I have Pro Claude plan, so I use that for deeper research. Not sure there's much difference now that Claude has web search. Perp is slightly more terse and also doesn't expend my tokens for Claude code
But I also haven’t used Google in years (I switched to other search engines) so their AI might be just as good.
Perplexity also has overlap with ChatGPT’s web search, but Perplexity is faster for me.
You get several models you can use (albeit they're not quite as powerful as their branded versions directly). you can setup projects and most other features you come to expect.
They do a good job with search though which is kind of their core value prop. They summarize well, cite sources, and have largely been pretty accurate with their findings. Being able to start a research task and provide more info and context while it is running is nice too for that "oh, I forgot about this!" moment.
With the pro sub you also get $5 in api credits a month so there's that.
Until Perplexity.
It was the AI product that actually reduced my Google usage. Even with AI mode directly built into Google homepage now, Perplexity is still better.
It has basically zero hallucination, each para/entry backed by a URL, and lower latentcy than any other LLM product.
I don't know why you find it bad. I use it daily, and for serious searches.
It has fundamentally changed the way I search the web/ask questions in the web. One particular way: I can now combine multiple subsequent searches into one. And I can trust the results.