No Brainer

There is a genuine cluster of useful tactical points - LLMs relying on mentions not links, the semantic triple concept, the ziptie.dev tracking insight for logged-in users, and SparkToro's audience-AI-usage mapping - but the episode is heavily padded with banter, tangents, and broad validation of obvious points. The ratio of signal to filler is mediocre for a 48-minute episode.

“the LLMs rely on mentions, not links. So you need to make sure that you've got your semantic triple”

“ziptie.dev was the first on the market and they have figured out a way to track logged in Google users...you're going to see seven times more AI reviews popping up”

The core contrarian claim - that AEO/GEO is 'horseshit' rebranding and it's still fundamentally SEO - is refreshing and practically useful, but most of the surrounding advice (allow crawling, HTML over JS, structured data, clean online presence) is recycled SEO 101. The user-embedding framing is mildly novel but not developed rigorously.

“I think it's horseshit. I don't know people who created those terms what they think the difference is”

“Google has 373 times more traffic than ChatGPT right now. So if we're talking about traffic, we're still talking about Google”

Katherine Ong is a genuine practitioner with a decade running her own boutique consultancy, a Ketchum background, and named client work at cancer.gov and National Academy of Sciences; she references real tools and real client data, which puts her well above the podcast-circuit thought-leader category, though she operates at boutique scale rather than enterprise.

“she helped one of my former brands really crush it. She helped us increase traffic by 300%”

“I have a client that wants their content stolen. They're an open data platform”

The episode earns credit for citing specific numbers (373x Google vs ChatGPT traffic, 67% cookie blocking, 7x AI overview undercount, 3% AI referral share, $12k/year user testing cost, 2% AI mode usage in iPull Rank study) and naming real tools and clients, though some figures are asserted without sourcing and several claims remain hand-wavy.

“2% of the people actually used it and most people ditched and didn't stay for very long”

“It's like 12,000 a year or something”

The hosts occasionally push productively - asking about traffic drop attribution, measurement under data loss, and the CEO personalization trap - but most questions are soft setups and the episode is dragged down by sustained off-topic banter, political asides, and the hosts frequently answering their own questions rather than pressing the guest.

“How much of that is due to the rise of AI versus some other wide range of potential?”

“to what extent are these AI answers just confirming what an individual already knows or parroting back their behavior?”