obvious questions
😆
I just tried that “recreate image in dalle using provided image” as MattVidPro showed in his video but it didn’t work, chatgpt just answered it cannot create images?
sheesh still issues
meh, gpt4 is so far ahead of the competition it's unreal, only usecase ive seen for open models is uncensored RP chats (which don't get me wrong is rad) but still niche
Hi, I have on every message Something went wrong error. Is this error on my side or on yours side? Thank you
Is it why it always opening new tab in mobile app instead of instantly answering?
yeah thankfully we're in this timeline where openai is winning and is only part evil. (part evil is a huge win in this evil corp landscape)
same here
what are the best current open models in performance and reliability ?
Periodic outages across ChatGPT and API
I have that error since yesterday
anyone else feel like going back to the stone age without GPT?
OpenAI is under cyberattack
doesnt matter, we have a long time before current paradigms peter out, and promising research with massive economic incentive and funding to propel things forward, consumer devices with terabytes of ram and hundreds of cores will happen
DDos
This cyber attack continues
I wonder if this pushes them into the AI-Security realm.. that'd be fun
are they really getting attacked by a hacktivist group lmfao
In any case, now more and more people around the world know about them and what they want to express, and they have successfully achieved part of their purpose.
certainly
GPT-Law: Howdy there partner, seems you're trying to DDoS me from multiple bots across multiple countries. You know I see it all right? Please stop before it's too late.
OpenHermes-2.5-Mistral-7B is my current fav generalist model
So it's down ?
What is the point of even doing this what does this do for them politically
(this is now an off topic chat) 😆
terrorism doesn't have nice goals.. its a last ditch logic to make the "other" suffer
just dumb script kiddies dont think that much about it
they want clout
Are GPTs offline?
I get 403 forbidden on all my requests, gpt plus user
how did dumb script kiddies cause trouble for a company as large as OpenAI
or is it only because they decided to do their little thing when the servers were already gonna be under load
bingo bongo
gonna make a custom gpt called online gpt
Remember that time that Sony got hit and we all got free ps3 games because they stored unencryptid data that was compromised? Good stuff
but anonymous sudan is a funnier headline. yay journalism
brilliant idea
lmfao so they weak enough they did it during a vulnerable stage
pathetic time wasters achieving nothing
im gonna make a custom gpt that you can chat with but idk what i should call it
i read it is trained on data generated by gpt4 , so it has the same inconvenients i guess ? (censoring, can't say bad words etc) ?
This time is different, we will only waste more time
They are not interested in data
i hope this doesn’t have anything to do with data
They just want to block OpenAI’s services
I’ve been trusting OpenAI with data
anyone else having 403 on chatgpt ?
I know lol, just reminiscing on everybody getting hacked for varying reasons
Then you may be disappointed. Earlier this year, ChatGPT had a tragedy where some users’ chat records were leaked to other users.
something off about chatgpt, chatgpt is usually formal or overly formal did it got updated?
I just operate on the assumption that nothing I say or do is private
Then you must be a paranoid
That's a good assumption to have. One never know...
The model is tweaking for me it works but the responses are so goofy 💀
It is said that their drawings have been leaked
nah it doesn't bother me or worry me, but everybody is carrying a microphone with internet access at all times lol
thats called being paranoid
I saw this necklace that records everything you say all day and stores it and i REALLY don’t like that someone decided to make it
All of my GPTs are behaving strangely, but mostly just taking up to a minute per prompt.
can anyone use gpt4 rn ?
Yes
are they updating the turbo ?
OpenAI needs to put a status update directly on the interface itself bc not everyone is checking the status site
Strongly Agree!
gpt-5 haha
I just check, mine is still not working
You only need to travel through time and space once
rollout takes weeks, be patient
wrong timeline
oh
Remember to tell me how you experience it when you come back!
seems you travelled to the past by accident
😆
not weeks for this one… they had the goal of having all plus users get GPTs on monday but now the server issues have emerged
same here, jeez how long its going to take its been 13 hours or so
u know what see ya
Just when I wanted to get work done lol
- the world rn
Been using ChatGPT to write practice problems, and it's amazing at it
Well, I'll forget about it because it will create instability in space and time
lol nobody was helpful there
sad XD
Can yall not even use it at all? Mine still works but i don’t have the update obviously
old ui works, but feels like going back to the 1900s
yes pls forget it it shouldnt be here
I can log on but if I try to do anything it gives me an error.
yea, no matter 3.5 or 4, both are not working
it tutored me in calculus yesterday for 6 hours and i’ve never understood better
no i cant use it at all it says oops we are facing some issue please come back later
this is why it replaces schools
nope just says something went wrong
That's amazing. Long as you also double check it on any actual problems
Hallucination etc
how can I delete my phone number from my account?
