I like that it provides sources, but I have to check them EVERY time because too often it hallucinates bogus solutions or protocols for me. I'm asking network questions, though, not asking for code snippets. I've had it hallucinate powershell modules as well. If you're willing to check it's work, then its useful maybe.
I tried to remember specific examples, plugged them into Phind, and it seems to have improved since I used it last.
We had asked it if Cisco LISP supports map server database replication. It doesn't, and maybe this is a bad use case, asking it about something that doesn't exist.. But it basically said "yeah it does, here's how" and spat out some irrelevant stuff (how to do manual database mapping, etc). Again, probably a bad example considering the prompt was kind of garbage, but I did expect it to "know" the difference between replication and manual data entry.
Had a procedure in a powershell script to query a name server for A records. It used a deprecated system class and I believe missed some syntax. Ended up getting what we wanted out of it by re-wording the prompt a few times. I can't replicate this now, I don't even have the original script, so sorry I can't help more.
Has Phind improved SINCE going to Gpt-4? It seems to have?
Anyways.. I still use Phind. Hopefully I haven't undersold it, because it's usually great and I recommend it to everyone at work. Like everything else AI, the prompt really matters, and any frustration I had with it was undoubtedly due to expecting too much of the backend (actual ChatGPT) and not the Phind interface.
