The deteriorating quality of Google searches has left me perplexed, but is the grass greener?

Browse any tech forum these days and you’ll find a recurring theme: people aren’t happy with Google Search. It’s not hard to see why – type in a query and you’ll often be greeted with shallow articles that barely skim the surface of the topic. It’s such a common complaint these days that entire AI startups like Perplexity are now trying to replace Google entirely, promising to offer simple, factual answers and no frills. But can Perplexity really outperform Google Search, or is it just another hype-driven AI platform? To find out, I’ve been testing it out over the last few weeks.

What is embarrassment and how is it different from Google Search?

Perplexity bills itself as an “answer engine” in an attempt to differentiate itself from traditional search engines like Google. Instead of displaying a list of links, it provides direct and concise answers to your questions by pulling information from multiple sources and quickly combining it into one answer. It uses generative artificial intelligence, but instead of creating an internal model, Perplexity relies on third-party language models such as GPT and Claude.

Confusion comes into its own when it comes to finding answers to questions like “why is Daylight Savings Time necessary?” or “should I soak or brine the turkey?” Summarized answers are ready in less time than it takes to manually sift through multiple search results. You still get a list of links you can click for more information.

Perplexity is not the first or only AI tool that promises to condense information from disparate sources. Google’s AI reviews do the same, but the feature attracts endless criticism for hallucinations and cherry-picking of sources. In one particularly egregious instance, Google’s AI referenced a decade-old Reddit thread that suggested using glue to prevent cheese from sliding off pizza.

Despite being valued at $1 billion, Perplexity is a bit of an underdog in the AI ​​race – you’ll rarely hear it discussed in the same breath as ChatGPT or Gemini. However, it has gained decent popularity among power users, mainly due to the Pro Search feature, which I will discuss later.

Google has had to restrict its AI review feature following a number of issues since its launch. So has Google’s failure to realize the technology’s potential opened the door for Perplexity and other AI startups like OpenAI?

Embarrassment vs. Google: Comparing Free Services

For Perplexity to be successful and useful in everyday use, it must offer accurate answers based on reputable sources. In theory, this is a simple task, but an unpredictable AI model is difficult to execute. Google has improved its AI review feature to prevent hallucinations like the glue-on-pizza story mentioned above, but I’ve noticed that it still references questionable sources — like TikTok in the screenshot below.

However, when comparing the actual quality of answers, Google and Perplexity are unmatched for this question. I like that Perplexity’s answer included some in-text quotes that I could easily click on. Google’s output was much shorter, with only one citation per entire paragraph. This meant I couldn’t verify the accuracy of individual sentences as easily as Perplexity. For such a simple question, it doesn’t matter, but it’s worth checking references if I’m using a search engine’s answers for studies or work.

For my next question, “next FOMC meeting,” Google didn’t post an answer or an AI-generated answer. In turn, Perplexity responded by providing the date and a brief explanation of the US Federal Reserve meeting calendar. However, I saw an AI hallucination here – Perplexity stated that interest rates have not changed since July 2023, which is untrue because the Fed has already cut rates by half a percentage point.

You may have noticed that the Google results page above does not include an overview of artificial intelligence. This is no coincidence – the search engine only displays AI summaries for a small percentage of responses. So if you care about AI-driven information density and the time savings that come with it, Perplexity is the much more consistent option of the two.

I could give more examples, but after a few weeks of testing, I came to the conclusion that Perplexity is indeed faster than Google in terms of fast searches. Whether I was looking for the cheapest way to travel between two cities in Thailand or I was curious about the impact of rate cuts on the stock market, Perplexity consistently provided me with the answers and quotes I needed. I used Pro Search when I needed the most accurate answers, but more on that below.

Perplexity also offers a chatbot-like interface, which means you can ask follow-up questions. Overall, it feels like a hybrid of ChatGPT and Google Search, while being more useful than either of them alone.

However, things eventually took a turn for the worse when my search terms became less clear. The prompt “Compare the latest smartphone models from Apple, Samsung and Google” yielded results describing the iPhone 13 and 14. Honestly, I wouldn’t enter such an open-ended question into Google, but the search engine at least returned results for the iPhone 15 and other modern phones.

Pro Perplexity Search: Where Google Can’t Compete

So far I’ve only talked about Perplexity’s core experience, but the platform has another trick up its sleeve: Pro Search. Basically, it performs a deeper search and analyzes more sources before answering a query. However, you only get five professional searches per day for free. Other than that, you’ll have to pay a $20 monthly subscription to Perplexity Pro. Subscription also unlocks higher quality responses with more intelligent AI models like GPT-4o.

A few months ago, Perplexity’s Pro Search feature was also significantly improved, allowing AI to perform complex human-like searches. This allows the AI ​​to answer extremely complex and multi-part questions that even some of the best AI chatbots will encounter.

For example, take the screenshot above where I asked Pro Search to compare the earnings of two publicly traded companies over a ten-year period, adjusted for currency differences. Confusingly, it split this query into four individual searches, mimicking how I would approach solving the problem. The result included mathematical calculations that Perplexity feeds to Wolfram Alpha. Everything worked impressively well, but it took over a minute for Perplexity to generate a response.

Will I switch from Google to Perplexity?

Given the impressive results, you may be wondering if it’s worth making Perplexity your default search engine instead of Google. I’ve considered this choice myself, but at least half of my online searches involve finding local businesses, finding real-time information like weather, or going to a specific website like Amazon. I could use Perplexity for some of these tasks, but it would be slower and much less efficient. I also wouldn’t trust generative AI to retrieve my flight status, given the risk of hallucinations.

All in all, confusion is a bit of an overkill for simply navigating the Internet, and Google is still your best bet if you’re just trying to get to your bank’s login page. With that said, Perplexity has replaced Google for about a third of my daily searches. The app is now available on my phone’s home screen, just above the Google search bar. So instead of replacing my primary search engine, Perplexity has almost completely replaced traditional AI chatbots like ChatGPT and Google Gemini. The latter will only selectively search the Internet, while Perplexity does it every time. This perhaps explains why OpenAI created SearchGPT, which is currently being tested behind closed doors.

Can you see yourself ditching Google in favor of an AI-powered search engine?