Live data from Hacker News

Show HN: Watermelon – copilot for code review

github.com

21–30 of 58 posts

Re: Show HN: Watermelon – copilot for code review

#22
post #17

Congrats on the release. If I plug this into my repo, does Open AI get to train on my proprietary codebase?

According to OpenAI, they only harvest data coming from ChatGPT (their app), but not from third parties using the OpenAI API.

> How we use your data. Your data is your data. As of March 1, 2023, data sent to the OpenAI API will not be used to train or improve OpenAI models (unless you explicitly opt in).

> Note that this data policy does not apply to OpenAI's non-API consumer services like ChatGPT or DALL·E Labs.

https://platform.openai.com/docs/models/how-we-use-your-data

Re: Show HN: Watermelon – copilot for code review

#23

Calling it a “copilot” when there’s clearly another product in this space, seems a little, provocative. Its kind of ballsy to host it on github :p

Thank you! Yes, we settled on booking to the brand, who knows, maybe they will add us to their portfolio.

Re: Show HN: Watermelon – copilot for code review

#24

The homepage ( https://github.com/marketplace/watermelon-context ) states that Watermelon supports "C, C#, C++ and 7 other languages", but I can't figure out what these other languages are. Has anyone found this info? Is TypeScript a supported language?

Oh! Will fix that. Yes JS and TS are even what we build it in!

Re: Show HN: Watermelon – copilot for code review

#26
post #13

open source wrapper around OpenAI or do other LLMs work?

We appreciate your curiosity about our app. Just to clarify, we're actually a RAG application, not a wrapper. Our focus is on reviewing PRs in a manner similar to human reviewers, by tracing code context from various sources. Currently, we exclusively integrate with OpenAI's models because they've been highly effective for our needs. However, we're definitely open to the idea of supporting open-source and self-hosted…

Good luck, but using the term open source here is indeed confusing. While your own code is open source, the core of your app is the OpenAI API and without it your app does not work. So the project as a whole does not have the main characteristics of open source apps as generally understood (one can inspect the innards to see how it works, good privacy by default, free). It is sad that so much of the recent AI hype distorted the meaning of the term open source, both due to opaque models that can be downloaded for free and web-apps wrapping the OpenAI API.

Re: Show HN: Watermelon – copilot for code review

#28
post #18

How does Watermelon compare to CodeRabbit ( https://coderabbit.ai )? Very curious!

I hadn't heard about them before. Their product looks awesome!

From checking out their website, it seems like they've developed a more advanced way of chatting with line diffs, whereas we've built a superior PR summary system by tracing code context from various systems.

Re: Show HN: Watermelon – copilot for code review

#30

Calling it a “copilot” when there’s clearly another product in this space, seems a little, provocative. Its kind of ballsy to host it on github :p

There's absolutely nothing "ballsy" about describing our product as a "copilot" as a lot of other people are doing, or about choosing to host it on GitHub – the go-to platform for open source repositories. Our product distinctly differs from GitHub Copilot X, and we're confident in our approach.
Post reply on HN