in case it helps, you’re not missing out too much on anything with not being able to use Claude 3 Opus, even the current ChatGPT 4’s problem solving is superior, but Claude 3’s coding ability with the right instructions is better - at the moment I’m using ChatGPT to create detailed instructions for what needs to be rectified and Claude 3 is writing the code to do it.
1 Like
I mostly use AI for coding which is why I wanted to try it, I’m pretty good with giving instructions to get the code I want. I guess the silver lining is that I got banned before I got to try it and find out if I like it, so they didn’t get any of my money and I got to find out they’re a idiotic company with implementing security features that puts the burden on their users that even with billions in funding for, they can’t muster up enough customer service to reinstate all the people they falsely banned or figure out how to fix the problem (more complaints of this show up daily). So I gues they aren’t a company I’d want to do business with anyway.
2 Likes
Fair enough! Working on neural network + machine learning related code for the last few months that’s well above my normal coding ability so require the AI I use to help me to be able to actually help diagnose and fix things I can’t since a lot of the technical terms + the way it’s meant to be structured are outside of my technical ability, going to feel like Xmas when GPT 5 comes out. Pray to god they have the continue generating button and don’t implement any idiotic limitations when it finally DOES come out. Wonder how much longer we’ll have this version of GPT 4 before they rectify whatever they changed
My best guess is that it’s because of their new stuff like video-generating AI. They have finite physical resources, GPUS, Servers, etc.
When they release new stuff that’s high resource demand, it’s almost always followed by technical problems and outages followed by a nerf bat to the model.
The last time they took a nerf bat to ChatGPT-4 Plus was when they released ChatGPT Enterprise. However, it’s not like they don’t know this and they learn the demand, get new equipment dedicated to provide the corresponding need, and once they’ve caught up hardware-wise, they seem to start reverting the throttling to ChatGPT which seems to harbor the most noticeable brunt of nerf inflicted trauma.
I suspect it won’t be too long, but when you think about it, there’s no way they can know before experiencing the demand how much resources will be used by their new releases, nor can they know in advance which companies they’ll sign to use them or how bad those companies will drain resources using it.
You can really only use deductive logic on this kind of stuff because the company has like zero transparency now, if ChatGPT went offline tomorrow we’d probably get something like “sorry, we had a technical problem” three days later.
I don’t think they intentionally keep it this way, so I’d say however long it takes them to get more dedicated resources to handle whatever the demand is for their new stuff. I haven’t touched a lot of it, but I saw the press announcement for video generation, I imagine that sucks some serious GPU uptime, like an exponentially huge amount.
I don’t think it’ll be like “turn key, it’s fixed”, but I think incrementally as they get those resources setup, we’ll see improvements. Probably within a month it’ll be better than it is now.
1 Like
Same here, reported twice, no solution what so ever.
same here, make the button “continue generation" again
Same problem here! I’m waiting for response emergency!
GOOD NEWS！Mine is back! Guys, try it!
1 Like
For more information on what OpenAI is doing to GPT read this thread:
EDIT: Per today “continue generating” is back for me too.
[I still don’t have the ‘continue generating’ button back. I still have to use a workaround by telling it where it stopped and ask it to continue. Very clunky.]
This topic was automatically closed 2 days after the last reply. New replies are no longer allowed.