The easy description of Perplexity is "ChatGPT with citations," and that undersells it in one direction while overselling it in another. It's not a better general-purpose assistant with a bibliography bolted on — it's a different tool built around a different job: finding and triaging information quickly, with a visible trail back to where it came from.
Where it earns its keep is Focus modes. Switching from general web search to Academic, Reddit, or Wolfram|Alpha isn't a cosmetic filter — it meaningfully changes which sources get pulled, and it's the difference between getting forum consensus on a practical problem versus getting peer-reviewed sourcing on a factual claim. For a quick fact-check that would otherwise mean opening five browser tabs and skimming each one, it's genuinely faster than Google.
The caveat most reviews skip: a citation link is not the same as a correct summary of that citation. We've seen Perplexity's synthesis confidently state something slightly stronger or more specific than what the linked source actually says — the link is real, the source exists, but the paraphrase has drifted. It's a smaller failure mode than a fully fabricated citation, but it means you still have to open the source for anything that matters, not just glance at the footnote number and move on. Its default web search Focus can also rank SEO-optimized aggregator content above the primary source it's actually summarizing, which is the same problem regular search has just wearing a cleaner interface.
It's not a writing tool and doesn't try to be — long-form drafting and multi-step creative work are noticeably outside its lane compared to ChatGPT or Claude, and it isn't a serious coding assistant either. Pro's "reasoning" search mode exists for harder multi-hop questions, but it's slow enough that reaching for it on a simple factual question is a waste of the wait.