Hey. I, on the other hand, get constant “Network Error”. I then only have the option to regenerate the response, and each time I do, it’s using a token.
Some days, I effectively get 3-5 responses and that’s it. Really frustrating that network errors are counted as a used token!
1 Like
The cap has been reduced to 20 per customGPT without telling anyone. Shady business that they didn’t even notify anyone. Same sleazy tactic as when they introduced a Plus version and dumbed down the Free one in 2023. Cannot forget this sleazy tactic. They keep doing it. And I’m guessing it works.
3 Likes
According to the ChatterClock ChromeWebStore page, the team plan has a ChatGPT4 message cap of 100 messages / 3 hours. Pretty impressive, I’d say!
Is the 100 messages / 3 hours ChatGPT4 message cap applied per user or per team?
I have the same problem and it pissed me off. I will not continue with this 4 version. Limiting me after 10 15 messages… Nah.
Did Teams also recently get really stupid and at the same time remove the continue button so you can continue generating code?
I’m trying to figure out if Teams will solve the issue I’m having or if I’d just end up giving them 3 times the money for the same headache.
From my 2-month experience with Teams, I don’t see a significant difference between the Plus plan and the Teams plan.
The main differences are that the message limit for the Plus plan is about twice as much, you can use two accounts (if you subscribe to two sheets), and your conversation history won’t be used for model training by default.
In my perception, text generation seemed slightly faster with the Teams plan, but the language model used and the context length are the same.
I received an email about GPT sharing, but I don’t think it’s relevant here.
I believe you can grasp what this implies.
Yeah, thanks.
So just to confirm, they took the continue button away to continue generating code on teams too?
I recently purchased the team workplace plan. I spent $600 and, although I am not entirely sure, I believe I didn’t receive a confirmation message for this transaction. I think you are to get 140 thousand tokens but I’m not sure . Nevertheless, I’ve been using it all day without encountering any problems.
However, using ChatGPT Plus has been very frustrating for me. It times out after 40 tokens, which doesn’t really allow for experimentation. Despite advertising on this platform that I was willing to share the other side of the teamspace, no one responded. This is okay because I understand that everyone is busy.
I’ve seen people complain about how frustrating it is to cap out at 40 tokens. But if they looked into how the team workspace functions, they would see that it’s a great deal. It amounts to $300 per person per year if you can get two people, which is much cheaper than the individual plan that costs $720 a year. Plus, it’s very private. You don’t even know about the other side of the workspace unless you voluntarily decide to share what you have. So, there you have it.
1 Like
I remember seeing a “continue” button to continue generating code in the teams plan.
Recently, like within the past few days?
I took a break for a few days where I wasn’t using it much, and then resumed working a few days ago, that’s when I noticed the continue generating button had vanished.
That also seems to be when the team assaulted the model with a nerf bat. So if Teams still has the continue generating button now, that’s an indicator that teams might not be impacted the same way as the Plus model.
It’s challenging, but I looked through my past screenshots.
ChatGPT responses are designed to fit within a single response, so the “continue” button rarely appears.
The last time I saw it was 8 days ago.
I didn’t renew my subscription because I personally didn’t think it was worth it, and now I don’t have access to the teams plan.
So, I have no way to confirm it now.
Thanks for trying anyway.
For me and the work I do, the continue button was an every day many times a day thing.
Working mostly with code, ChatGPT’s code output becomes exponentially unreliable when it is fragmented, so when you have it modify code, if you don’t have it modify the whole thing it mostly just breaks the parts if they have any sort of integration requirement.
So myself, I probably saw the continue button 20-50+ times a day.
I could usually tell just by looking at the markers when it would have to start filtering out the useless chit chat in responses where it tries to explain the obvious to me or tell me how it’s going to do what I just told it to do and how to do it like I don’t know, and just output the code.
