I found Phind and it blew my mind!

It is better than ChatGPT, uses GPT-4 under the hood, and is made for developers!

Questions that would have defeated ChatGPT, are no match for Phind!

More info on my blog post


Did you try copilot yet?

I have not!

How well does that integrate with Katalon?

I haven’t tried. All I can tell you is, the folks I’ve passed it on to are all impressed (but they’re not Katalon users, just regular coders).

Phind seems okay (when it’s not crashing on seemingly random keypresses). Good … uh… find :wink:


Whit chat-gpt’s, you only have to ask the right question:

Thanks for sharing. I think I’d seen Phind mentioned a while ago but neglected to check it out at the time. I’m impressed by their results, and I really like their citing sources. I’ll have to play with it some more.

FWIW, I’ve been checking out Cody from Sourcegraph. I really like what they are doing in terms of augmenting an LLM with code search. (Like you and others, I’ve been burned by hallucinations of classes like TestObjectConverter when chatting with unaugmented GPTs.)

I was turned on to Cody by this post from Sourcegraph Head of Engineering, Steve Yegge:

Cheating is All You Need

I think it is worth a read.


FWIW, I’ve used Copilot a bit. I’ve found it useful for code that is relatively self-contained and uses well-known libraries. Oftentimes it is useful just as smarter Intellicode/autocomplete within a single file. What I would really like, though, is more project-wide awareness and awareness of external resources like databases, web pages (particularly for testing), etc. Cody (see above) aspires to this.

Copilot doesn’t integrate with Eclipse or Katalon, yet, but you can check it out in Katalon scripts if you open up a Katalon project in VSCode.

Similar to my dream: feed it (et al) with a ton of stuff, then ask it questions, prompt it for tangential material…

Not just code, you understand… anything.