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 191 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
Thummala: AI enterprise buyers now demand specific workflows and ROI
“In the last year, people are now sharpening up their budgets. They're sharpening their pencils, and they're like, okay, we know AI is effective. We know it's a hundred percent going to be transformative. How? Like, what are the specific workflows, outcomes, RO…”
Matze: AI coding tools produce errors without architectural understanding
“The thing that people have to understand is that with cloud code or any of these coding tools, they're just that they're tools. If you don't know how to swing a hammer, the nails could go in a bent. If you know how to swing a hammer, the nails to go in straigh…”
Apfelbaum: Internet users prefer done-for-you services over software tools
“The headshot software that I built is actually a service for people because people don't want to use software. People, especially people on the internet, what they want to do is they want you to just deliver headshots for them.”
Apfelbaum: People Will Pay to Learn Basic Prompting on Suno and Sora
“People will pay you to teach them how to make AI music. People will pay you to put in a prompt into Suno and teach you how to make AI music. People will pay you to teach them how to go into Sora and make a prompt and make a video on Sora too. People will pay y…”
Apfelbaum: Front-End JavaScript Prompt Generators Eliminate AI API Costs
“And by the way, it's not even using AI credits cause it's all JS. It's all JavaScript. So you create a framework that you don't even need a backend for that you sell access to that they're either paying monthly, 10 bucks a month, or they're paying a lifetime a…”
Santos: Giving people creation tools creates novel behaviors, not linear extrapolations
“And so I think if you give people the tools to creation, they will do things that are not necessarily an extrapolation of what we do today.”
Vice Santos: Multi-character consistency in generative AI is exponentially harder
“If you look at most comparable products out there usually they have one character just because the whole thing about having, if it's already hard to keep consistent with one character character, it becomes actually exponentially harder with multiple, not becau…”
Santos: AI tools make product distribution harder and more critical than execution
“Particularly in the time where you have all these tools like you know, vibe coding, lovable, so on, so forth, where everyone can create the product. You can, you know, it, it's a great time in the sense that, you know, if you ever, you know, if you were ever t…”
Foster: AI Enables Small Businesses to Automate Previously Cost-Prohibitive Tasks
“AI allows you to put brains in your business in places that was really hard to put brains in the past. It was very manual, very expensive, and mostly, especially for small business, you're just like, I'm just not going to do it.”
Foster: AI is valuable for building workflows because people struggle with systems
“This is the other place where AI is really exciting, is it's not just helpful as a tool inside your workflows. It's also really helpful for building workflows, because building workflows is kind of hard. Like, not everyone thinks in systems, thinks in structur…”
Foster: Giving AI agents more tools increases their error rates
“The more tools you give it, The harder it thinks, the more mistakes it's gonna make.”
Foster: Speed-to-lead automation increases conversion rates to box out competitors
“Speed to lead. You, it improves your conversion rate massively. And when you improve your conversion rates, it means you can now be more efficient in how you deploy your advertising dollars, because now you're getting a higher customer lifetime value, higher A…”
Foster: Top automators route queries across ChatGPT, Claude, and Gemini
“What I see the best automators do is they know ChatGPT is really good at this type of queries. Claude is really good at these types of queries. Gemini's really good at this type of stuff, and so they're using Zapier to say, hey, go route this request here, and…”
Foster: AI makes an exceptionally good coach because it is neutral and patient
“I have found that AI is an exceptionally good coach. It is infinitely patient. It is neutral. It just observes what it sees and it tells you what it thinks it has observed.”
Hastings: Accounting is largely understanding business context, not just numbers
“People think that accounting is just numbers. Actually, so much of it is just understanding what's going on in the business. And so much of that is actually human written.”
