Get Your Weekly Digest

Receive the best podcast summaries and insights delivered to your inbox every week.

We respect your privacy. Unsubscribe anytime.

My First Million

5 AI Tools I'd Use to Make $1M (w/o employees, capital, or time)

Published October 21, 2025
Visit Podcast Website

About This Episode

Sam and Greg walk through a series of AI-powered tools and workflows that they personally use to boost productivity, create content, and make better business and financial decisions. They demonstrate concrete use cases for deepfake-style video generation, AI-first web browsing, voice dictation, AI spreadsheets, TikTok automation, and automated job applications, while also touching on the risks and ethical concerns around scams, spam, and brain-rotting content. The conversation balances excitement about arbitrage opportunities for people going from zero to one with unease about how pervasive and manipulative AI-generated content could become.

Topics Covered

Artificial intelligence Entrepreneurship Content creation User-generated content Online safety Investing Hiring philosophy Marketing and customer education Side hustles

Disclaimer: We provide independent summaries of podcasts and are not affiliated with or endorsed in any way by any podcast or creator. All podcast names and content are the property of their respective owners. The views and opinions expressed within the podcasts belong solely to the original hosts and guests and do not reflect the views or positions of Summapod.

Quick Takeaways

  • AI-powered deepfake tools can now create highly convincing video avatars from consumer hardware, enabling both powerful marketing use cases and serious scam risks.
  • AI-first browsers like Perplexity Comet can observe your screen, manipulate webpages, find codes, surface exact video moments, and even draft and send emails, dramatically compressing many workflows.
  • Voice interfaces such as WhisperFlow let users talk to their computers at over four times normal typing speed while automatically correcting and formatting text, making traditional typing feel slow.
  • AI-native spreadsheet tools like tryshortcut.ai act like an on-demand financial analyst, turning plain-language prompts into analyses, models, and reports directly inside a spreadsheet.
  • Automated TikTok slideshow and UGC-style video generation with tools like real.farm can be used to quickly spin up niche content accounts that funnel attention to software products.
  • Job seekers are already using services like AIApply to auto-apply to massive numbers of jobs and generate interview prep, forcing a coming arms race on the employer side to filter AI-generated applications.
  • Greg frames these tools as a practical kit for people with no special connections or capital to make their first million by exploiting current AI arbitrage opportunities.
  • Both hosts feel tension between excitement about AI leverage and concern that ubiquitous AI content will contribute to scams, low-quality media, and addictive, brain-rotting experiences.
  • Greg spends about 10 hours per week experimenting with new AI apps and has become truly fluent in around 10 of them, seeing this as a key part of staying ahead.

Podcast Notes

Introduction and episode goal: AI tools to get from zero to a million dollars

Framing of the podcast and audience

Podcast identity and purpose[0:00]
• Sam states that the podcast is called "My First Million" and positions it for people trying to get from zero to a million dollars.
• He says the episode will cover tools that he and Greg are using to make a million bucks.
Target listener skill level with tech[0:47]
• Sam describes himself as a "really good Googler" and "pretty good at ChatGPT" but not deeply technical.
• He contrasts himself with Greg, who he says is about three steps more technical and an early adopter, but still able to explain things in simple terms.

Greg's approach to choosing tools for the episode

Accessibility of the tools[0:04]
• Greg says he will try to explain clearly and has picked apps that literally anyone could use that will make a big difference.

Anticipating AI skepticism

AI haters and deepfake backlash[1:15]
• Greg notes there are many AI haters who see things like AI-generated Taylor Swift and accuse users of ruining the world.
• He frames the episode as being for people who want to use this technology to their advantage rather than reject it.

Deepfake video tools and open-source AI avatars

Demonstrating an AI-generated Taylor Swift-style video

Greg's warning about disappearing boundaries between real and fake[2:33]
• Greg narrates over an AI Taylor Swift clip, saying humanity is crossing a line where the boundary between reality and fiction is vanishing.
• He emphasizes this was not created in a Hollywood studio but with open-source tools on a regular computer at the click of a button.
• He advises listeners to be safe, question what they see, and verify what they believe.
Quality of the AI avatar[2:36]
• Greg points out that the AI voice sounds just like Taylor Swift and that the mouth movements match his friend "Serial" whose face was used as reference.
• He says it looks pretty much like Taylor, though Sam comments that the jawline appears stronger than normal, so it is not a perfect likeness.

