Learn 80% of Perplexity in under 10 minutes!

Perplexity for Beginners

Perplexity is an AI search tool best suited to finding current, source-backed information. ChatGPT and Gemini are described as stronger choices for brainstorming, writing, and other creative tasks, while Perplexity is recommended when accuracy and up-to-date information matter.

Key Features and Guidance

• Privacy and profile settings: Perplexity lets users exclude their data from model training while retaining conversation history in the Library. Profile instructions apply across conversations, much like custom instructions in ChatGPT. The presenter recommends setting these up once to tailor responses.
• Choosing a search focus: The Focus feature narrows which sources Perplexity uses. Examples include the whole web for current news, academic sources for supplements, social sources for workout recommendations, and videos for exercise demonstrations. A separate Writing option generates text without searching the web, though the presenter prefers paid ChatGPT or Gemini for creative writing.
• Search results and sources: Perplexity gathers information from multiple sources and provides linked citations. Users can expand the sources and remove ones they consider outdated or unreliable. Responses may also include supplementary videos or images.
• Search comparison: In the presenter’s example about the most recent Olympics, Perplexity returned the answer directly, while Google Search results required more clicking. The presenter notes that Google AI Overviews would be a closer comparison than ordinary search.
• File uploads: Free users can attach PDFs, but not images. Converting an image to PDF is mentioned as a workaround.
• Collections: Collections organize related threads and can include a prompt that applies to every conversation in the collection. Examples include:
  • A Japan travel collection tailored to local, tourist-friendly recommendations.
  • A presentation collection that generates outlines using the SCQA framework.

The presenter values citations in presentation outlines as a way to check figures and claims.

• Limitations: In a thumbnail-brainstorming example, the presenter found ChatGPT’s ideas more creative and relevant. The overall recommendation is to use Perplexity for research and fact-focused tasks, and ChatGPT or Gemini for demanding creative work.
• Free vs. Pro: The presenter says most users will be fine with the free version. Pro adds model selection, image generation, and Pro Search, which uses more sources and gives more detailed responses. The presenter states that Pro Search is available up to 600 times per day, compared with five times for free users. Access to other models through Pro does not provide the same experience as using those models in their original products, because each service applies its own tuning and search approach.
• Other features: Perplexity can generate a page, such as a blog post, and publish it to the Discover feed. The presenter does not use this feature. Perplexity Labs is described as a sandbox for developers to test its language models before integrating them into applications.

Main Speaker and Sources

• Main speaker: Jeff Su.
• Sources and products discussed: Perplexity, ChatGPT, Google Gemini, Google Search, Google AI Overviews, Apple and Marriott help pages, Reddit, and Perplexity’s linked web sources.