Hastings: Waiting for monthly financial close is too slow for decision-making
“If you are waiting until the end of the month to understand what your finances are, the time has already passed to make the decisions that are important. You really need to be making decisions on a day-to-day basis, and that requires understanding the fundamen…”
Hastings: Hybrid software-services model is best for launching AI startups
“Well, that's one reason that the hybrid software and services model is a great way to get started on a business like ours because that last mile edge case, we do have humans do, and it's helpful for onboarding and understanding the, our customers' businesses, …”
Hastings: Founders must do manual dirty work to build great products
“And I think that, by the way, I think that's the best way to just build a great product. Just really see the dirty parts of it instead of just handing it off to someone else.”
Hastings: Bookkeepers misclassify LLM API costs into OpEx instead of COGS
“The past 15 years of SaaS businesses, everyone just assumes that their margins are incredible, but now companies are spending so much on the LLM API providers, and because traditional bookkeepers don't understand those new vendors and how to treat them, they o…”
Hastings: AI software margin profiles fundamentally differ from traditional SaaS
“Really becoming a requirement in the AI era because your costs are so volatile and changing every day, and the margin profiles of software businesses have completely changed with how expensive their costs are and how those costs scale when their users use them…”
Hastings outlines a 3-step playbook for AI automation startups
“I would say for anyone looking to use AI to start a company to displace what was traditionally done very manually by humans, first, see what they're doing really understand it then, two, do it yourself. As you bring on your first customers. So do the work your…”
Tossell: No-code tool workflows are fundamentally programming in a visual format
“The thing I didn't realize until, I don't know whether it was within the MakerPad life or after it, was like, I'm still teaching programming, but in a different format. Like, there's different, the tools just look different. Like, they don't look like a code b…”
Tossell: Generate agent instructions by asking AI to document successful executions
“So you can see all the different tool calls and things it did. So I kind of just copied all of that and then said to Droid, like, give me step by step instructions exactly on what you just performed to get to this output. I said, okay, great. Now put that into…”
Tossell: Community leaders must ask basic questions to drive engagement
“When anyone asks, asks me for advice on community, the number one, I think, number one thing I say is, you're going to lead by example. Like, if you're not in there talking, no one else is going to be in there talking. If you're not in there asking seemingly s…”
Tossell: Raw screen recordings outperform polished corporate videos on Twitter
“I think the best videos on Twitter are the ones that are just, hey, I've just clicked record, I'm just doing this thing, I've just posted it, and that's kind of it. And like, I like that. That feels like that's a community member talking to a community member.…”
Singla: Standalone AI tools must integrate deeply into workflows to compound growth
“Just standalone tool won't help. You have to intricate this Deeply into the workflows. So like I've been giving examples this entire podcast about make it workflow oriented, which can give value to the customer so that it can compound growth. You don't solve f…”
Ethan Smith: Video marketing works best for unentertaining, high-LTV B2B products
“And the less entertaining your product is, the better, the more interesting video is.
And the reason why I say that is because video is made to be entertaining and entertaining categories like food, travel, beauty are the most crowded and the least entertainin…”
Smith: Help center optimization is critical for AEO, unlike SEO
“Help center is actually a great place to do that because help center is you talking about how your product works and what it does when it does not do. And so help center optimization is not something you would focus on for SEO, but you would for answer engine …”
Smith: AEO does not require verbatim exact-match pages for every query
“Yes, you do not need to create millions of pages for every single question. You need to generally answer the question... It's grouping and theming them. And the AI is indeed smart enough to just find the answer. So you don't need it explicitly verbatim exact m…”
Ethan Smith: AEO testing requires control groups because LLM growth inflates baseline metrics
“The reason why the control matters is because LM's are growing a lot, and charts are just going to naturally go up because they're growing a lot. So if you don't have a control, you don't know whether or not you caused an effect.”
Pujji: Output-based pricing requires narrow focus to protect margins
“When you think about costing based on outputs, what you're going to find pretty soon is you're going to want to narrow the things you start to focus on, like your example of the SDR thing, because then, you know, you can keep delivering it.”