Underlying model and accessible front-ends

WAN 2.2 model and origin[3:41]
• Greg says the demo is using an open-source model called WAN 2.2 created by people at Alibaba, noting it is a Chinese app and implying users should beware.
Enhancer.ai and FreePick as easier interfaces[4:08]
• Greg explains that many people find open-source tech hard to use via GitHub, so SaaS platforms like Enhancer.ai and FreePick let them access similar capabilities more easily.
• He notes that these apps allow people to pick presets and avoid technical setup.

Use cases and risks of deepfakes

Creating various celebrity or public-figure avatars[5:09]
• Greg says users can record a video of themselves talking and then choose a target character such as Taylor Swift, Mark Zuckerberg, or Jensen Huang to deepfake.
• Sam jokes that the AI looks like "Tyler Swift" rather than Taylor due to the slightly off appearance.
Real-world scam example using deepfake-like tech[3:54]
• Sam recounts a case where a wealthy acquaintance was robbed: criminals learned they were out of town, used a voice/video fake to call the building, claimed a handyman would come, and got access to the apartment.
• He says this type of incident makes deepfakes very nerve-wracking, especially for vulnerable people.
Greg's concern about elderly victims[5:16]
• Greg says he is especially concerned about grandmas and parents getting scammed by such technologies.
• Sam adds that he himself almost gets scammed weekly, so it is not just grandmas who are at risk.

Using AI avatars in ads and content

Current quality is already ad-ready in Greg's view[5:44]
• When Sam asks when this will be good enough to use in ads, Greg answers that it is already 100% good enough.
• He cautions against using fake Taylor Swift in Meta ads due to approval issues but notes that people are using public-domain figures like JFK and Plato successfully.
Example of scripting and production workflow for AI ads[10:31]
• Greg references "PJ Ace" as an AI video ad expert who gets hundreds of millions of views and works with very large companies using AI-generated personas.
• He outlines a workflow: use ChatGPT or Claude to generate a script, use an image model like Enhancer or FreePick to create storyboard frames, then animate those frames to form a 20- to 30-second cohesive video.
• For advanced users, he suggests editing the resulting video in software like Final Cut Pro.

AI-first web browsing with Perplexity Comet

Why switch from Chrome to an AI browser

Sam's initial use of Perplexity[11:48]
• Sam says he recently used Perplexity after watching one of Greg's videos about Sora, using Perplexity for research on making a viral video, then passing that into Claude for the script and into Sora for video creation.
Greg's claim that an AI browser is a game-changer[12:12]
• Greg acknowledges it is a hassle to move browsers but says having an AI-first browser is an "absolute game-changer."
• He notes there are multiple AI browsers (Perplexity Comet, Dia, Opera, others) but he prefers Comet and wants to demonstrate workflows that will convince people they need an AI browser.
Greg fully replaced Chrome with Perplexity Comet[13:29]
• Greg confirms that he uses Perplexity Comet instead of Chrome and says he cannot go back because traditional browsers now feel too slow.
• He compares the shift to using ChatGPT instead of Google for most queries, estimating people may now use ChatGPT about 95% of the time.

Using Comet to find discount codes automatically

Shopping example with clothing brand OneBone[14:01]
• Greg shows an ad he saw for OneBone, a clothing brand for big and tall people, and notes that while he is tall (6'3") he is not big (165 pounds) but clicked anyway.
• He describes getting to the checkout page where there is a discount code field and wanting to find a valid discount code to use.
Prompting Comet to find and apply a working code[15:49]
• Greg opens Comet's Assistant and types a prompt asking it to find a discount code that works and add it to the form.
• He points out that Comet can "see" what is on the screen, searches multiple sources, and then discovers a code "Thrift10" that saves him $66 on two jackets.
• Sam jokes about "girl math," saying Greg has effectively made money by shopping and treating the savings like profit.

Using Comet to navigate and summarize YouTube content

Finding exact moments in videos[18:36]
• Greg demonstrates opening a Steve Jobs YouTube video and asking Assistant to find and play the exact moment where Jobs talks about Apple's intersection of liberal arts and technology.
• He explains that Comet downloads the transcript, searches it, and quickly returns the timestamp where the quote begins, which he says was at the seven-second mark.
• He then instructs Comet to play the video from that point, and it opens a new tab and starts playing automatically.
Maintaining flow state versus manual searching[20:50]
• Sam says that in his normal workflow he might vaguely remember a good hook from a Steve Jobs speech but not know which talk it came from, and manual searching would be slow.
• Greg emphasizes that doing this in a regular browser would require opening multiple tabs, Googling, and scanning, which pulls him out of his flow state when working.
• He credits tricks like this with helping him run five successful companies by saving time.
Extracting lessons and essays from video content[20:46]
• Greg shows that after locating the clip, he can ask Assistant, as a founder, to list the most interesting takeaways from the talk and write them in an essay.
• He says YouTube contains an incredible wealth of information from world leaders, top founders, and podcasts, and tools like this help unlock that content more efficiently.
• Sam mentions he had installed a Chrome plugin to get better YouTube transcripts so he could paste them into ChatGPT, but notes that Comet will make his life much easier.

Using Comet for talent sourcing and outreach

Finding a specific type of LinkedIn profile[23:24]
• Greg says one of the hardest tasks as a founder is hiring good talent and shows a prompt asking Comet to find and go to the LinkedIn profile of someone who worked on Apple AI and is now at Meta.
• Comet returns the profile of Mark Lee, identifying him as a match who worked as an Apple AI research engineer and now is an AI research scientist at Meta.
Drafting and sending a recruiting email automatically[25:14]
• Greg prompts Comet to craft an email that is likely to get Mark Lee to work for him, specifying that he is the CEO of Hampton and providing the URL joinhampton.com.
• Comet generates a cold email introducing Sam as CEO of JoinHampton, describing assembling an elite group, and making a recruiting pitch.
• Greg notes that he can then ask Comet to send the email from his Gmail; Comet spins up an agent to send it once Gmail is connected.
• He does not actually connect or send, joking that he is not Sam Parr and is not trying to spend $100-500 million hiring the candidate.

Voice control and WhisperFlow for faster text input

Sam's workflow with WhisperFlow

Talking to the computer instead of typing[27:21]
• Sam says that, although some might mock him, he uses WhisperFlow and talks to his computer a lot because he hates typing and his fingers sometimes hurt.
• He has a hotkey mapped to the "glow" button on the bottom left of his screen; holding it lets him have a conversation that WhisperFlow transcribes.
Correction and formatting features[29:08]
• Sam explains that if he changes his mind mid-sentence, e.g., "Can you pick me up a burger? I mean, I actually want a taco," WhisperFlow will only output the corrected request.
• He says he can verbally structure multi-point responses (e.g., three points) and WhisperFlow will output a formatted list with numbered paragraphs.

Greg's productivity framing for WhisperFlow

Speed comparison and irreversibility of adoption[28:34]
• Greg remarks that once you start speaking to computers with tools like WhisperFlow, you cannot go back, just as with AI browsers.
• He compares typical keyboard typing at about 45 words per minute versus around 220 words per minute using Flow, calling keyboards too slow afterward.
Using WhisperFlow snippets for teams[31:22]
• Greg suggests using WhisperFlow for teams by leveraging its snippet library, where users can create reusable snippets like "calendar," "hours," "support intro," or "FAQ."
• He notes you can simply say the keyword (e.g., "calendar") and the prewritten text will be inserted, accelerating repetitive communications.
• He calls WhisperFlow a productivity upgrade comparable to moving from Chrome to an AI browser.

AI-powered spreadsheets and financial analysis with tryshortcut.ai

Sam's desire to "talk" to spreadsheets

Existing frustration with Excel and Google Sheets[31:36]
• Sam says he uses Google Sheets rather than Excel and has been trying to find a solution that lets him talk to sheets.
• He mentions demoing a company funded about 12 months ago that attempted this but found it underwhelming and "only five" on a quality scale.

Greg's description of tryshortcut.ai

Positioning as AI-first version of Excel[32:08]
• Greg admits he hates Excel and does not know what a macro is, finding Excel as intimidating as using a terminal.
• He says tryshortcut.ai, built by top-tier researchers, is an AI-first version of Excel where the left side is a cloud spreadsheet and the right side is a prompt pane.
• Users can use natural language to instruct the sheet, similar to how Comet controls web pages.
Example analysis using uploaded financial data[33:22]
• Greg says he asked Comet one minute before the call to find a financial data sheet, then uploaded it to tryshortcut.ai; the sheet includes company names, market cap, revenue, and gross profit.
• He shows that by prompting "analyze my data for key insights" the tool returns all key insights about the dataset.
• He explains that from there he can ask it to build a DCF analysis, an income statement, or other financial statements.

Checking which stock Warren Buffett might pick

Sam's question about Buffett's criteria[34:58]
• Sam asks Greg to query which stock Warren Buffett would pick based on the uploaded data, explaining he once bought a book about Buffett on accounting to understand what Buffett sees in financials.
• He says the book was complicated, implying that AI could simplify extracting Buffett-style insights.
Context mixing: sheet data plus internet[35:22]
• Greg notes that the model not only uses the data on-screen but also sends agents across the internet to figure out what is happening with each company.
• He says that combining local context with external context leads to good outcomes, and mentions that the result in this example is Buffett picking Apple.
• He demonstrates asking the tool to create a new document based on this that shows operating cash flow.

Local Excel plugin versus cloud data concerns

Sam's past practice with ChatGPT and financials[36:38]
• Sam says he has previously uploaded his business P&L or personal net worth into ChatGPT and asked questions based on that data.
• He asks whether Greg would feel comfortable doing that with Comet and whether it would give good advice.
Greg's preference for local use with sensitive data[38:32]
• Greg says he would feel comfortable but would be more comfortable using such tools locally in Excel so data is not uploaded to the cloud.
• He recommends using Excel locally for really sensitive data since most people already use Excel.
Early-stage Excel plugin[36:38]
• Greg mentions that tryshortcut.ai now offers an Excel plugin so users can get AI features even if they avoid the cloud version.
• He notes the plugin only has 36 ratings, indicating how early users are to this technology, and jokes that Sam would be number 37.
• He says the tool is magical and offers a limited number of free credits, around four.

TikTok slideshow and UGC-style content automation with real.farm

Sam's skepticism about AI TikTok content

Initial resistance to TikTok and AI "slop"[40:02]
• Sam says he does not use TikTok and thinks a lot of AI content is garbage; he calls himself a purist who wants high-effort creators like MrBeast.
• Greg predicts Sam will hate the next tool but wants to show it because of an arbitrage opportunity.
Acknowledging some entertaining AI content[40:46]
• Sam admits he has enjoyed some AI-generated content, including a fake "potato lunch" shirt prank on a grandma and a Stephen Hawking rap video.

Strategy: own TikTok accounts to funnel users to software

Arbitrage of AI-driven slideshow accounts[45:11]
• Greg says many people are "vibe coding" software but the hard problem is getting customers.
• He suggests owning numerous TikTok accounts, especially slideshow or meme-style accounts, and using them to direct traffic to software products.
• He believes there is current arbitrage in using AI to create such accounts cheaply and at scale.

How real.farm automates TikTok slideshows

Basic workflow and Pinterest sourcing[45:44]
• Greg explains that real.farm is a tool to automate TikToks to drive traffic to a website, focusing on TikTok but with videos that could be uploaded elsewhere.
• He describes sourcing images from Pinterest for a slideshow like "top 8 protein sources ranked by bioavailability and cost," then adding a prompt and generating the slideshow.
• He notes that TikTok is currently promoting slideshow content, making this format particularly effective.
• Within about 30 seconds, users can create a piece of content in their niche and schedule it for publishing.
Sam's reaction to real.farm and its clients[47:20]
• Sam mentions a friend who built a men's kegel exercise app that is hard to promote because users do not want to publicly brag about using it.
• He says that on real.farm's site he sees logos like a16z and Substack and expresses surprise that such companies would use this tool.
• Despite teasing and mocking some of the AI content, Sam admits that he personally watches slideshow-style content such as "top 10 strangest missing person cases" when they appear in his feeds.

UGC-style AI avatars for ads within real.farm

Template-based AI presenters reviewing products[50:52]
• Greg points out a UGC ad avatar feature in real.farm where users select a template avatar (e.g., a sleepy guy in a hoodie) to present an ad.
• Users write a hook such as "my top three matcha brands," choose background music, and real.farm generates a product-review-style video.
• Sam observes that the generated matcha looks like green paint rather than tea, though Greg notes they are scrutinizing it more carefully than the average viewer would.
• Greg argues that typical users quickly swiping in a "brain rotted" state are unlikely to notice such imperfections and will just consume the content.
Economic value of niche review accounts[53:42]
• Greg asks rhetorically what a TikTok account reviewing matcha brands with thousands of followers might be worth, implying it has real economic value.
• Sam expresses mixed feelings, joking about capitalism and then saying he sometimes feels like becoming a socialist when imagining a world of pervasive AI ads nudging people to buy dumb stuff.

Automated job applications with AIApply and ethical ambivalence

Overview of AIApply features

Auto-applying to thousands of jobs[55:26]
• Greg introduces AIApply, a service where job seekers can automatically apply to thousands of jobs via AI.
• He calls it "absolutely crazy" and highlights it as a powerful tool for job seekers, especially in a tough job market.
Additional capabilities: interview prep and documents[55:38]
• Greg mentions AIApply has an Interview Buddy to help users answer interview questions in real time.
• It can also help build resumes and cover letters, beyond the core auto-apply function.

Fairness for job seekers versus burden on employers

Helping disadvantaged job seekers gain an edge[55:48]
• Greg says he perceives in the tech bubble that the economy seems good, but many people are suffering and having a hard time finding jobs.
• He supports giving job seekers any "unfair advantage" they can get to reach the interview stage.
Downside for HR and company recruiters[57:02]
• Greg acknowledges that AIApply makes life harder for HR managers and companies receiving large numbers of AI-generated applications.
• He notes the product's CMO replied to a tweet saying the next version will help companies filter out AI applications, effectively addressing both sides of the problem.
Scale of adoption and business potential[56:44]
• Greg says AIApply is loved by 1 million users, indicating substantial traction.
• He compares it to Nestlé owning hospitals, joking about a system that fattens people up then sells them medicine and repeats the cycle, implying a potentially problematic but lucrative loop.
• He anticipates AIApply being a great business because it addresses both job seekers and, soon, employers.

Reflections on AI, generational shifts, and Greg's mission

Emotional reactions to pervasive AI tools

Sam's discomfort and generational perspective[57:38]
• Sam admits that although his own business relied on advertising, some of this AI-driven arbitrage content makes him feel bummed out and like older generations complaining about lost craftsmanship.
• He says this is the first time in his life where he feels old enough to have to consciously decide whether to "get on or get off" a technological wave, unlike with the internet where he was born into it.
Greg's focus on zero-to-one builders[58:52]
• Greg reiterates that the podcast is called "My First Million" and says that if someone wants to go from zero to a million, these are tools they can use without capital or needing to be in Silicon Valley.
• He shares that he grew up with no connections and not much money, giving him a soft spot for people starting from zero.
• He says he has dedicated his career to helping such people, describing himself as being in the "light bulb business"-creating lightbulb moments that change life trajectories.

Balancing excitement with responsibility

Historical parallels with past technological shifts[1:00:52]
• Greg argues that older generations likely said similar things about the internet and the industrial revolution, suggesting that upheaval and anxiety are recurring patterns.
• Sam contrasts brain-rotting on TikTok with watching TV for eight hours, implying that similar concerns have existed with prior media.
Concerns about addiction and family exposure[1:00:28]
• Greg notes that he and Sam now have kids and do not want their families to be addicted to AI-laden products, so he feels some responsibility.