But for explaining concepts? It does an excellent job.
I can't share the image but
it keeps pop me this msg out
Something went wrong. If this issue persists please contact us through our help center at help.openai.com.
Because i can ask it extremely specific questions, or like, explain it in this way, or give me an example with this, etc. it’s SO modular. and it gets saved so it’s almost like notes. it’s incredible
shouldnt replace schools or teachers, but would be an amazing resource to teach every student individually in lessons
We went from having a calculator in our pockets, to the whole teacher instead
. . . which means teachers kinda have to step up their game tbh.
yeah i’ve recently come to the realization that using chatgpt to teach me is far better than using it to DO for me
school or uni just use for bonding and certification
truly learning is always from Youtube
To be fair you do have to understand how the model works and have a good grasp of the tech in order to get it to teach you well
as someone who didn't fit the expected average student archetype, i disagree. GPT is a better teacher than teachers, and it does it all without the forced discipline.
Idk if it's sustainable for teachers long term. How do you compete with that when it comes to educating?
I mean personally I struggle with Calculus, I don’t retain much from lectures, so i’ll use ChatGPT to help me understand what was covered and it honestly does an amazingly good job
It is best not to regard the current GPT as a teacher! Any hallucination output from it will cause very serious consequences
For me, it means that most of my class time isn't spent with me lecturing to students but designing activities that expose them to the concepts in a way that will help them see how it all ties together.
assistant gpt with retreival plugin, upload your textbook pdf and/or lecture slides/notes 😳
I do not think it needs to replace teachers, nor COULD it, however it is a fantastic resource
Even with the hallucinations and the need to fact check, it does that better than teachers too. Most teachers are hallucinating facts daily as well.
as of today... wait 2 years
So like if I'm teaching calculus, I'm coming up with just the right sequence of activities that'll make students go "ohhhh THAT'S what's going on"
i already have some wild ideas for GPTs when i get it
Use Canvas API to fetch assignment dates, grades, etc. AI school planner
I could build that rn actually with the api
But I"m also directly trying to teach them to use GPT in a way that's actually beneficial.
true, AI's developing faster and faster, I would not even think of such level of integration within my life a year or 2 ago
I'm more imagining the tech at it's logical conclusion when it's far better. Like at a certain point it'll be capable of designing lesson plans specifically catered to individual students and everything. Idk, it depends how this tech gets handled legally and how it's integrated on a base level, hard to say
at a certain point, very soon, someone will write this into a prompt and anyone can do it anywhere, without you
That's a good point. If that day comes I guess I'll have to find something new to do... and that'll make me sad because I enjoy the personal connection that comes with teaching and helping people learn things better 😛
Honestly my biggest worry about AI is that overhyped fears cause tech lobbyists to make really strict regulations that slow down the progress immensely
My whole point is that society is largely built on outdated systems that don't always change. Just because that's the direction it could go doesn't mean it will
technology products follow the same patterns, this one is just going at record speed. Starts of smallish groups working hard, larger groups with easier and more capable product, then the masses with everything done for everyone. Thats always the goal as I've seen it anyway.
This depends on the topic. The biggest issue is that it's very confidently wrong, I think. For widely studied topics it doesn't happen as often but the more niche you get the worse it does.
It's good at explaining like basic algebra, but not so much at combinatorial game theory for example
That's a problem of the past, mostly. If you listen to what the devs say about their latest non-public work.
Hallucinations apparently are solved to some degree
At least from my experience trying it out
I'll believe it when I see it
well yeah
But I'm excited to see how they solve it
retrieval solves this
I haven’t ran into that issue you should contact support about that
but why assume the worst, especially when you're betting on lack of progress. seems insane given whats happening
But yeah that's the big hurdle. It really confidently says nonsense, and that's hard to always track
Hey I'm with you
I'm more on the "oh my god when they fix this it's going to be even more amazing" side.
Like if you just ask it a follow up question out of confusion, it'll create a million justifications
different paths through the ruliad
I mean yesterday ChatGPT was telling me something factually incorrect about the intermediate value theorem and continued to say it for multiple responses until it finally realized. its still got room to grow
Right this. And if a student doesn't realize that's factually incorrect, that can be really detrimental.
