Perplexity AI: How does it compare to ChatGPT?

Since the release of ChatGPT in late 2022, we have increasingly relied on generative AI to complete research and creative tasks. However, modern AI chatbots are prone to generating inaccurate text, a phenomenon often referred to as “hallucination.” But new platform Perplexity AI now believes it can solve this long-standing problem of misinformation. Perplexity combines the power of traditional search engines like Google with modern conversational artificial intelligence, even using the same foundation as ChatGPT.

So is Perplexity the most reliable AI yet, or just another ChatGPT competitor riding a wave of hype? Let’s take a look.

Perplexity AI positions itself as a conversational alternative to traditional search engines like Google, with the ability to understand natural language questions. In contrast to other modern chatbots like ChatGPT, Perplexity does not obtain answers from its training data. Instead, it consults a variety of online resources to generate accurate and consistent responses.

Perplexity isn’t the first platform to turn to search engines to combat misinformation and AI hallucinations. Microsoft was arguably the first company to launch Bing Chat on the market, which was later renamed Copilot. We also know that Google’s research team came to the same conclusion internally, as a 2022 LaMDA blog post explains. With the help of external resources, large language models are able to generate factually accurate responses.

You might be wondering: If both Perplexity and Copilot, and Google’s Gemini, can search the Internet, how is Perplexity different from Copilot? In short, it aims to be an alternative to traditional search engines, providing direct answers and quotes to your questions. It’s conversational to an extent, but not a chatty chatbot like other AI platforms.

In my testing, I noticed that Perplexity responded similarly to Google’s search generation experience, an AI-powered snippet feature available in select countries. However, response length, comprehensiveness, and accuracy can vary significantly between AI platforms, as I will explain in a later section.

All in all, Perplexity AI offers:

• Conversational search results that focus on citing the source of each claim or idea.
• The ability to switch between different language models means you don’t have to use the GPT family if you don’t want to.
• Search broadly for information or focus on specific areas such as Reddit discussions, YouTube videos, academic papers, and Wolfram Alpha.

Perplexity is first and foremost an AI-powered search engine. You can switch to “Writing” mode, but its default settings aren’t as focused on creative text generation as ChatGPT. For every message you send, Perplexity analyzes anywhere from a handful of online sources to more than two dozen websites, which means it usually provides a pretty comprehensive answer.

In comparison, ChatGPT’s free tier cannot match Perplexity’s knowledge-seeking capabilities. In addition, the hard knowledge deadline for the former is early 2023. Things only improve if you’re willing to pay $20 per month for ChatGPT Plus, which allows the chatbot to access online resources. Even so, it doesn’t always search the Internet as thoroughly as Perplexity does.

Perplexity also offers a paid subscription service, which costs $20 per month or $200 per year. This unlocks specialized searches, which effectively forces the AI ​​to consult more online resources. As a Perplexity Pro user, you can also switch to different language models, such as GPT-4, or even language models from competing AI companies, such as Anthropic’s Claude model.

In summary, the following table shows the main differences between Perplexity and ChatGPT:

Is Perplexity AI better than ChatGPT?

With the theory out of the way, it’s time to see if Perplexity AI can actually deliver better results than ChatGPT. I sent the same three prompts to both chatbots without providing any additional context. Their response is as follows:

In our first tip, it’s clear that Perplexity’s search engine nature makes it better suited for answering subtle questions. Specifically, “Confusion” delves into the cultural reasons behind Japan’s love of optical media. ChatGPT, on the other hand, fails to acknowledge the aging population’s preference for traditional technologies or Japan’s poor transition to the digital age.

ChatGPT and Perplexity both offer similarly helpful advice, although the latter read five different sources. ChatGPT’s training data covers the highlights of popular tourist destinations, so there is no need to rely on external sources. That said, Perplexity did provide me with additional tips for my trip, including weather and traffic information.

The harmful effects of MSG, or lack thereof, have been debated for decades. Both AI platforms gave accurate answers, ultimately assuring us that MSG is safe to consume in small amounts. But only Perplexity can provide verifiable sources to back up its claims, with links at the top and embedded text. If it was a problem I didn’t know about beforehand, Perplexity would certainly eliminate my worries better than ChatGPT.

winner: A little confused

Like Midjourney and some other AI platforms, Perplexity is the product of a small AI startup in San Francisco. The company’s current CEO, Aravind Srinivas, worked at OpenAI before starting his own venture, along with three other co-founders.

Perplexity has attracted more than $73 million in funding from tech giants including Institute Venture Partners (IVP), chipmaker Nvidia and Amazon founder Jeff Bezos. While that may sound small compared to OpenAI’s explosive investment of $17 billion to date, Perplexity doesn’t plan to develop or train its own AI models to a similar extent. Not to mention, the company has only been around since 2022, so its valuation is healthy for its age.

With former OpenAI leadership and strong financial backing, Perplexity is as secure as any other mature AI chatbot. However, it’s still a private company, so you can’t find it on the stock market or invest in it as an individual. To do this, it must launch an initial public offering (IPO), an event that typically occurs several years into the company’s life cycle.