Shipper: Prompting LLMs on internal meeting transcripts generates honest company handbooks
“I have a recipe on there that I made that I really liked it, that was really cool, which is, Go through all of my internal meetings and write our company handbook from what you see. And I want you to capture all of the unspoken ways that every works as a compa…”
Shipper: Effective ghostwriters interview subjects instead of instantly generating output
“Good ghost writers. When you say, Hey, I want you to tweet about this or blog post about this ghost, ghost writers. Don't go ahead and say, great. I did it. Here it is. What they do is they say, okay, well, what was interesting to you about this article? What …”
Shipper: AI evaluations smuggle in human intelligence through task setup
“And to do all of the things you need to do in order to set up the eval is a job. And so the eval is what I'm saying is smuggling a lot of human intelligence into it because in order to take a like complex and dynamic activity, like doing a job and turn it into…”
Patel: Google AI Overviews favor concise answers over multi-paragraph text
“If you look at AI overviews, the last thing that they want to do is answer the question with a three paragraph answer. It defeats the purpose of an AI overview, just giving you the answer as quick as possible.”
Patel: Structured table comparisons help websites rank better in LLM citations
“And let's say if you want to compare best CRNs for construction companies and you write your own article for that on your own blog and you go super in depth, that's a great way for you to drive more signups, right? And you would do it in a table format. Pricin…”
Patel: AI products suffer higher churn without direct financial ROI
“If you can't help people save money, Why should they pay you? If you can't help them make money, why should they pay you? Because what we see is, if you can't help with those two things, you can still get signups, but your churn tends to be higher.”
Patel: Centralized Business Dashboards Fail Without Prescriptive Actionable Insights
“Telling people all your data in one place. Cool. No one really will log in or do anything with it. That's the real reality.”
Patel: Boutique AI agencies must verticalize to compete against legacy consultancies
“If you want to do well as a smaller business, we think it is a very tough market to get into unless you have a lot of connections on the sea level and you're better off verticalizing and going after specific categories like the medical industry, like plastic s…”
Shim: Basic transcription and summarization are commoditized table stakes
“It's very easy to build a basic model. So transcription, you can do open source. You can buy some stuff. You can drop it into chat GPT or Anthropic and get a summary against it. And those are good. And you'll see there's this kind of table stakes of you should…”
Shim: Chronological meeting summaries are ineffective
“So when you get notes that summarize it in chronological order, those aren't effective notes.”
Shim: Real-time sentiment metrics cause cognitive overload for junior sellers
“The second thing was after the fact, taking those coaching metrics, because a lot of time it is going to be too much cognitive overload, especially for a more junior seller where they're not used to taking all these inputs because you've got a PowerPoint prese…”
Shim: Frequent interruptions after five minutes indicate poor conversational synchrony
“You and I might interrupt each other because we're still getting a cadence. We call it synchrony, but after five minutes, if you're still interrupting each other, that's almost intentional. That's, it's not a good conversation. You both aren't paying attention…”
Rob Howard: Private practice doctors are prime early adopters for AI consulting
“Somebody who has a limited amount of time, they're essentially a solopreneur or small business owner as a private practice doctor. They've got some money to invest in speed. And automation, and they value their time a lot. They were very early adopters, even t…”
Howard: The primary value of RAG chatbots is internal team drafting
“Where we find that chatbots and RAG, which is essentially Taking documents and putting them into a vector database so AI can read them, right? AI that can read your documents in a custom way is RAG. When you think about a chatbot in RAG, the value there is giv…”
Rob Howard: AI automations do not require complex integrations to deliver ROI
“You don't even need it to be a super fancy automation because the amount of time that's being saved, even with the final copy step, and you want the human in the loop anyway, to do that review, you're still saving such a huge amount of time that it's worth it …”
Howard: Software founders face a squeeze from Big Tech and weekend coders
“Anybody who's spent 24 working hours with Claude Code can also build a prototype of your thing pretty much over a weekend as one person, right? So you've got competition from the big players. But you've also got competition from the tiny do-it-yourself players…”
Howard: Bespoke AI consultants are insulated from big tech commoditization
“The other thing is that consultants who are building bespoke projects for clients repeatedly are not exposed To those same challenges, because by virtue of hiring a consultant, the client is essentially saying the Google Gemini built-in tool is not appropriate…”