3 AI "Secrets" 99% of People Don't Know

Helena Liu

31,105 views • 3 days ago

Video Summary

This transcript explores three underutilized AI techniques to enhance productivity. The first method focuses on personalized learning by using Perplexity to gather resources on a topic, then feeding these into Notebook LM to generate custom lectures and even design a full curriculum. The second technique details how to create professional reports and interactive dashboards rapidly. Perplexity can analyze SEC filings, and this data can then be transformed into websites or infographics using Google Canvas and Gemini. The third approach refines interaction with AI models like ChatGPT by first prompting it to conduct deep research on a specific subject, then uploading this research as a PDF to a dedicated project folder. This ensures future interactions within that project are informed by the comprehensive research, effectively creating a specialized AI expert. An interesting fact is that Google's Notebook LM can learn from provided resources in seconds, and ChatGPT can generate an 18-page document in just 10 minutes

Short Highlights

Create custom lectures and design your own curriculum using AI by feeding resources into Notebook LM.
Generate professional reports and interactive dashboards in minutes from raw data using Perplexity and Gemini Canvas.
Train AI models like ChatGPT on specific research by uploading documents to dedicated project folders for specialized expertise.
Perplexity can analyze SEC filings with a finance tab, even on the free plan.
ChatGPT can generate an 18-page document of research in just 10 minu

Key Details

Personalized Learning and Curriculum Design [00:00]

AI can create custom lectures and entire curricula for any topic.
The process begins with using Perplexity to find a curated list of 20-50 online resources for a chosen subject, like Python, ensuring URLs are for educational content, not product pages.
These resources are then imported into Notebook LM, a free tool from Google, by pasting the URLs.
Notebook LM rapidly learns from these materials, allowing users to generate a custom lecture.
Users can prompt Notebook LM to create lectures in a specific style or for particular goals, effectively designing a personalized learning path.
This technique is useful for self-teaching or supplementing educational content for others.

"You can now teach yourself anything whether you're a visual learner or audio learner. Now AI can create any lecture that you want on any topic that you want."

Professional Reports and Interactive Dashboards [04:19]

AI can quickly generate professional-looking reports and interactive dashboards.
The process involves using Perplexity with its finance tab enabled (accessible even on the free plan) to analyze SEC filings for specific companies.
A prompt can be used to direct Perplexity to focus on key areas such as financial health, growth, profitability, risks, competitive position, and evaluation metrics.
Perplexity generates a detailed report within seconds.
This report can then be copied and pasted into Google Canvas via Gemini.
Within Canvas, the data can be transformed into various formats, including a complete website or an infographic, complete with charts and data visualizations.
This allows for the creation of impressive presentations with minimal effort.

"Imagine taking this to your next board meeting. Like, how impressive is this going to look? It looks like you put in a whole bunch of effort into this, but you've only like spent two seconds on it, right?"

Specialized AI Expertise and Research [08:49]

A more effective way to leverage AI for expert knowledge is to have it conduct deep research on a topic first.
Instead of simply instructing ChatGPT to "act as an expert," the approach is to prompt it to "do research on..." a subject, such as compelling copywriting techniques.
ChatGPT can then perform extensive research, diving into the psychology behind the topic.
The research output, even if extensive (e.g., an 18-page document in 10 minutes), can be copied into a Google Doc.
This document is then downloaded as a PDF.
The PDF can be uploaded to a specific project folder within ChatGPT.
This process trains ChatGPT on that specific body of research, ensuring that future chat threads within that project are informed by that specialized knowledge, effectively creating a custom AI expert for that domain.

"So instead of telling chat GBT act as an expert copywriter in the future, I can put all of my text in this specific folder or project here so that it's all organized."