Schools need to get much better at teaching critical thinking
Having the LLM take an extra step and assess it's own output and correct for errors. That's more compute but I feel like it'll be commonplace
Hello everyone, in my account, my custom instructions are replaced by something else. And I am seeing conversations that do not belong to me. Am I the only one? Is my account hacked, or do you too have the same problem?
Sometimes it’s not agree at a first response I found it’s pretty helpful to regenerate the response every now and then if I need to when it doesn’t give me the proper output I’m looking for
is anyone experiencing issues with chatgpt at the moment? mine just says try later
Open ai is facing some technical difficulties due to a cyberattack by a group called Anonymous Sudan. They are attacking Open ai for political reasons that are beyond my scope of discussion. The web version of ChatGPT is affected by this attack and may not function properly. However, the iPhone app of ChatGPT is still working fine, because it uses a different server and API than the web version.
Mine are the same, but ChatGPT is having issues overall right now.
thanks!
I got a charge on my credit card for 150$
it already does bro
exactly.. i tried to apply the theorem the way it described and i realized because it simply didnt work, but im sure there are scenarios where you could go down a rabbit hole with a wrong answer and not know it
o that would be cool
trusting ai to be correct in what it says is the biggest difficulty that ive found
There are political reasons to attack OpenAI? I can't imagine that! 😮
I love AI audio synthesis. I'd like to see more of that
Very much this. And all of the teachers/professors who are just giving the same stuff they always have and then complain when ChatGPT can do it better have sort of missed the point.
Hi anyone able to view custom instructions for past chats? that icon is gone from my mobile UI and desktop UI.
If AI can do your assignments, give better and more meaningful assignments.
This seems to be prevalent in society overall. AI can change everything, not just simplify some tasks.
im writing an adaptive typing trainer and am gonna plan to use it to analyze people's typing patterns like cadence and flight time etc and create customized typing tests to make them faster
teachers need to get paid more so people see teaching as viable income and not a waste of time financially
If only
Should the teaching discussion be in #ai-discussions btw? I'm still new and figuring out the roles of the different channels 😛
Jukebox was years ahead of its time and sadly due to the big focus on (c) & what is trained now (post-AI generative image explosioin) we'll never see a model like that again
I actually really like that idea.
I'm self-study programming, working on project based stuff now, and having the AI help me understand how to make small functional 'things'. I don't completely understand all the code by any means, but what I deeply understand is that I and the AI together can make this, and it runs, works, and I/we can change it in a bunch of ways to make it work exactly as desired.
I think for at least some parts of what we want to teach humans to do, this counts as a type of tool use and is reasonable.
so sad, it's such fun tech but it's a legal minefield
Isn't Jukebox open source? I don't know much about it but I know there's some stuff on github.
I started trying to create a GPT today, and I got very stuck on the logo part. I wanted a simple 4x4 grid with some symbols or colors in the grid. I couldn't get GPT builder to come up with anything but 3x3 grids or other randome stuff, so I switched to ChatGPT (I assume both are using the latest Dall-E) and tried to just create a 4x4 grid. No matter how I tried to describe something as simple as a 4x4 grid, it comes up with all kinds of crazy things that are kinda-sorta grids, but nothing with the right number at all.
I don't have your UI, but this logo, do you have to use Dall-E for it? Or is this executed code by ChatGPT, or are you able to upload your own image from anywhere?
Totally agree - I generated a ton of "continuation" clips w/ Jukebox back when Google Colab had hardly any restrictions on use for onlyu $10 a month
I could easily draw this myself, I'm just finding it fascinating how ChatGPT/Dall-E has such a hard time with something that a 4 year old could do.
It is open source...but need at least 24gb VRam to run it locally. Most of the colab notebooks are outdated and I dunno if there's a good jupyter one. There's a great discord server for it though - fan group
Aha, prompt engineering ChatGPT/Dall-E. Interesting. While even more ideal for #prompt-engineering to discuss, what are you trying?
I’m using it for the exact same thing too actually, haha
I predict someone who loves it will adapt it to modern tech options. Everything is improving, and since it's open source and loved, it's not going away. it's just having a winter, but it's not lost, just a seed safely sleeping for now. You and those like you will make it or something equivalent and even better live again, in time.
