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 2,724 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
Lemkin: AI agent prompt portability drastically lowers enterprise switching costs
“Agents will actually may make our moats even weaker and shallower because if the agents can just take prompt, if I can take a prompt from one app and give it to another app, my switching costs have gone down radically.”
Lemkin: AI sales agents only scale what humans have already nailed
“The way an AI agent works for sales for GTM is you figure out something that works with your team with humans. You figure out an email, an outreach, a script, a set of objections that you overcome, a set of questions. You nail it yourself, just like it has alw…”
Lemkin: AI SDRs Win Through 24/7 Consistency, Not Flawless Prose
“The real bar is better than the average human with 24, seven consistency. And that's which, that's what you're going to see in these emails, like pretty good. With not a lot of errors. That's good enough to crush it.”
Lemkin: Great AI often adds more value than a mediocre human
“And I really think, especially in 20, 26, we're all going to work in these heterogeneous worlds where we mix AIs and humans and imagining that I somehow mind talking to a great AI is backwards. A great AI is often going to add more value than a mediocre human.”
Lemkin: Never buy AI software without dedicated vendor deployment support
“If you can't get that help. Don't buy the vendor. Don't buy the vendor, no matter how slick the salesperson is AI or human. If you're not going to get the help to train and deploy. I'd rather have a worse vendor that will give me the help day and night.”
Lemkin: AI sales agents only require six months of actionable data
“You just need enough data for these tools to work. You don't need 10 years of data. Six months is enough. You need a volume of actionable data.”
Queener: Legacy SaaS needed sales overhead because products did not do the work
“Like all of this is because when you delivered the product, it didn't deliver value. A customer would have to use your product and then go do their job. But you have a product now with a what the fuck interface to look at and they realize in their head, oh my …”
Queener: Agentic AI Enables Multi-Product Expansion in One Month
“The reality is that you're not building screens. If you have the metadata and your agent interface understands the context, you can do it in a month.”
Queener: Enterprise Workers Will Refuse to Use Multiple AI Agents
“Customers,
Don't want to use multiple agents. A salesperson, an individual functional role does not want to use seven agents.”
Queener: Modern AI Startups Replace PRDs with Interactive Prototypes
“And I'd say our most leading edge companies like design and product is just building prototypes. There's no doc. There's heading prototypes to dev success or support that we moved out of dev normally is back in dev because the iteration loops to hear from supp…”
Queener: Customers make great PMs for agentic software solving pure pain points
“What is interesting now in a world, which you're not building screens, but you're solving problems, as long as you're smarter customers, at least they're actually very good product managers. They're giving you use cases or pain points that I want you to solve …”
Lemkin: CIO Budgets Are Entirely Consumed by AI and Price Hikes
“No matter how great your app is, there just isn't any budget at the CIO office. It's all being consumed by AI and price increases.”
Lemkin: 20,000 Words of Recent Content Are Enough for AI Agents
“You don't really need twenty million, right? You might need 20,000 great words of content that are recent on your product. It might be enough to make the AI great.”
Lemkin: Deploying AI SDRs Costs Founders Time Instead of Saving It
“This is, you will get more benefits, but you do not get time back. You lose time. You lose time with these agents.”
Lemkin: AI agents are not 'set and forget' in 2025
“There's no set and forget, no matter what anybody tells you in 20, 25, maybe in 20, 26. But there's no set and forget.”
Lemkin: B2B companies should buy off-the-shelf software rather than build custom AI
“The main learning is only do this if you can't buy it from 99% of you. Like it's cool, but we did this for content review and other things, and we'll show you where there is no off the shelf app, but don't roll your own. Buy, buy, don't build, right? Use for B…”
Lemkin: AI agent success depends on daily deep training, not vendor choice
“None of these names are as important as putting in the reps and the work. And Amelia will show you the actual date in a second. These are all great apps. We stand behind them. We recommend them for sure. But you could pick another two. The most important thing…”
Lemkin: Implementing custom OAuth in vibe coding creates major security leaks
“All of these apps have their own OAuth built in. That is secure. That has been hardened. And everyone goes in and I know Replit the best because that's where I've spent time, but they all have the same ultimate. They're all more similar than they're different.…”
Lemkin: AI agents fabricate unit test passes and break code on failures
“If you build unit tests and this is, if you go back to the craziness that happened to me on social media, the agent, when it, first of all, the agent wants you to be happy. So we'll make up data. That's number one issue. We'll say it passed the test when it di…”
Lemkin: Building a production vibe-coded B2B app takes a month, 60% QA
“If you seriously want to put a simple, but real B to B app into production with real users, collecting real information and charging for it for real, and you want it to be any good budget a month. A month and 60% of your time will be QA and testing.”
Cabane: Marketing is splitting into human-touch events and AI agent optimization
“There's really two broad categories, things that are AI proof, right? The AI resilience. So that's going to be like social proof and influencers is going to be like some of the UGC stuff. A lot of that, the webinars, the in-person, the events, the dinners, all…”
Cabane: AI content workflows match or beat the output of large teams
“And I'm now encouraging CMOs to like have much leaner content teams, and just a few people that can run content workflows at scale, because the output is just as good, if not better.”
Bessemer's Talia Goldberg: COGS Is the New CAC in AI Business Models
“I have a saying at Bessemer that I started saying a year ago, which is that COGS are the new CAC, which is that you can spend a lot on, on COGS, but it means that you also can't then be spending a lot on CAC in customer acquisition.”
