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 12 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
May: OpenAI and Anthropic will be pushed to raise inference prices
“I think it's going to get worse before it gets better, because my guess is as open AI and anthropic IPO, wall street's going to scrutinize their unit economics. And it's probably going to ask them to raise prices, push them to raise prices on inference.”
May: Many senior executives are imitators rather than adaptable problem solvers
“You know, a lot of people that have climbed really high in their career are just great imitators. They just take a playbook that they saw somewhere. They don't always fully understand it. They don't always know how to change it for changing circumstances, but …”
May: Best GTM strategy for IT buyers is open sourcing core value
“One of the things I've learned having built companies that sold to like sales and marketing people and companies that sold to IT people is like your comment on, well, you can't give things away for free for people who don't value it. Very true in the marketing…”
May: Massive markets forgive mistakes, unlike winner-take-all smaller niches
“The reason you wanna do that is because when you're in a big market, you know, you can make a lot of mistakes and still Come out on head. You could be the seventh biggest player in the industry and still do really well, you know whereas in, in, in, you know, i…”
Rob May: Good executives should immediately know who they would hire next
“I think as an entrepreneur, and also I expect this of the executives that work for me, that they should constantly be cultivating talent. You know, a couple of times a month, like, you know, going to events or taking people out to dinner or having coffees with…”
May: Founders de-risk talent recruiting by letting prospects observe gradual progress
“A lot of people think it's risky and the best way to de-risk it is for them to see your progress.”
May: Startups Lacking Access to Stage-Experienced Talent Should Bet on Raw Intelligence
“Sometimes you don't have access to those people or there's not enough in the market that you're in. So that, that is ideal if you can find those people. But you know, sometimes you can't, and you just have to give somebody who's really intelligent, like a shot…”
May: Acquirer failed at shifting his startup's product to channel sales
“I mean, this happened with the company that bought my company. They were very channel sales centric. I didn't think it would work well for our product. And they tried to move us over and after a year they had to switch back to direct, but like their philosophy…”
May: Companies address LLM inference costs at $20k-$50k monthly spend
“Somewhere when you hit that sort of like 20 to 50,000 dollars per month in inference spend, and you realize it's growing at 1012, 20% per year, whatever it is, that's when you start to go, oh, we should do something about this.”
May: Reasoning AI tactics increase token usage 20% of the time
“Sometimes when you try to use like one of these reasoning models where you add chain of thought or one of these tactics, they'll actually use more tokens about 20% of the time. It'll be more expensive.”
May: AI architecture will mirror software transition to microservices
“Most companies start with a monolithic code base, everything in one repo and one giant thing. And then they start to parse off the pieces that need to be optimized and make them microservices. And I mean, it's, that's how businesses are going to go with AI as …”
May: AT&T cut AI token costs 90% using small language models
“We just saw an article in late February from AT&T or AT&T was doing eight billion tokens per month through their core AI systems. And it was costing them too much money, and they rewrote their entire architecture to take advantage of small task-specific langua…”