Some positive feedback & getting other contributors #160
Replies: 4 comments 5 replies
-
@florianfelsing That’s awesome! Thanks for sharing this. That was one of my original motivations for it as well. The funny thing is that my wife also thinks it’s a really good idea because she only uses ChatGPT maybe 2-4 times per month and doesn’t want to pay $20 for that, but she wants GPT-4 as well. It’s on my list to set her up with her own instance this week so she can stop using my account. :) So it’s actually both extremes—using it a lot and using it a little—that make it nice to have a different payment model. |
Beta Was this translation helpful? Give feedback.
-
Oh, and related to Kamal, yes! Eventually I wanted to make sure we got Kamal setup and also easy Heroku integration, since enough people still use Heroku. It was actually one of the reasons that Rob went ahead and got docker compose integrated already. We knew we’d eventually want to support docker deploys. If you want to tackle this because you’re doing it anyway then by all means. If you decide to prioritize it, just comment on this so others know you’re on it: It’s also fine to wait until later, totally up to you. There is quite a long tail of things to improve to get a really solid 1.0. :) And speaking of which, if you have any ideas for finding a couple more people to contribute, I’d love to hear. Keep that in mind. |
Beta Was this translation helpful? Give feedback.
-
google_us_rails-open-source-pr_matching-terms_2024-02-28_20-34-02.csv So there is definitely some volume for these terms, check the CSV attached to this comment. It should not be hard to be everywhere when people search for these. For now I'd probably try to:
Let's coordinate about the outreach - as I said, feel free to delegate some of that to me. The other things, I think we can handle independently. Edit: Data source for search volume is Ahrefs. |
Beta Was this translation helpful? Give feedback.
-
I've been thinking about this quite a bit over the past few days and I think I finally have clarity on how to approach this. First, I believe the best way to find people who want to help make this project better is to first get people to use this project. At the the point we have 100 rails engineers all using HostedGPT, people will run into lots of little issues and have new ideas and this is what will create the motivation to fix those issues. My own motivation to help out on other open source projects almost always comes from me using it and running into an issue that I care about. So the most important question is: what will motivate a lot of rails engineers to want to personally use HostedGPT. This made me realize something important. I've been focused on getting to feature parity between HostedGPT and ChatGPT. This will take many more weeks. But 100% feature match will only create a minimal motivation to use HostedGPT. The primary motivations can really only be: (1) better control over your data and (2) better control over billing/usage. But I think the thing that will most motivate someone to want to use HostedGPT is that if it has a few compelling features above & beyond ChatGPT. I'll continue to chip away at the Issues in the 0.6 bucket, but I'm going to shift my focus to adding the first big bonus feature: the ability to verbally converse with the bot within your browser & to let it see your screen. I already have a working prototype of this, but it's going to be a chunk of work to get the code cleaned up and integrated into HostedGPT. That's what I'm going to focus on. I think this will be a pretty wow feature. I've used it myself for a bit and it's really useful. After this is done, then I'll use this "bonus" feature as the best motivation for convincing people to try HostedGPT for themselves. In the meantime it's worth reaching out to any other rails engineers you know that may be interested in using HostedGPT as-is. But in a couple weeks I think it will become a lot more compelling. |
Beta Was this translation helpful? Give feedback.
-
Hey guys,
just got some really positive feedback about our project from my wife. She says that something like this is exactly what she had been looking for, because she's constantly running into the message limit using ChatGPT.
I think it's pretty cool to see end users excited about this project!
As for deploying the project - if interesting to us at all - I could at some point add instructions for deploying HostedGPT using Kamal as well. For now I'll use Render, but this shouldn't be hard to accomplish with Kamal and it would give our users some flexibility in terms of their infrastructure.
Beta Was this translation helpful? Give feedback.
All reactions