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 15 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
Cheung: Sending Calendly links is often perceived as disrespectful
“And I know there's this weird stigma around Calendly. I don't know how many people who watch your podcast know about this. I don't know why. I don't agree with it. But like if you send a Calendly link, it's kind of known, I don't know, it's kind of disrespectf…”
Cheung: 1% AI Agent Error Rate Is Too Risky for Enterprise Inbound
“I'm just a little weary of using agents right now because There are like occasional times where like it could go wrong, and just that like one percent Knowing that it could go wrong one percent just worries me for those like really high tier clients.”
Cheung: The Rundown's 15-Person Team Operates Like 50 People
“We have a team about 15. I think we operate as like a 50 person team in terms of speed, but in some cases we might actually operate faster because like the headache of having 50 people is like really burdensome.”
Cheung: The Rundown Gained 150K+ Followers In One Year Using AI Avatars
“We started doing it and it's been under a year and we've grown to over a 150,000 followers.”
Cheung: Limiting AI Avatars to Three Seconds Maximizes Visual Realism
“The avatar is only really shown for the first like three, four seconds. Cause the avatar portion of it isn't that great, but if you just show like 3:04 seconds, it looks more real.”
Cheung: ElevenLabs Direct Generation Outperforms HeyGen's Native Voice Integration
“HeyGen has like an 11 Labs integration. But you actually want to do it separately, because if you do it on 11 Labs, you get higher quality and you just stitch it together.”
Rowan Cheung: Human Authenticity Is the Moat Against AI Content
“Because like, as we get into the age of AI generated content and there's just like seas of AI generated content, you want to push for like that human layer. Cause that's really your moat.”
Cheung: AI Exaggerates And Ultimately Fails At Delivering Authentic Human Insight
“The insight part, AI sucks at giving like human insight and experience, comparing certain things to like other models, for example, and just like tweaking up little things that just don't seem right. Like, Claude likes to exaggerate a lot. Like I say, like wor…”
Cheung: Co-editing with AI increases editing time but produces superior content
“It actually takes more time to edit with AI as a co-editor, but your content gets so good. Because it's trained on your best writing and you're just going to find these small details that you wouldn't have otherwise seen.”
Cheung: The Rundown requires employees to query GPT before asking humans
“Basically the model then becomes don't ask questions. You haven't asked GPT first.”
Cheung: The Rundown drafts newsletters manually, using Claude as editor-in-chief
“When we write our newsletter, we write it all by hand, But we like to use Claude as like an editor in chief.”
Cheung: The Rundown built custom Lindy scheduling agent Ava
“Lindy actually provides like a template for this, but I've modified it a little bit because I wanted like my workspace email. So I actually created an agent and I called it Ava. So Lindy has this, like they give this own template and they use like their email …”
The Rundown Built A Custom Internal GPT Trained On HR And SOPs
“We built this GPT that's trained on SOPs literally loom transcripts, meeting notes, our policies on AI usage, HR docs, any like admin really related question. And it also knows everyone on the team and their role.”
Cheung: The Rundown Landed Two Major Sponsors Via AI Inbox Triage Agent
“And we've actually landed two sponsors through this. Like, it's really weird that they don't go through actual sponsor form, but these are legit sponsors from like big companies that went through a support inbox.”
Rowan Cheung Uses ChatGPT Voice While Reading To Boost Concept Retention
“So for example, I use ChatGPT a lot or ChatGPT voice when I'm reading. So I'll just have it open next to me. And whenever there's a topic I don't really understand, I'll just ask it, oh, hey, what does this mean? And I'll go deeper with it before I move on to …”