I complained about it here on these forums, but I might have been the first, because I’ve never seen anyone else talk about how the continue button only worked once. I’ve often had to generate responses that went beyond even 1 continue, but the second time you pressed it, it’s a 100% guaranteed network failure, I think I’ve even posted screenshots of it.
But the trick was you could say “continue from…” before and it would usually continue, and I’ve even had a few times where it continued once (with the button), then I had to continue manually, then it continued again with a button. But that was a couple of months ago when I wasn’t on here complaining because everything was running great so I didn’t have a care in the world.
Until this week, I have not had a problem with it refusing instructions and condensing its output to fit in one reply since the last time I posted complaining about it some time back.
I’m testing some new customization instructions, I want to see how it impacts things with this current state, but largely the problems I’m having right now it’s been a while since I’ve had them.
If I thought for a moment Teams would give me 2 weeks ago ChatGPT-4, I’d pay the $60/month and use both accounts myself or have a friend sign on for the 2nd one.
I put in an application for Enterprise again but my team is small, we’re not nearly rich enough for OpenAI to acknowledge we exist, let alone respond to a sales inquiry.
As far as the image, 8 days ago I had the continue generating button to. It’s only been this last week that things have changed for me.
I guess maybe I should update this and see if I can figure out how ChatGPT locally condenses/refines context to reduce the burden on the model.
That’s the thing OpenAI has done so good with ChatGPT that makes it, IMO, better than any of the other models. A normal chat model on many platforms, and including my own, mostly just sends your history with each message.
It has to be more than that with ChatGPT, because I’ve had conversations so long that there’s no way I’m not vastly past the point where it could do that, so it seems to me that they have some way of refining relevant context.
When they did their last good update, before they beat ChatGPT with a nerf bat, that last context update was amazing. I’ve had conversations so long that the token counter would start locking up, but it did great at remembering what we were doing.
The way ChatGPT-4 remembers context and keeps the conversation once that conversation has a decent size history blows anything I’ve been able to reproduce through the API or any other site using their API away. I’ve had conversation histories recently that were so big I can’t even submit the whole conversation to GPT-4 Turbo because I get an exceeds limit error. I’ve NEVER had any conversation, no matter how long, even when it started to get amnesia, get rejected in ChatGPT, so they must have some sort of localized filter for what they submit to the model with your prompt.
I really don’t understand why I am paying for something that is giving me limitations. Give me at least a chance to pay so I don’t have limitations without having to go to an additional tier or sign some crazy paper or anything like that. I just tried to use the ChatGPT Agent on my phone And mid conversation it just stops. Then I can only wait before I’m able to use it. That seems like the most unproductive that I can think of. If somebody is paying for something and they want more of it, why not just give it to them and ask for the money too.
When you say “100 dialogs,” do you mean separate chats including perhaps dozens of individual messages? Or are you referring to a dialog as each time I hit the enter key sending a message to the chat box? Thx for the clarification!
The cap is as it is, but the STUPID user interface makes users angry. Why?
- There should be information CLEARLY visible how much messages is left till you reach the cap! All the time!
- The messages not fullfilled (ie network errors) should not qualify for cap.
- The cap mechanism and internals must be open - we shouldn’t be surprised, I assume, that it even qualifies for class action suit in the US.
- There should be an option to temporarily increase the cap.
- I feel cheated if I pay for a month of access, and then I use chatgpt for maybe a few days. So, what about all those hours where I could use my 20-40 messages per hour and they are not cumulated into my monthly cap? It feels like a scam, because I do not pay for an hourly limited access but for monthly access!
1 Like
You need to upgrade to the team plan i left plus and i now never get any of those reach the limit message. It’s $600 for the year but for me it was the best option
What happens, if say, you ask your brother together with you join the team plan, you pay all the money and control his password, end up using all cap?
What happens, if 3 team members, each have only 33 out of 100 limit, then it is even worse than personal plan at 60 limit for yourself (on paper)?
1 Like
I am having the same issue. Very frustrating. I think they changed it without telling anyone.
1 Like