By playing, you agree to uListen's Terms and Privacy Policy and YouTube's Terms of Service.

The Calum Johnson Show

AI Reporter: The NEW Way To Use A.I. To Get Rich in 2025 (Chat GPT, Gemini & Perplexity Labs!)

Ready to see how Square can transform your business?

Matt WolfeguestCalum Johnsonhost

Jul 14, 20251h 40mWatch on YouTube

Summaries, chapters, transcripts, and FAQs on this page are made by uListen, not YouTube.

Great conversations run long. Their best ideas don't have to.

uListen is free on iPhone, iPad, Android, Chrome, and Claude.

CHAPTERS

Jump to publisher timestamps

Chapter summaries

Why Three AI Tools Are Enough and What Feels Eye-Opening
Calum frames the episode around people overwhelmed by AI and asks Matt why ChatGPT, Perplexity and Gemini are his daily tools. Matt says knowing these three keeps you well ahead of most people, and cites Veo 3 and Perplexity Labs as recent eye-opening developments.
- •Calum describes two camps: those overwhelmed by AI and those who overestimate it
- •Matt says the big three tools are still the most valuable
- •Veo 3 video generation is getting close to fooling people
- •Matt worries social platforms may cut out creators by generating content themselves
- •Perplexity Labs feels like a $5,000–$10,000 consultant
Becoming the AI Orchestrator
Responding to a Dan Martell quote about top earners being AI power users, Matt agrees and compares the future skill to a symphony conductor. He says the winners will know the tool landscape, prompt well, and direct AI agents across tasks.
- •Matt agrees with the claim that AI power users will be top earners
- •Value moves up the chain to those who conduct many AI tools
- •AI may eventually act as salesperson, coder, designer and video editor
- •Human craft roles won't disappear entirely
Prompting Like a Teammate and Feeding Context
Calum walks through a three-step prompt framework he got from ChatGPT. Matt thinks prompt engineering and role-priming are overvalued, prefers talking to the AI like a Slack teammate, and stresses giving as much real context as possible, such as analytics screenshots and writing samples.
- •Matt finds "you are an expert" priming overvalued
- •He prompts the AI like a team member on Slack
- •More context, such as analytics screenshots, always helps
- •AI is a strong correlation-finding machine for titles and metrics
- •Feeding it your own writing samples gets it closer to your style
Why AI Is Moving So Fast
Calum asks why people are so fascinated with AI for business. Matt traces the exponential pace from 2021 to today, from no image or video generation to ChatGPT, GPT-4, Midjourney, Sora and now far better video models.
- •Capabilities leaps are coming closer together
- •Timeline runs from GPT-3 through ChatGPT, GPT-4 and Midjourney V4 to Sora
- •Today's video generators make Sora look weak
- •The pace is why people feel both freaked out and behind
What Each Tool Is Best At
Matt explains the nuances of Gemini, ChatGPT and Perplexity and where each fits. He also advises paying for ChatGPT and using the free tiers of AI Studio and Perplexity, starting a business by brainstorming a roadmap with o3.
- •Gemini: huge context window, Deep Research, NotebookLM podcasts, coding
- •ChatGPT: brainstorming partner, human-sounding scripts, image generation
- •Perplexity: the new Google, plus Labs and troubleshooting
- •If paying for one tool, Matt would choose ChatGPT
- •Start a business by asking o3 for a roadmap
ChatGPT Demo: Fixing an Underperforming Video
Matt shows a past o3 prompt using two screenshots and a simple request to optimize an underperforming video. They discuss turning the output into learning for the next video, treating ChatGPT as a conversation rather than Google, and counteracting its agreeableness.
- •A simple prompt with screenshots produced consultant-level suggestions
- •Matt acts on title and thumbnail tests and applies lessons to future videos
- •ChatGPT memory supplies past context
- •Treat it as a back-and-forth conversation, not a Google search
- •Asking for brutal honesty counters agreeable answers
Building a Shorts Writer with ChatGPT Projects
Matt shows how he uses ChatGPT projects with custom instructions and example transcripts to write Shorts scripts and explain research papers from just an uploaded PDF. He also compares o3 and 4o, and Calum likens projects to departments inside ChatGPT.
