CEO Michael Jackowski on AI, cloud, and digital transformation in…

Transcript

I just can't impress upon you enough how much is technology is changing. I want everyone to understand that the models are being used to create the next generation of models. How many of you remember when Deep Sea got announced? Raise of hands. Remember Deep Sea getting announced? It was everywhere in the headlines. Here's one thing I find astonishing. I feel like when Deep Sea got announced, it was like years ago. It was in January of 2025. Like think about it and AI time. That's like a freaking decade. I'm like blown away. Invidia lost $600 billion of market cap that day. Do you understand how significant that was? Now here's the thing that you have to understand, by the way, the Chinese like there are cranking. They're trying to figure all this out. One thing to understand though, DeepSeek would not exist without ChatGPT. I want everyone to understand that, right? Like everyone's like, ohh, how is this working? But they did come up with these new techniques. I won't go into the details like distillation, but it's a process of basically taking the knowledge out of these big models and using them as a jumpstart for training all kinds of ridiculously awesome, you know, mature models. So, and what you have to understand is in creating these models, it's a ton of manual labor, you know, wrangling data, getting data where it needs to be, setting up the. Parameters. What should the parameters be? And what's happening is the models now are doing all that work, not humans. So we're taking the human labor out of the the capacity limitation here. And what's all of this like? I know it's probably way too techie, but I just love this stuff. So I thought I'd go into it to impress upon you how significant this jump and leap is. But in the end, remember this thing called Moores law? What was Moore's Law? Processing will double every 18 months, right? That was Moore's Law and we thought that was super profound. And there are some that say that this is going to 10X every year in terms of knowledge and capability, 10X every year. I want you to understand this. It shows 100 X up there. That means in two years it's not 20X. That means in two years the improvement is 100 X. I want you to think about that. These models are getting smart, they're getting capable. And what I love is there are so many different ones of different sizes and open source and the big ones that we have this technology that we can use something that from a cost perspective is fit for purpose for what we want to do, really powerful stuff.