Stephen O’Grady, co-founder of RedMonk, and Kate Holterhoff, senior analyst at RedMonk, discuss the 2024 IBM Analyst Forum in this quick take video.
This was a RedMonk video, sponsored by IBM.
Transcript
Kate Holterhoff
Hi there, Kate Holterhoff, Senior Analyst at RedMonk, and with me I’ve got Steve O’Grady, Co-Founder, Principal Analyst at RedMonk, and we are here at IBM’s AR Forum, and we’re just going to talk a little bit about what we’re seeing here. So Steve, you want to go first? What are your impressions of the forum?
Stephen O’Grady
A lot of AI, as you’d expect. You know, we’ve been updated by a variety of IBM folks over the past day and a half. Had the opportunity to hear from customers and partners and all that fun stuff. So I think the biggest takeaway for me is that the IBM’s AI vision, I think, is one that at least makes sense, I think, from my perspective, in that for enterprises, when they’re going out and trying to tackle specific problems and issues with AI, a lot of times they’re going to be doing so with not the sort of general foundational models that get sort of all the flash and attention from a market standpoint, but they’re going be doing so with some smaller, local, fine-tuned models on enterprise data. So one of the facts and figures surfaced today was that in IBM’s estimation only 1 % of the data in AI systems currently is enterprise focused that you know, the overwhelming majority of it has nothing to do with the enterprise and so their goal is to change that and it makes sense, you know to me on paper, but that’s me. What do you think?
Kate Holterhoff
Yeah, I mean, I’d say one of my big takeaways is the emphasis on Granite. So when I attended IBM Think in the past, there was sort of a push towards the multi-model strategy, but I see that they’re really pivoting so that Granite is front and center, which, you know, looking at some of the charts that show how it compares with some of the other models that it’s competing with. Yeah, the benchmarks suggest that it’s actually doing fairly well. So it makes a lot of sense. It seems like a decent strategy,
Stephen O’Grady
Yeah, yeah. we’ve seen a variety of vendors take different approaches, right? So some vendors touted their single model. Amazon famously stood up, was it two years ago, I think, at re:Invent and touted, I think, 141 models they had available. so there’s very much a paradox of choice in this sort of issue in the sense that enterprises are probably not going to want to sort through dozens and dozens of models. I think there’s, you want to have some choice, you don’t want be locked into a single model. But these are my perspective, I think, to focus on, hey, we have a model that’s going to work for what you want to do for your enterprise needs. Yeah, it makes sense.
