Perplexity CEO Aravind Srinivas, Thursday Nights in AI

[@dukensonguerrier5369]

A reason I prefer perplexity over other LLM applications is because of organized searching. I am able to make collections of my queries which creates a friction-less experience when referring back to them.

[@michellegill1278]

Love it: "Only say what you can cite"

[@trinitybricks9522]

Right but the material that is cited is not dependable then what.

[@srini9653]

I just tried Perplexity AI, wow it is amazing. Well done, team!

[@shyama5612]

The Minerva answer (last 2 min) is brilliant - as I had the question myself on how to fine tune yet preserve the generalizability obtained with a lot of effort from training with a large/diverse corpus for the foundational model.

[@shrek22]

This is awesome. Thank you!!!! 💯✅☮️🦾

[@vivekanand3551]

Perplexity is future

[@mehdismaeili3743]

Excellent.

[@TheRickAustin]

Was it Nate Freeman that Aravind quoted?

[@yogendramiraje]

Nat Friedman

[@Ved3sten]

The product sounds really cool, but I don't understand how this is supposed to rival Google's AI Search? Google's product does the same thing, though not perfect, has good scope for improvement in the future since Google already has the index infrastructure. Is there some additional uniqueness or problem that Perplexity is addressing that Google isn't? Is is all about "product feel"? Because from a technical and ML standpoint, Google can surely come up with something in the future

[@pranayroy]

👏

[@RajanKannan11]

👏👏

[@Ring0--]

We're all doomed 😅

[@striker44]

8:13 that woman is seeing $$$$.