Eliot Prince - power user of Perplexity

---

---

https://www.youtube.com/watch?v=eJ0FczURNNI

The speaker in the video highlights Perplexity as the one essential AI tool, claiming it can replace the need for separate subscriptions to platforms like ChatGPT, Claude, and Gemini due to its comprehensive features. Here’s a detailed breakdown of the features and capabilities discussed:

1. Integrated AI Model Selection (0:35):
   • Perplexity allows users to choose from a variety of top large language models (LLMs) within its single interface. This includes:
     • Perplexity’s own Sonar (fast model)
     • Anthropic’s Claude 4.0 Sonnet (advanced model)
     • OpenAI’s GPT-4.1 (advanced model)
     • Google’s Gemini 2.5 Pro (latest model)
     • xAI’s Grok 3 Beta (latest model)
     • Specialized “reasoning models” like Reason1.lng, R1 1776, and o3.
   • A “Best” option automatically selects the most suitable model for each query (e.g., Sonar for research, Claude for coding). This eliminates the need to subscribe to multiple AI services.
2. Advanced Source Filtering (2:06):
   • Users can precisely control where Perplexity pulls its information from by toggling specific sources on or off. Options include:
     • Web: Searches across the entire internet.
     • Academic: Searches academic papers.
     • Social: Focuses on discussions and opinions (e.g., Reddit, forums).
     • Finance: Searches SEC filings for financial data.
   • An example showed searching “what is the best AI model right now?” with only the “Social” source enabled, resulting in an overview and comparative table based on Reddit community consensus.
3. Deep Research & Labs Functionality (3:30):
   • Perplexity is presented as a purpose-built research tool that goes beyond simple queries.
   • Researching Pain Points (3:30): The speaker demonstrates a query: “I want to come up with proven micro saas ideas and tools I can build. This requires finding peoples pain points and problems which they are struggling with. Then we know from the start that the tool/saas will solve a specific problem. It’s pre-validated. Can you research Reddit for problems and pain points that could be solved with a micro saas or tool. I want to focus on the parenting niche.”
   • Output and Validation (5:06): Perplexity performs a deep search across 90 Reddit sources and threads, identifying real pain points and potential SaaS solutions. Examples include “Sleep Training Support Tool” and “Family Calendar Coordination Platform,” with validation pulled directly from user discussions.
   • Project Creation with Labs (5:57): The “Labs” feature allows Perplexity to act as an AI agent to build out projects based on research. From the parenting pain point example, the speaker prompts Perplexity to “start bringing to life with a business plan, brand mood board, and mvp features to get us going.”
   • Comprehensive Business Plan (6:34): The Labs feature generates a detailed output including:
     • A project name (“FamilyFlow”)
     • Market Opportunity (TAM, SAM, SOM with estimated annual values)
     • Competitive Gap analysis
     • Core Value Propositions
     • MVP Feature Set (with detailed functions and validation metrics)
     • Brand Identity & Mood Board (including color palette and image suggestions)
     • Technology Stack for building the app
     • Roadmap beyond MVP
     • Conservative Financial Projections
     • Key Risks & Mitigations
   • App Mockup Generation (7:39): Labs can also create visual mockups of applications and dashboards. An example shows a “cool modern techy glowing vivid magenta and vivid green ‘SEO competitor’ dashboard” complete with interactive charts and key metrics.
4. Image Generation (8:00):
   • Perplexity can generate images directly from text prompts, either integrated into Labs projects or as standalone creations in the chat. The speaker demonstrates creating YouTube thumbnails and even a “professionally lit and edited photograph” of himself, showcasing its ability to combine user photos with generated graphics.
5. Spaces (8:54):
   • Perplexity offers “Spaces,” which are akin to custom GPTs. Users can:
     • Define custom instructions and roles for the AI (e.g., “Brainstorm Buddy”).
     • Upload local files or add URLs to provide specific context for the AI’s responses.
     • Select any of the available AI models (GPT-4.1, Claude, Gemini, etc.) within that specific space.
     • Choose specific sources (Web, Academic, Social, Finance) for the AI to consult within the space.
     • Mention and link other “Spaces” together for complex, interconnected projects.
6. Automated Tasks (10:00):
   • Users can set up recurring tasks to automate research and content delivery. The speaker shows a weekly task to research “most important AI technology trends, product launches, and innovations,” delivering the curated information automatically.
7. Voice Input (11:45):
   • Perplexity supports voice prompting, allowing users to interact with the AI hands-free.

In conclusion, Perplexity is presented as a powerful, all-encompassing AI platform that centralizes various advanced AI models, offers detailed source control, enables deep research for business development (including comprehensive business plans and app mockups), generates images, allows for custom AI “spaces,” and supports automated tasks and voice input.