At first, Perplexity got popular because it was the AI for research. There was a time when ChatGPT didn't have built-in search, and even when it did, it wasn't as good as Perplexity's.
Perplexity had the best product for asking a research question and getting a report with cited sources. They kept making it better, adding Deep Research and tools for things like finance. This is what I view as Perplexity's peak.
Then they started spreading themselves too thin.
They introduced Labs to build apps, trying to compete with Lovable and Bolt. It's now called Create files and apps. Then came Comet, their own browser. When OpenClaw came around, they started doing Perplexity Computer to try to own the personal assistant space as well.
Instead of being a first-rate Perplexity, they became a second-rate ChatGPT, a second-rate Lovable, a second-rate Dia, and a second-rate OpenClaw.
Comet was useful to me, but I stopped needing it once Codex had good computer use and browser use. Perplexity Computer was dead on arrival for me because open-source agents like OpenClaw and Hermes Agent were already better for what I wanted. Plus, I could customize them to my heart's content.
Coding is already a crowded market. Adding another app builder doesn't give me a reason to choose Perplexity.
Research still does. It's an unsolved problem. Frontier models routinely miss insights that are obvious to experts in the field. There's plenty of work left in making an AI that finds the right information, follows up on it, and produces a better answer.
That's where I want Perplexity spending its effort. I want the best research product. I don't need it to own my browser and my personal assistant too.
If Perplexity returns to research and makes their product the best at that, they can still win. If they don't, they'll fizzle out as the company that failed to pick a lane.