- •Projects hold custom instructions so no extra prompt is needed
- •Shorts scripts are drafted from news articles or PDFs using sample transcripts
- •AI handles research and scripting, while recording and editing stay human
- •o3 is analytical; 4o is more creative and human-sounding
- •An Explain This For Me project simplifies papers
ChatGPT Image Generation in Practice
Matt demonstrates conversational image editing in ChatGPT and discusses how he actually uses it. Currently it produces assets, not finished products, for thumbnails and B-roll.
- •Edits are made by telling it what to change, not rewriting prompts
- •Background removal works with a single short follow-up
- •Thumbnails still involve Photoshop and real photos
- •AI images are used for B-roll gaps and animated in video tools
- •It's a component of the workflow, not the whole workflow
Perplexity as a Google Replacement
Matt explains that he uses Perplexity instead of Google for most questions. It performs the search and reads the articles for you, giving a summarized answer without the ad-heavy results.
- •Matt can't remember his last Google search
- •Real-time answers, such as sports scores
- •Even lazy searches like "best digital camera" get synthesized reviews
- •Perplexity runs the search and reads the top articles for you
Perplexity Labs Deep Dive on YouTube Strategy
Matt shows a Perplexity Labs report generated from one prompt about viral AI YouTube channels. It searched 62 sources, produced charts, keyword effectiveness scores, a production workflow and action steps. Calum connects this to researching competitive landscapes for a new business.
- •Labs takes about 10 minutes and delivers a consultant-style report
- •One prompt with no uploaded context produced charts and strategies
- •Keyword analysis showed which title words correlate with better performance
- •Competitive research that took weeks can happen in minutes
- •Labs requires the paid plan with a monthly query quota
Fast MVP Testing Over Perfect Planning
Matt notes that research doesn't replace testing, because it is now easy to build minimum viable products and see if people respond. Calum adds that the guests he's interviewed share a habit of firing off small tests and going all in on signals.
- •AI makes MVPs like simple apps or newsletters quick to create
- •Test with ads and move on if there's no interest
- •Calum sees constant small tests as a common trait among successful guests
- •Feedback and data come faster with AI
Gemini Deep Research for Interview Prep
Matt demonstrates Gemini Deep Research, which he uses to prepare for interviews with guests such as Demis Hassabis and DJ Sampath. It pulls information on a person and exports a structured Google Doc, while he and his producers still decide the angles.
- •Deep Research takes about 10 minutes and writes an essay-style report
- •Matt uses it to find novel angles for interviews
- •Results export directly to a Google Doc with a table of contents
- •He doesn't use its questions verbatim, only the concepts
- •No dedicated researcher on his team; AI does the digging and producers narrow topics
Google AI Studio and Long-Document Analysis
Matt introduces the free AI Studio, where he uses Gemini most, with its million-token context window. He shows pasting a YouTube URL for a summary, the token limit on long videos, and swapping in a transcript to break down a long Naval Ravikant interview.
- •AI Studio is free and offers Gemini 2.5 Pro, Flash and Gemma models
- •The context window is roughly 750,000 words
- •Gemini actually watches pasted YouTube videos rather than reading transcripts
- •Videos over the token limit require pasting the transcript
- •Nothing else handles such large text as well
One Weekend Challenge: Journal Into ChatGPT
Calum asks for one action listeners can take this weekend. Matt recommends using ChatGPT 4o to journal and brain dump about business and life, then seeing the insights it returns. Calum dubs it the Wolfe Challenge.
- •Start with ChatGPT 4o via voice or typing
- •Brain dump what's working, what isn't, and what's causing anxiety
- •It can suggest ways to automate, systematize and reframe
- •Matt calls it the first layer of the onion that shows AI is useful

Publisher timestamps

Get more out of YouTube videos.

Read along on your phone, search it beside the video, or ask your favorite AI app what was said.