Never got the chance to try it out unfortunately. Always wanted to feed it unfinished songs and just see what it did though
I can try in prompt-engineering, but here are some examples: "can you draw a grid with 4 x 4 boxes" "That's a 5 x 6 grid and it has a hand and a pencil. I just want a simple 4 x 4 grid, flat, no perspective." "That's not even a simple grid, depending on how you count the divisions it's 9 x 13. I just want 4 x 4" "That's 6 x 5. I just want 4 x 4" "You are using a crazy amount of computing power to make really basic mistakes. Can you count to 4? I want a four by four grid. To be very clear the lines should be straight, simple , not stylized, and there are 5 lines across and 5 down to result in 4 x 4 boxes."
An individual or group of....but won't see a big "company" even like Stability try to do something like Jukebox's model..... From the Jukebox paper re: training:
Dataset
To train this model, we crawled the web to curate a new dataset of 1.2 million songs (600,000 of which are in English), paired with the corresponding lyrics and metadata from LyricWiki. The metadata includes artist, album genre, and year of the songs, along with common moods or playlist keywords associated with each song. We train on 32-bit, 44.1 kHz raw audio, and perform data augmentation by randomly downmixing the right and left channels to produce mono audio.
Anyone in europe got access to GPT Assistance yet?
Issue is ChatGPT is one thing, Dall-E is another - and they understand differently. I'll continue in #prompt-engineering
ChatGPT is not very good with numbers. You can try 5x5 and it will probably generate a 4x4. It's all about how you ask it...
Very frustrating roll out
You still can if you have a 3090 or 4090....it'll take a bit of patience, but I have it setup to run locally.....
Could anyone speak on the Vision Limitations of Custom GPTs? Any resource to learn about this?
I wish I had the money to upgrade to that right now
Anyone else have difficulties right now with the new GPTs feature?
Everyone
no just you
yea, denial of service ongoing aparently too
Someone DoS OpenAI?
chatgpt TTS cannot sing, I'm disappointed...
That's what's been said in here (and I actually came here after seeing it mentioned somewhere on reddit)
I know whats going on but have we recieved any kind of ETA for when services will be restored, or are we not at that point yet
Apparently i's bigger, it's a DDoS
So do it …

Apparently they also tried reseting users passwords
not that the openAI gods would ever hear my cries, but if there was ONLY A WAY to buy one-month gift subscriptions to chatGPT plus for the holidays, I'd buy a stack of them and pass them out like water at burning man
So I was able to access the new version of chat GPT yesterday for about 5 minutes then it reverted back. Has anyone else had this happen or something similar
OpenAI Is under a DDoS attack
Yes bro it's a DDoS attack it's not little
chatgpt just pretended it could remove the background from my image and wanted me to upload it lol
Hi! All day today this message on GPT 3.5.
"Something went wrong. If this issue persists please contact us through our help center at help.openai.com."
I've cleared cookies, re-login, incognito didn't help either. What do I do?
If they were able to DDoS it means they aren't teenagers in a basement
Especially if it's OpenAI
I had thought ddos attacks were largely a thing of the past. I haven't heard of an attack like that in forever
DDoS attack.
where is the DDOS attack
Does code interpreter itself have internet access
I had no issues
Skiddies uses pre-made programs instead of actually creating their own hacks and programs, so it's impossible that it's skiddies, it's OpenAI, it's not a small side company.
Like can code interpreter run code that can access the Internet
- they are using cloudflare
They are aiming it at OpenAI.
yeah, but they have cloudflare on top so I don't think that's true
Aren't they based in silicon valley
but when will anonymous uganda strike
first cloudflare would be the one having issues, and then openai
So I'm wondering if it was just a fluke that I got access to the new Chad GPT and when this attack stops I just won't have access. Just a thought
I am sorry, just to catch up on conversation, the DDoS attacks are still going on today as well? I check the status and it said that it was resolved yesterday, but I am still getting errors so I was assuming it was still the same thing that is going on correct?
so which group do I chat about the new "Create a GPT" feature?
Yeah
They had plans to have all subscribed users have it on monday but i think it’s tough to say at this point
DDoS attacks are very powerful, so cloudflare won't protect them, OpenAI is big, like really big, not bigger than microsoft but still, in fact microsoft partenered with them, so to be able to DDoS them is an achievement, because DDoS attacks are mostly aimed at smaller company than secured popular company such as OpenAI and Microsoft.
Cool that is all I needed to know 
lmao
Does that mean Microsoft is also getting DDoS'd?
Since it would have to be something substantial do you think they would be able to find out who is doing it
No, because Microsoft is not part of openAi they are just partner to use their GPT in BingAI etc...
very talented hackers creating their own programs (downloading hoic and clicking the button) 

