I like it too. The basic use I like to think of it as "2-in-1." Previously with Google I'd type a query, Google posts back a page of results with links and short summaries. Than on that 1st (or 2nd) page, I go through the links, possibly clicking on some following to the content, reading some while skipping others, trying to find the info I need, and or keeping a running summary in the process too. So the process was - step 1 computer, step 2 human. Now Perplexity has taken over step 2 too, it has collapsed those two steps into one. Like Perplexity, have made it a default search engine on almost all browsers I use daily.
Great article. While I also use Perplexity now a days instead of Googling, would like your opinion on Grok 3 model. Would that be at par with Perplexity when it comes to the same use cases you have described in your article?
Just a small note on your comment, a lot of folk use Grok 3’s Deep Search in addition and combo with Perplexity as Google replacements. Genspark is also another good option.
Some of the citations/sources Perplexity uses are not great.
I use perplexity, but I also use Google's deep search, and OAI's deep search, and Grok's deep search, and other models like DeepSeek, Llama or Gemini 2.5 with websearch enabled. All of them have different pros and cons that determine best fit for different use cases, and they all share the problem of hallucinating sources that don't exist.
No doubt, search in general is changing, but Perplexity doesn't have any great advantage over alternatives in this, and it certainly doesn't have the potential for any kind of moat that might be accessible to Google.
At this point, none of them are replacements for having good research skills, and trusting in them without taking time to dig deeper and dyor is taking a substantial risk for the credibility of any knowledge worker whose value is in providing analysis based on sound reasoning.
Collections are a must-have. I don’t know why others haven’t implemented them yet. It remains a pain to find back an old discussion, almost like searching in WhatsApp...
If we consider 'the context' as the context of the pain point—being overwhelmed by too much noise from a feed or discussion—collections in Instagram, X, or even Pinterest should be considered.
Even Slack has a bookmark feature to facilitate finding back important messages, though it is not yet organized into collections.
Thanks for the post and some clarifying comments here.. Im used to copilot, when I use perplexity sometimes I appreciate that it avoids this conversational feel going more straight to the answer but other times I feel like they need to redesign the frontend of the results like more minimalistic
I know totally how you feel AC, I find the interface of both Perplexity and ChatGPT extremely bloated and confusing now. I wish too that it was more minimalistic. It’s one of the reasons I am using Genspark more for certain kinds of summarization tasks.
There are certainly many valid alternatives to Perplexity. The citation quality can be poor and the way it attributes citations and data from sources is full of hallucinations at times. Given those errors for me, it's hard to warrant a significant subscription.
Liner for me wouldn't work for me personally, it's way too slow.
I was hoping you did a perplexity deep dive, Alex! Thanks
I use Perplexity daily, no ads or sponsored results
Thank you for advice
https://substack.com/@martinzuzak/note/c-130455047?r=3en26e
I like it too. The basic use I like to think of it as "2-in-1." Previously with Google I'd type a query, Google posts back a page of results with links and short summaries. Than on that 1st (or 2nd) page, I go through the links, possibly clicking on some following to the content, reading some while skipping others, trying to find the info I need, and or keeping a running summary in the process too. So the process was - step 1 computer, step 2 human. Now Perplexity has taken over step 2 too, it has collapsed those two steps into one. Like Perplexity, have made it a default search engine on almost all browsers I use daily.
Is it really just an AI web search tool? What’s the competitive advantage with others doing that well now.
Great article. While I also use Perplexity now a days instead of Googling, would like your opinion on Grok 3 model. Would that be at par with Perplexity when it comes to the same use cases you have described in your article?
Just a small note on your comment, a lot of folk use Grok 3’s Deep Search in addition and combo with Perplexity as Google replacements. Genspark is also another good option.
Some of the citations/sources Perplexity uses are not great.
I use perplexity, but I also use Google's deep search, and OAI's deep search, and Grok's deep search, and other models like DeepSeek, Llama or Gemini 2.5 with websearch enabled. All of them have different pros and cons that determine best fit for different use cases, and they all share the problem of hallucinating sources that don't exist.
No doubt, search in general is changing, but Perplexity doesn't have any great advantage over alternatives in this, and it certainly doesn't have the potential for any kind of moat that might be accessible to Google.
At this point, none of them are replacements for having good research skills, and trusting in them without taking time to dig deeper and dyor is taking a substantial risk for the credibility of any knowledge worker whose value is in providing analysis based on sound reasoning.
Perplexity has been my search engine of choice the last two months.
Perplexity has been my search engine of choice the last two months.
Collections are a must-have. I don’t know why others haven’t implemented them yet. It remains a pain to find back an old discussion, almost like searching in WhatsApp...
That’s a really good idea actually! Being able to save, bookmark and tag research and searches is so important.
What is a good example of a product that has implemented collections in this context?
If we consider 'the context' as the context of the pain point—being overwhelmed by too much noise from a feed or discussion—collections in Instagram, X, or even Pinterest should be considered.
Even Slack has a bookmark feature to facilitate finding back important messages, though it is not yet organized into collections.
Closed to what we discuss, OpenAI released Canevas to help user keep note of their ai discussions.
Thanks for the post and some clarifying comments here.. Im used to copilot, when I use perplexity sometimes I appreciate that it avoids this conversational feel going more straight to the answer but other times I feel like they need to redesign the frontend of the results like more minimalistic
I know totally how you feel AC, I find the interface of both Perplexity and ChatGPT extremely bloated and confusing now. I wish too that it was more minimalistic. It’s one of the reasons I am using Genspark more for certain kinds of summarization tasks.
checkout KMF MAGAZINE SUBSTACK for the Golden Age
Amazing work ! Let’s connect 🚀
One of the biggest challenges I’ve seen with perplexity is that it many times gives incorrect citations or summaries especially in research contexts.
I’ve found GetLiner as a better source for research work.
There are certainly many valid alternatives to Perplexity. The citation quality can be poor and the way it attributes citations and data from sources is full of hallucinations at times. Given those errors for me, it's hard to warrant a significant subscription.
Liner for me wouldn't work for me personally, it's way too slow.
This is the future of search!
Have you tried A/B testing against Google’s new Gemini Advanced?