Lemkin: VCs now target trillion-dollar outcomes instead of billion-dollar exits
“VCs are now looking for trillion dollar outcomes. People have decided that every couple of years, there'll be another open AI. Not every week. There'll be a trillion dollar company.”
Lemkin: SaaS companies growing under 20% at $100M ARR lost product-market fit
“After a hundred, if you're growing less than 30%, it's big yellow light, right? And if after a hundred million, you're growing less than 20%, you fell out of product market fit. A hundred million. You have so many thousands of customers. They should be telling…”
Jason Lemkin: Most sales reps want 80% earnings for 30% work
“Most sales reps want to earn 80% of what they can for 30% of the work.”
Jason Lemkin: Avoid marketers lacking a strategy for the AI search shift
“If they don't understand what they're going to do about the fact that search is going from SEO to chat CPT, and I don't know the answer, but I know the issues probably can't hire them.”
AI Replacing Outsourced Business Processes Faces a Pricing Race to Bottom
“Some of the easiest things to sell AI products to today or automate are things that maybe have already been outsourced that business process organizations are doing already. And the interesting thing about that is they might actually have a real race to the bo…”
Pre-Training Proprietary Models Does Not Build Moats for AI Application Startups
“Training your own model, certainly pre-training it, not helpful. Yes, you need data to fine tune, but it's not actually a ton of data. And so a lot of people have access to that. And so in many ways the differentiation that I think will continue to compound is…”
LaRue: Prospects Prefer Trained AI Outreach Over Inexperienced Human SDRs
“Again, if you can get your AI to be decent to the level of you yourself, that's probably the quality you should aim for. They will prefer that to, you know, a AE or SER that is one day in at your company saying, hey, have you heard of Saster?”
Lemkin: High-Quality GTM AI Creates More Work, Not Less
“You get higher output. You get higher quality, but you, but there's no less work trade off here. That does not seem to exist in high quality GTM AI today. More work for more, for better output.”
Levie: AI coding tools will build strong moats via deep codebase context
“I'm actually very bullish right now. You know, there's like a little bit of a meme of like these AI coding tools. You can just swap them out. I think that's like a temporary thing that you'll see on Twitter or X or whatever. I think the mode is actually fantas…”
Ramaswamy: Partnerships are fake unless a partner employee's career depends on you
“And the thing that I always tell any partner is unless there's someone in snowflake whose career and future depends on you succeeding, you don't have a partnership. And by the way, that's true for every company.”
Rangan: Software pricing models are shifting to volume of AI work completed
“Everything in this age, and we'll talk about this, but the business model is changing. It's not like what product you shipped and how much you can get for that product. It is where is their value and how much work is getting done because of AI.”
Norton: Relationship-only CROs will not survive the next decade
“You need to be a highly systems engineering focused CRO to survive over the next two to five to 10 years. Yeah. If you don't have the acumen to be able to understand how customer journey connects to systems and infrastructure and connects to the rep process, A…”
Benioff: Current CEOs Are the Last Generation With Only Human Employees
“I'm the last generation of CEOs to say, maybe you are too, that we only had human employees.”
Benioff: Establishing a Standalone Enterprise AI Budget Is a Mistake
“I think that where there was an AI budget, that was a mistake. And that where it should be is like, you should be holding, at least this is what I do. I should be holding our service team accountable. Our sales team should be held accountable for that. And tha…”
Lemkin: Modern SaaS IPOs require 50% growth at $500M ARR
“The IPO bar today is 50% growth at five hundred million, okay? Now that's not a billion dollar IPO, that's a much bigger IPO.”
Frankel: VCs often push startup sales to generate fund liquidity
“Often there is an investor on the cap table that's been there for a long time, and their model, it's like the tail wagging the dog here, their model is they need to see some liquidity for the next fund, so they're often very problematic reasons.”
Frankel: Founder-only secondary liquidity creates friction with executive teams
“Sometimes like the mistake we make is we take pressure out of the system by giving the founders some kind of liquidity. But we don't think enough about the rest of the team. And then the team's playing on the founders.”
Moreno: Delaying RevOps until $50M in ARR creates massive technical debt
“Generally, my experience has been when it's frustrating, it's because at around 50 to a hundred million in revenue, someone's like, oh, maybe we should use data to make decisions. And then it's like a multi-year process of unwinding a bunch of stuff that happe…”
Zinman: Upmarket SaaS firms become irrelevant by abandoning SMBs
“People don't realize that the SMBs and mid-market of today are the enterprise of tomorrow. So they getting used to use another product, they're getting used to use another platform, and eventually they become the enterprise companies. And then those enterprise…”
Zinman: Scaling past $1B ARR requires going multi-product
“If you want to scale a company past a billion and even to a billion, you have to go multi-product. There's no other way you can do it.”
Zinman: Multi-product integration beats standalone best-of-breed CRM solutions
“And because of that, you don't necessarily have to have the best CRM in the world. It might be second place, but because of that integration and because of that data flows between the different products, you get such huge benefit from that, that it's a huge ad…”
Clayton: Modern SaaS IPOs demand $400M ARR and 30% growth
“We think that you need to be at least four hundred million of ARR. You have to be growing at least 30% a year. You're probably going to be worth three to five billion dollars and trade anywhere from that eight to 12 X range.”
Clayton: SaaS companies cannot expect public valuations above 10x ARR today
“And regardless of private market valuation, at least in today's interest rate environment, companies cannot expect to be worth more than 10 times ARR.”