Thanks to the development of AI, new possibilities have emerged for Internet browsing and information search. One of them is Perplexity. By combining a search engine mechanism with LLMs (Large Language Models), it is possible to get better results and a summary, so you will no longer have to click through the various sub-linked pages to find answers to your questions. If you have a moment go to https://perplexity.ai/pro?referral_code=ZRHZX3A2, sign up for a free account and see how much time you save on your daily searches.
It’s hard to find an Internet user these days who doesn’t use Google. This search engine was founded in 1998 and, interestingly enough, has not undergone major changes since then. Abstracting from the recently hot topic of Google’s monopolistic practices, it is worth noting that quite a number of competing solutions have emerged that proudly stand up to such a well-known and proven search engine. We will be presenting and comparing the various solutions in more detail in the pages of this blog, but we would like to start by focusing on Perplexity, which is a leading competitor of Google using AI to reduce the time it takes to find necessary information on the Internet.
How do the mechanisms of Perplexity work?
We are all familiar with the flow of working with Google. We type in key phrases from the topic we are interested in, and then browse through the various pages of results by clicking on successive links and finding the content we are interested in. However, the development of LLMs and AI in general has led to the emergence of new possibilities not previously available. It is these that the company Perplexity has decided to use in its proprietary search engine.
The operation of the Perplexity search engine, in simple terms, boils down to 3 steps:
- Paraphrasing the text entered by the user so as to better understand it and increase the likelihood of returning more relevant results.
- Processing the results obtained from the traditional search phrases established in step No. 1.
- Generate a summary in a language the user can understand, using LLM.
The search summary thus generated is returned to the user along with a hint of other related queries in which he or she may also be interested.
Use cases
Let’s start with the most classic topic dreaming of senior front-end developers at night. Let’s ask Google and Perplexity “how to center a div in html and css”.
As we can see in the gif above, in Google case, just after scrolling the result page we know nothing. We still don’t know how to center a div. We have to click on the best-looking links and check if they are useful for us or not. My view isn’t the worst, because if you don’t have a plugin to block ads, you will see additional clutter with ads for programming courses and etc.
Now, let’s check the same query in Perplexity:
As you can see, you have to wait a few seconds for the query results, but after that, you get concrete code which you can choose and use in your application. You don’t have to check any sources. Of course, if you want, you can click on links connected to the selected part of the answer and see more information. This feature makes it better than ChatGPT, where (in the standard use case, without SearchGPT) you wouldn’t have the possibility to go to the sources and verify or extend the result. As you know, LLMs sometimes have hallucinations (and can provide incorrect information)—that’s why it’s so important.
What’s best: you are not limited to a specific LLM. You can set up your favorite LLM in the settings, and Perplexity will use it by default, or you can rewrite the response. This is especially useful at the beginning. You can compare results and see what works best for you.
Battlefield conclusions
Of course, not all problems can be boiled down to the question “how to center a div in html and css”. Despite this, the answers generated by Perplexity are usually enough so that you don’t have to click through search engine results and check individual links. This saves time every time. 🙂 If you need to go deeper into a topic-you still get links to the sources on which the answer was generated, so you can click on them and check the details. Sometimes, for slightly more complex topics, however, I go back to Google. If the issue is not solvable in a simple way that can be formulated after reviewing a few links, it is still worth reaching out to Google and going through the various links. Sometimes the most popular and obvious answer does not solve our problem. Only the juxtaposition of, for example, several side answers with documentation can help us solve the problem.
Another category for search purposes can be when we want to check some opinions from people on the internet and use these thoughts in our situation. In a case like this, Perplexity can show its best side. It can check what the internet thinks about your topic and summarize all this knowledge. You get concrete information in a few seconds, without clicking and checking all this stuff.
Pricing
Perplexity offers its services in two plans: free and paid ($20 per month).
The free version is full-featured, i.e. we have access to an unlimited number of queries, and I already encourage you to use it and see if such a solution will increase your efficiency. In the free package, however, you can only do 5 Pro searches per day. How do they differ from “non-Pro”? Generally speaking: they are of slightly lower quality. The search and processing process is simplified and contains fewer steps, so the answers may be inferior in quality to the Pro version. Here, however, I will reassure you: even the free version can handle the problem of “how to center a div in html and css.”
The paid version offers more than 300 Pro (most advanced) searches per day. In addition, it allows you to choose default models and rewrite answers using a different model. Interestingly, in this version you also have the possibility to upload your document (e.g. a pdf) and query Perplexity for its contents. Besides, there is the possibility to generate images using, for example, DALL-E or Stable Diffusion, but this topic seems to be the least refined. If you care about advanced graphics generation, Perplexity will not solve your problems.
The question arises: is it worth paying extra for the Pro version? As you can probably guess, there is no single right answer to this question. It all depends on how often you use the search engine and how complex problems you solve with it. Certainly, the worst solution would be not to take advantage of emerging opportunities to work more efficiently. Click here and create an account for yourself: https://perplexity.ai/pro?referral_code=ZRHZX3A2 First see if the basic version can speed up your work. If it does? Great, switching to Pro will help you even more! If not? Delete your account and go back to Google/DuckDuckGo or whatever you use. 🙂
Perplexity offers a lot more useful features, which we will be happy to tell you about in future posts. Add a comment, leave a rating or share the post and we will see that this topic is interesting for you as well and we will prepare another article expanding on Perplexity and similar solutions in general.
Ps. If you use provided link to Perplexity, you will get a $10 discount during your possible first payment.
[…] hope you are already familiar with perplexity and its basic capabilities. If not, I invite you to my previous post. In today’s post, we will talk about the slightly more advanced capabilities that this tool […]