Can you provide an estimated time frame for that so I don’t have to check three times a day?
1 Like
No one here can do this. This is a community forum for developers, there’s nearly never any OpenAI staff present here.
1 Like
Imagine the amount of items, news and features that appear every day and if every person in the world opened a ticket or post or something like yours.
7 billion’s x %N
1 Like
I think I’d treat it as a technical exercise, I’d get a spare laptop or a 2$ a month VPS and i’d put a small python script on it that calls the openai chat completions endpoint once every 6 hours with a call specifying the “gpt-4” model with the prompt of “test” and then a check to see if I get back an error message or a valid completion, then i’d make it fire me off an email if it detect the valid completion.
3 Likes
If you’re going to do something like that, I’d just poll the models endpoint and check if gpt-4 is present.
2 Likes
Hello I want to mention that I have received an email informing me about the availability to the GPT-4 API today.
1 Like
@bha443 can you share what you learned?
And if anyone can tell me the set of steps I need to do to become registered as an API developer, I’d love to know. I am on the waitlist and have been probably the same time as everyone else, but this will be my first thing to ever develop so I don’t have a prior API record with them. I’m assuming I may need to change that?
As @Foxabilo mentioned, it seems that the updates are done in batches based on unknown criteria (maybe sign up date, Geo region, etc…) but I can now confirm that there is progress for what it’s worth. I was also on the wait-list but did not do anything special other than what I mentioned in the post.
1 Like
To me it’s obvious that they prioritize users that use their product, in this case the API. I was lucky to get access to GPT3 API early, it was september 2020 if I remember correct. Since then, I have been using the different models via the API and have reported bugs I’ve found and tried various things.
To me, the criteria that makes the biggest difference how fast you get access whenever they release something new, is how active you are as a user. Sign up date, region etc. isn’t something I’ve noticed makes any real difference. I have friends that signed up earlier than me and they still waiting to access GPT4 using the API. While I got access in March. I live in Sweden, so not really a country that usually get prioritized.
The only difference between me and the friends that signed up early is that I have been using the API ever since I got access. Even though I’ve used the API for a long time, the amount of money I have been charged isn’t much.
I didn’t even have to sign-up for a waiting list to get access to GPT4 API. The only thing I can think of that made any difference is that I’m an active user.
I’m not sure what you mean by “become registered as an API developer”?
Add a card to your account so they can charge you based on how much you use the API. Then you can start using the API. To become familiar with the API there’s no need to have access to GPT4. How the API work is the same anyway (almost the same).
If I were in the situation you are I wouldn’t wait, instead I would start using the API, simple as that.
What is a “developer”? Anyone with OpenAI account? Do I need to register somewhere?
You are searching for this:
I don’t think there is an “official” definition of what a developer is, but I think it’s fair to say that if you use the API for building products, you are now a “developer”. I think it would be impossible to detect someone building software around the API and someone just using the API to get around the usage limits on ChatGPT, as long as you are using the services regularly… I’d say that counts.
1 Like
I’m a developer (stackoverflow .com/users/10133797/overlordgolddragon). I don’t use OpenAI regularly because GPT3.5 is insufficient for my needs, and I won’t spam tokens for X minutes a day since I value my time.
That’s my money that OpenAI won’t be getting. It’s also wasteful since someone could be using ChatGPT for memeing or other non-dev stuff and have no real need for GPT4, while actual devs are still waiting.
Consider adding an explicit registration, and prioritizing registrants. Regardless, thanks for clarifying.
1 Like
Who said that GPT-4 is a developer tool? Why do you think that personal entertainment is less worthwhile than what you work at?
A bizarre sentiment from someone coding for 30 years. The answer is obvious enough that the only question is in reverse - how is personal entertainment as worthwhile as scientific computing? Nobody’s saying bar GPT4 for good - but if there’s need to prioritize, the way is clear.
1 Like
I could probably talk hours about that. But think of it like there are two humans and both are worth the same.
Sounds pretty strange, I know.
One has the priority to work on whatever is so important to bring humanity into the next era and the other equaly important person is just chillin after a day in an elderly care where someone died with his hands on the head and now needs some way to work through it and creates memes on GPT-4…
I’d say your “scientific” programming can wait.
Human worth has nothing to do with this. Hospital beds are reserved for the sick, not those who have a kink to lie with needles injected. You want memes, scroll Reddit or watch YouTube.
1 Like
The point on “emotional support” in form of conversation, which cannot be acquired elsewhere, is valid, and indeed it may not wait unlike computing. I’ve seen stories of it helping people. But I don’t see why GPT-3.5 isn’t sufficient for that, nor know whether such needs are common enough to justify completely randomizing access.
Regardless, this isn’t on-topic - it appears OpenAI’s policy is to prioritize developers, and I am but pointing that they don’t appear to be actually delivering on it. If OpenAI wants to prioritize no one, then let them say so.
1 Like
Well, here is another chance (if your research might have any impact and you think you meet the standards of ethical usage):
The process is like this: you get microsoft certified, open a company, apply for microsoft partner company, then ask for a meeting with microsoft ai team and fill the first application. After ~4 weeks you should get an appointment and have to wait ~2-4 weeks to get access to GPT-3.5. Then you fill the second form and after ~6 weeks you get a message like this:
