Episode Overview

Podcast: The My Wife Quit Her Job Podcast With Steve Chou
Episode: 618: AI Tools Are Getting Scary Good – Here’s The Latest With Toni Herrbach
Date: December 10, 2025
Guests: Steve Chou (Host), Toni Herrbach (Guest)

Main Theme:
Steve and Toni dive into the rapid evolution and recent breakthroughs in AI tools, particularly those relevant to ecommerce and online business. They discuss how AI is changing user behavior, privacy perceptions, marketing, content creation, search, and workflow automation. The conversation also explores public skepticism around AI, the tension between automation and personalization, and practical ways entrepreneurs can leverage AI to gain a competitive edge.

Key Discussion Points & Insights

1. The Pace and Impact of Recent AI Announcements

• AI Tool Surge: Steve notes continuous, rapid announcements of new AI tools and capabilities nearly every week, making it challenging but exciting to keep up.
  • Steve (01:20): “These past couple weeks I feel like there's been this huge bonanza of AI announcements… every week a new tool, a new version of something comes out and it's very impressive.”

2. Real-World Use of AI: Technologists vs. General Public

• Different Comfort Levels: Steve and Toni share that while they and their professional circles embrace AI for complex business applications, many non-tech friends (doctors, lawyers, etc.) are skeptical, fearing job displacement and privacy risks.
• Privacy Hesitation:
  • Toni (04:23): “My friends were like, are you kidding me? No way. I would never put my medical information into AI because then it's out in the world... But you're right. Like, my friends are all like, oh, I don't really want AI knowing a lot about me.”
• Contrast in Behavior: Outside of tech/business circles, many are not even on social media, illustrating a broader public lag in adopting new technologies (05:06).

3. AI and Personalization: The Promise and the Fear

• AI-Powered Shopping: Discussion on OpenAI’s integration with platforms like Etsy and soon Shopify, raising questions about hyper-targeted recommendations based on private data (05:24).
• Comfort with Targeting:
  • Toni (05:56): “I want AI to know more about me so that I get better tailored responses... Serve me all the ads of things that I want... But I think we probably skew way to the other side.”

4. The Future of Search and Ads: Google Gemini’s Role

• Gemini’s Disruption: Google’s Gemini AI is set to transform how ads and search results are delivered, potentially overtaking traditional search formats (06:45–08:31).
  • Steve (08:00): “Google's going away like the traditional search and then Gemini, which was just announced is gotten really good reviews... I imagine will happen is that Gemini will take over because Google has like this huge installed base.”
• Amazon’s Vulnerability: Discussion on how AI search could cut into Amazon's ad revenue and dominance in product discovery (09:30).
• Meta (Facebook/IG) Less Impacted: The panel sees social platforms retaining their ad niches despite these changes.

5. AI for Content Creation and Analytics

• YouTube Analytics: Kevin, a friend of the hosts, uses AI to summarize and analyze complex YouTube analytics, saving hours of manual work (10:58–12:21).
  • Toni (11:59): “What was probably a two to three hour job... I had not everything I needed, but like a pretty good start in under 30 minutes.”
• Gemini for Video Critique: New features let users upload videos for AI critique, showing the potential for AI to automate and improve creative feedback loops (12:28–13:14).

6. The Manipulation and Validity of AI Training Data

• Reddit’s Role and Issues: The co-hosts discuss information manipulation on Reddit—“gaming Reddit”—and growing distrust in its reliability as an AI training source (13:27–14:35).
  • Steve (13:36): “All sorts of people are gaming Reddit right now... So I don't trust Reddit at all.”
• China’s Regulations: China’s efforts to restrict expertise on certain topics (health/finance) to credentialed posters highlights global responses to misinformation risks (14:36).

7. AI Content Proliferation and Human Creativity

• AI Video Channels: The ease of generating faceless, AI-powered TikTok channels has led to a glut of generic content, raising questions about quality and authenticity.
  • Steve (17:01): “If everyone just starts pushing all this AI stuff... I'm going to want to see human. I already want to see human.”
• Human vs. AI Scripts: Experienced users can often tell when a script is AI-generated due to repetitive phrasing (“chat has its favorite, like, sayings and words”) (19:42).

8. Emerging AI Video Ad Tools

• Rapid Content Production: Tools are being developed to stitch together fully automated video ads and UGC-style content, with implications for both genuine brands and spam (21:17–21:50).

9. AI-Generated Music

• Music as a Neutral Ground: Acceptance of AI in music seems higher if the tunes are good—people care less about authorship in this context (22:10–23:41).
  • Steve (22:38): “...If it puts out good music that I just want to listen to while I work or whatnot, I wouldn't care whether it's human or AI.”

10. Personalization Pitfalls in Ecommerce

• Collecting Data vs. Acting on It: Toni recounts an email marketing example where customer data was collected but not used to personalize messages—a wasted opportunity (26:43–28:55).
  • Toni (28:33): “It's like you're collecting this data about me, but you're not doing anything with it... If you're not taking the time to act on it, then you're kind of wasting your time on the front end.”

11. How AI Is Changing (and Ruining) Email Outreach

• Email Fatigue: Easy-to-use AI tools have vastly increased cold email, making even poorly targeted emails look professional—but also causing users to tune out (29:08–29:37).
  • Steve (29:22): “I don't know about you, but I've been getting a lot more email like in like the last year. Just random emails that are clearly AI written.”

12. Privacy, Security, and Agentic Browsers

• Risks of New Browsers: Agentic browsers can unknowingly leak sensitive info if prompted by hidden prompts on malicious websites; security is a major concern (29:37–31:15).
  • Steve (29:37): “The problem with using those browsers is that if you visit a website that has like hidden prompts in it... The agentic browser will actually send that information out.”

13. AI Scams and Enhanced Phishing

• Smarter Scams: AI makes scam emails and messages much more convincing, potentially tricking more victims, especially less tech-savvy users (31:15–32:41).
  • Toni (31:15): “So I'm like, where did this, this come from? Right. Like people are just emailing like crazy. But I feel like what AI does in a negative way is it makes all those spam emails better, right?”

14. Public Sentiment vs. Business Adoption

• Youth Embrace, Older Skepticism: Younger generations are more positive about AI, while much of the public remains cautious (33:12).
• Lack of Positive PR for AI: Steve laments that AI's transformative upsides are not effectively communicated to the mainstream (32:41).

15. Fabrication and Animation of Images/Video (Vibe Coding)

• Ethical Questions: Using AI to animate still images for storytelling and ads blurs lines about authenticity (34:07–35:38).
  • Steve (34:15): “It did such a good job. It looked real, actually. And so I included in the video. It's just crazy what you can fabricate now.”

16. Fake Testimonials & Avatar Selling

• Deepfakes for Ads: AI tools can easily create fake customer testimonials and realistic-looking avatars, raising new issues of authenticity in marketing (35:39–37:47).
  • Toni (35:53): “It holds more weight than if I read it. Right. Like if I talk to you and you're like, oh, I went to that hotel in Vegas... That holds more weight than like me just reading it... So are you—I don't see. I don't know.”

17. Invasion of "Live" AI Avatars

• AI as Host/Performer: In Asia, always-on AI avatars are already live selling products 24/7 on platforms like TikTok, making millions—and most people can't even tell they're not human (38:17–39:07).

18. Workflow Automation Dreams—What AI Can (and Can’t) Do

• Voice and Video Automation Challenges: Steve’s experiments with 11 Labs and avatar clone videos demonstrate the limits of current AI for natural, human-like narration and presentation (39:38–41:15).
• Short Form Video: Efforts to “auto-produce” content often fail due to lack of nuance/emotion; hosts prefer to keep their creative “human touch.”

19. What Hosts Want from AI in 2026

• Steve: Focuses on workflow automation, especially ad creatives, but is wary of using AI for organic content creation due to noticeable “AI-ness.” (41:45–43:03)
• Toni: Excited to level up “vibe coding” for rapid prototyping, drawing parallels to how she once wanted to master Excel and now leans on AI to close knowledge gaps (43:19–44:49).
  • Toni (43:35): “I feel like the gap is getting narrower... The more time I spend... It will be the same outcome.”

20. AI: A Learning Tool Versus a Shortcut

• Learning by Doing: Both hosts reject the idea that AI necessarily “makes people dumber;” for them, it sparks deeper thinking about solutions and structures (47:29–48:46).

  • Toni (47:36): “It made me think through the entire process... for me, it makes me stop and think a lot harder...”

• Experimentation for All: Closing, both stress the need for everyone, especially those in online business, to get hands-on with AI, predicting its continued, transformative impact on business models and productivity (49:37–50:09).

Notable Quotes & Memorable Moments

Human vs. AI Attitudes

• Toni (04:23): “My friends were like, are you kidding me? No way. I would never put my medical information into AI because then it's out in the world.”
• Steve (05:24): “AI recently announced shopping... your medical report, like, all that stuff is fair game, right?... OpenAI might recommend you fish oil supplements.”

On Overwhelming AI Content

• Steve (17:52): “I already want to see human. I've actually started marking a lot of these videos as spam, actually.”

On AI’s Impact on Email

• Steve (29:22): “I've been getting a lot more email… Just random emails that are clearly AI written... now they're all kind of well written... to the point now where I'm like desensitized... I just close it and don't even read.”

Security & New Threats

• Steve (29:37): “Agentic browsers will actually send that information out. There’s so many security holes right now...”

How AI Changes Learning

• Toni (47:36): “For me, it makes me stop and think a lot harder about things because also I hate, I hate waiting for chat to figure stuff out. And if you give it a bad piece of information, then you’re waiting for like three minutes.”

Timestamps for Key Topics

| Segment | Timestamp | Topic Summary | |------------------------------|---------------|-------------------------------------------------------| | AI Announcements Bonanza | 01:20 | "Huge bonanza of AI announcements" | | Real-World Friends & AI | 01:42–05:06 | Technologists vs. public; privacy apprehensions | | AI Personalization | 05:24–06:29 | Shopping, recommendations, privacy divide | | Google Gemini & Search Ads | 06:45–09:30 | Gemini’s role; Ads/Meta/Amazon impact | | AI for Analytics & Video | 10:58–13:14 | YouTube, VidIq, Gemini critique features | | AI Training Data Problems | 13:27–14:35 | Gaming Reddit, distrust in training data | | Tiktok/AI Video Content | 15:00–17:52 | AI-powered/automated content; human vs. AI channels | | Automation Tools for Video | 21:17–21:50 | Tools for auto-generating ads/video content | | AI-Generated Music | 22:10–23:41 | Music as a test ground for AI creativity | | Personalization Pitfalls | 26:43–28:55 | Email flows, collecting data but missing the mark | | Agentic Browser Security | 29:37–31:15 | Browser-based AI, hack risk, vendor bug bounties | | Enhanced AI-Driven Scams | 31:15–32:41 | How AI makes phishing and spam more realistic | | Vibe Coding & App Building | 34:07–35:38 | Animating images, fabricating stories | | Fake Testimonials/Reviews | 35:39–37:47 | Deepfake advertising, blurring lines of authenticity | | 24/7 AI Avatar Live Selling | 37:56–39:07 | Asian markets, AI avatars, live social commerce | | Future AI Projects | 39:38–43:03 | Hosts' plans for AI, what excites and worries them | | AI as a Learning Tool | 43:19–48:46 | Vibe coding, closing the tech skills gap | | Practical Experimentation | 49:37–end | Hands-on encouragement, inevitable AI integration |

Tone & Style

The episode maintains the conversational, witty, and lightly skeptical tone characteristic of Steve and Toni—balancing awe at AI’s progress with pragmatic concerns and a healthy dose of humor about tech adoption, hustle culture, and the intricacies of running digital businesses.

Summary Takeaways

• AI is evolving at an unprecedented pace—business owners who don't engage risk being left behind.
• Privacy and authenticity are major concerns; even as AI tools get better, skepticism outside tech/biz circles persists.
• AI’s greatest strengths today: workflow automation, rapid content creation, analytic insight, and hyper-personalized recommendations—if used responsibly.
• Some AI-driven trends (spammy emails, generic content, manipulated data) risk harming trust and user experience.
• Future jazz: Prepare for a world of live AI avatars, deepfake testimonials, fully animated B-roll, personalized shopping—alongside the next wave of security risks.

Final Note:
Steve (50:09): “If my 80 year old mom is using it, there is literally no excuse.”

For More:
Episode resources and the free ecommerce mini-course are at mywifequitterjob.com/episode618.
Seller Summit 2026 tickets at sellersummit.com for those wanting to network and learn tactical strategies in person.