Greg's time investment in learning AI tools

Exposure through design agency work[1:02:28]
• Greg says one reason he loves his design agency is that it works with many top AI apps (e.g., Character AI, Jasper AI), so he sits in meetings and asks to be taught everything about these tools.
Weekly experimentation and proficiency level[1:03:28]
• Greg estimates he spends about 10 hours a week playing with new tools every single week.
• He says he is genuinely proficient-"native"-in around 10 AI tools.

Sam's gratitude and closing remarks

Feeling pulled toward the AI epicenter[1:03:52]
• Sam recalls once being the young guy his parents asked for help with tech, but now feels further from the AI epicenter.
• He expresses appreciation that he has friends like Greg who can pull him toward AI, show him what is cool, and not make him or the audience feel stupid.
Greg reiterates his mission and signs off[1:05:06]
• Greg repeats that he is in the light bulb business and hopes listeners had at least one to three lightbulb moments, even if they found some tools cringe.
• He encourages listeners to adapt the tools in ways that work for them, thanks Sam for having him on to share nerdy stuff, and says he will see them next time.
Episode wrap[1:05:28]
• Sam closes by saying, "Alright, that's it, that's the pod."

Lessons Learned

Actionable insights and wisdom you can apply to your business, career, and personal life.

1

AI-first tools that understand context on your screen or in your data (like AI browsers and AI spreadsheets) can collapse multi-step workflows into a single natural-language instruction, giving you leverage that compounds over time.

Reflection Questions:

  • • What recurring digital workflow in your day (research, reporting, outreach) could benefit most from an AI agent that "sees" your screen or data?
  • • How might your weekly output change if tasks that currently take you 30-60 minutes could be handled by an AI tool in under five minutes?
  • • Which one workflow will you commit to redesigning around an AI-first tool in the next two weeks, and what is the first step to do that?
2

Speaking to your computer using high-quality voice interfaces can dramatically increase your effective bandwidth, especially for drafting, outlining, and ideation, making traditional typing a bottleneck.

Reflection Questions:

  • • Where do you notice yourself slowing down or procrastinating because typing or formatting feels like friction?
  • • In what kinds of work (emails, documents, notes) could you safely experiment with voice-driven drafting over the next month?
  • • What simple setup (hotkey, app installation, microphone choice) could you put in place this week to make talking to your computer as easy as opening a browser tab?
3

Current AI content tools create real arbitrage opportunities-such as auto-generated TikTok slideshows or UGC-style reviews-that can cheaply capture attention and route it to products before the market fully saturates.

Reflection Questions:

  • • Which niche audiences are you trying to reach that already consume short-form or slideshow-style content?
  • • How could you test an AI-generated content channel for your product with a constrained experiment (e.g., 10 videos over two weeks) instead of a big, risky bet?
  • • What metrics (click-throughs, email signups, trials) would you track to know whether an AI-driven content funnel is worth scaling for your business?
4

Tools that level the playing field-like automated job applications or AI-driven recruiting and research-can be powerful equalizers for individuals without connections, but they also increase noise and force everyone to raise their bar.

Reflection Questions:

  • • In what ways are you currently at a structural disadvantage (where you live, who you know, how much capital you have) that AI tools could help offset?
  • • How might the increased "noise" from AI-generated resumes, emails, or pitches require you to differentiate yourself more clearly in your domain?
  • • What is one concrete way you could use AI to open doors (jobs, clients, intros) that currently feel closed, while still preserving authenticity in how you show up?
5

Staying ahead in a rapidly shifting AI landscape requires deliberate time investment in exploration and practice so that you become truly fluent in a small set of high-impact tools rather than dabbling shallowly in many.

Reflection Questions:

  • • How many hours per week do you currently devote to actively learning and experimenting with new tools versus just consuming content about them?
  • • Which two or three AI tools, if you mastered them over the next six months, would most dramatically improve your earning potential or effectiveness?
  • • What recurring block of time can you schedule in the coming weeks specifically for hands-on experimentation with one chosen tool, and how will you measure your learning?

Episode Summary - Notes by Harper

5 AI Tools I'd Use to Make $1M (w/o employees, capital, or time)
0:00 0:00