Hacker Newsnew | past | comments | ask | show | jobs | submitlogin

Depends on your usage, I guess. I personnaly couldn't justify it. In the last few weeks, I've used the API extensively and total cost comes to a few dollars max.

Admittedly, the API doesn't provide nearly as many features as ChatGPT Plus does. To each their own !



> Admittedly, the API doesn’t provide nearly as many features as ChatGPT Plus does.

A lot of the ChatGPT Plus features are features of the wrapper rather than the model, sure. Then again, that’s what things like Langchain are for, right?


Not even kidding, I would pay $1000 USD per month for GPT4 access.


You know you can use phind.com with GPT4 access in both expert and creative mode? It's fine-tuned for writing code, but it's good for other stuff, too. Or do you mean API access? There is a form to fill out and they seem to be pretty generous with handing out access.

https://openai.com/waitlist/gpt-4-api

(edit: They didn't email me, I checked one day and I was able to use model="GPT-4", but not the 32K model.)


Wow, is phind completely free? How do you know that they're using GPT under the hood for their answers?

edit: found their ShowHN https://news.ycombinator.com/item?id=34884338 Anyway thanks for this recommendation!


Got off the waitlist for GPT4 the other day and have been experimenting with it a bunch for coding tasks (generation, review, planning, bug fixing etc.). I prefer 3.5 for the speed and cost and move to 4 for accuracy / detail when necessary.


Well, soon you'll can. Just select the GPT-4 32k and voila. 30 dollars a day.


I'm on the waiting list for GPT4 API access. Giving this thing an entire projects codebase will be super interesting!


i have been using it for some time. Great, but it has a tendency to give short answers and don't use the full context available. even if you set the max response size to 20k.


Comparing it to pasting code into GPT4 (the chat interface), how much better is it? I can't wait to see what this can do when an entire codebase can be held within the context window.


Yeah, it can help with some things, but if you ask it to, say, convert a python app to Go, even with plenty of context it will give you hints without full implementation.

I suppose they favored concise answers in the fine tunning process and reused the model to train the 32k.

Things that I have been doing: paste full table DDL with entities and other classes involved and asking for some refactoring of a class. Also that and some stacktraces. It definitely helps, but the price (at least on azure) make me think twice before using.

Also azure playground is awful.


How do you use it? I’ve been trying to change my habits to use it more, but some of those grooves are pretty deep and it just doesn’t come to mind.


It's been a watershed moment for me and honestly, even after just a few months I can't imagine life without it.




Guidelines | FAQ | Lists | API | Security | Legal | Apply to YC | Contact

Search: