PodcastAugust 2026
SEO Is Officially Dead
Hosted by Born In Silicon Valley (Jake Aaron Villarreal)
Episode 301 of Born In Silicon Valley, recorded in San Francisco the week after SaaStr. Jake Aaron Villarreal works through whether a company needs a decade of SEO equity to show up in AI answers, where the line between AEO and GEO actually falls, and what changes inside a company when the growth channel stops being a search results page.
Watch or listen
Also available on Amazon Music, which have no episode-level player to embed:
Key takeaways
- Domain age is not the moat it was. Customers who launched six months ago outrank competitors who have been publishing for ten years, because focus and clarity matter more to a model than accumulated backlinks.
- AEO and GEO are not synonyms. AEO is whether your own site gets cited. GEO is that plus the third-party network around your brand: Reddit, LinkedIn, review sites, YouTube, forums.
- Inconsistent messaging across platforms is the worst failure mode. Twenty different descriptions of the same product means the model cannot resolve what the company does, so it recommends someone else.
- Prompt data is purchased, not guessed. GrackerAI sources real user prompts from roughly 27 data providers and integrates with Search Console, Bing Webmaster, Ahrefs, and Semrush rather than inventing prompts for customers.
- Hiring criteria flipped. The previous company screened for protocol depth and years of coding. Now the interview asks whether the candidate has actually built with AI-assisted tools and can translate a requirement into something a model can execute.
- "AI hoarding" is a culture problem, not a tooling one. The fix is a standing session every other week where the team shares new prompts, tools, and skills, plus a dedicated channel for it.
- The scarce role is the GTM engineer, not another backend engineer. Building a product got dramatically cheaper; getting it in front of buyers did not.
- The platform is agentic because it has to be: roughly 45 to 47 agents, each with one job, because a system that sources from many third parties cannot be a static request-and-response API.
In his words
“Even as a consumer, if you want to find a restaurant, you go to ChatGPT or Gemini and say give me the top 10 restaurants. Nobody is using Google anymore.”
“If you have different messaging on different platforms, the AI is going to get confused about what the company does. That is one of the worst scenarios.”
“Every single day when I talk to customers, the only thing they say is: Deepak, can you do the GEO for me that you have done for your own company?”
“We don't know when Google might turn off normal search. Maybe a year, maybe two, maybe three. But they are going to fully transition into the chatbot model.”
Quotes and chapter marks are transcribed from the episode audio and lightly edited for clarity. Nothing is paraphrased.
Jump to a section
Timestamps open the YouTube cut at that point.
- 1:37From software engineer to serial founder
- 4:46Can a brand new company rank in LLMs?
- 7:10How much of this is technical?
- 7:57AEO vs GEO, defined
- 10:22Spotting the shift two and a half years early
- 11:57Why vertical beats horizontal in GEO
- 15:08Business model and prompt monitoring
- 18:17Hiring for AI fluency, not protocol depth
- 21:27AI hoarding and the culture fix
- 26:09Eating your own dog food
- 29:24Why GTM engineers are the scarce role
- 31:49Why the platform had to be agentic