The Ledger
Every statement that passed quotation and attribution checks. Mix any filter with any other: certainty 1/5, debate potential 5/5, or both at once.
why aren't all 8 resolved? a statement only gets an assessment when the public record can support or contradict it. opinions and what-ifs never can, and 0 checkable ones are still open, waiting for their date. predictions held up or didn't; assertions are supported or contradicted. on every card: ▮▮▮▮▮ certainty · ▮▮▮▮▮ debate potential. speakers are clickable
Smith: Over 90% of ChatGPT users still use Google
“Similar web I have data showing that the overlap Basically everyone who uses ChatGPT uses Google. Nearly all of them. I think it's like 90, 90% plus or something. So people are not using it instead of search.”
Smith: TechRadar Is Among Most Cited LLM Domains Despite Affiliate Disclosures
“TechRadar is a good example. TechRadar shows up A huge amount. It's one of the most cited domains in LMs. And you'll see a snippet at the top saying we're an affiliate site and it does really well. There's no evidence that a paid for mention does worse than a …”
Smith: Half of newly published internet content is AI-generated
“We just posted an original piece of original research showing that half of the content published on the internet is generated by AI.”
Smith: The llms.txt protocol does not work
“So the intention of LMS.txt is to get more traffic and to send information about, it's not to block, it's to send more information to the LMS so that you get more traffic, but it doesn't work.”
Ethan Smith: LLMs frequently cite YouTube, TikTok, and LinkedIn, but rarely X
“It generally, so LMs, the classes of domain types of domains that are cited frequently, video and UGC are huge.
Instagram, TikTok, LinkedIn, I am not seeing X that much.
And then video and YouTube shows up a lot.”
Smith: Search volume is 5x to 10x larger than LLMs
“I think it's five X, at least five X LMs, maybe 10 X, depending on the category. So it's way bigger than AEO.”