Jason M. Lemkin (speaker) argues that Palantir always integrated machine learning from origins in PayPal fraud mitigation, highlights Palantir’s dramatic turnaround despite being founded in 2003, and uses the example (and comparison to Shopify, founded 2006) to urge established companies with brands and customers to adopt AI quickly — including firing staff who resist — because inaction risks gradual decline. The tone is urgent and prescriptive.
Questions this video answers
2003
2006
Chapters
Now they always did machine learning. This is the whole point. Palantir comes originally out of learnings from PayPal about how to mitigate fraud. And so, you know, classic AI ML was part of their DNA. But this company was on terminal, on the terminal ship to nowhere, right? And it's the greatest turnaround so far in the history of software. So that's why I keep putting these numbers up. Not to tell you to go build, turn Your E Commerce B2B into Defense Tech or go software to kill people, but because, you know, this is a company founded in 2003. So what's your excuse? It's not a company founded last month. It's not a. It's not lovable. Founded six months ago. 2003. Right. And even Shopify, which in some ways is still early ish, right on the journey. But I mean, you know. When was Shopify founded? 2006. I don't think Toby quit, did he? As near as I can tell. So, and, and so, and so maybe this isn't the question you're asking. In a way, if you're behind, it can feel like it's too late. All of a sudden you felt really stable 20, 24 months ago. Now you're surrounded by AI enemies that are getting all the press and getting everything and growing faster than you. And yes, if you do nothing, you will slowly die. Like, let's not make any bones about it. But it's not too late. If you have a brand and happy customers, it is not too late. But you gotta go now. And you gotta fire everybody that doesn't want to be on that journey. You have to fire all of them. Fire them all. If so many folks don't want to, don't want this change, they don't. I would say most, most don't want them. Most founders are willing to do it, but.
