00:31:15 · Kevin B. Boyd
[00:31:15] S1: Yeah, there are some really big differences. And part of it is that our organization is so very decentralized. When you're at United Airlines, the CEO says, alright, we're going here. Like, need you... We need the the the business processes to go here. Here's our business strategy. And then it's really a question of, okay, so how do we align the technology for the business strategy? Make sure that we're building the the infrastructure, the applications, the other tools to support that business strategy. It's really different in the university. It's more like you have, you know, 2,000 CEOs each going in their own direction. And this one over here is researching ancient languages, and this one is trying to come up with something that will help a particular type of disease. This one over here is in astronomy and they're ingesting enormous amounts of data from the Arecibo telescope in South America. Another is ingesting enormous amounts of data from Argonne National Lab or CERN, trying to understand the makeup of the universe at a really tiny level. You know, bringing all that together and saying, okay, so what do we think they're going to need? Like, I'll give you an example of where it's really challenging to be responsible for technology at a university. There's a lot that we do that actually requires physical GPUs. It becomes not terribly cost effective to try and do it with the cloud resources, just because some of this research just will run twenty four seven for a really long time. And so it tips the balance to doing things on prem with our own our own hardware in the research space. But then it becomes a question of, all right, so how much do we need to buy? And then you're actually going out to hundreds, if not thousands of faculty and trying to get a sense of what they need and where's the funding gonna come from. And so that's been a real challenge and it's a particular challenge with need for additional GPUs and obviously the costs for those keep increasing. So we're very similar to others in that we need more and more data center space, we need more and more compute to support the amazing research that's going on across campus. And it's figuring out, all right, over the next twelve months, what's the compute capacity that we actually think we need over the next two years and so on.
00:34:09 · Kevin B. Boyd, Grace Gong
That's what makes it most challenging and interesting. [00:34:14] S0: When I guess when you're thinking about, I think that's like so interesting to like think about like which department needs more GPUs. Like, is there any particular department that needs more than others? Like, obviously you don't have to give out like the specifics, but like, I guess I wouldn't assume the computer science department is probably needing more than, I don't know, biochemist or... I mean, you know? [00:34:36] S1: It depends. There are particular types of research that are very GPU intensive. And interestingly, like there are, there's GPU intensive research in the business school. There's a lot, a lot that goes on in our physical sciences division. That's where computer science is, data science, physics, and others like that. But there are also others doing a lot with AI research that's GPU intensive even within our biological sciences division.
00:34:36 · Kevin B. Boyd, Grace Gong
[00:34:36] S1: It depends. There are particular types of research that are very GPU intensive. And interestingly, like there are, there's GPU intensive research in the business school. There's a lot, a lot that goes on in our physical sciences division. That's where computer science is, data science, physics, and others like that. But there are also others doing a lot with AI research that's GPU intensive even within our biological sciences division. [00:35:12] S0: How does the budget comes down? Like, it's not like, for example, like there's like grants from, let's say, like for sector X. So like, oh, let's fund the research of like this thing. And, you know, I was like reading the report of like, I know that UChicago recently invested into building out like the part of the campus, like, you know, to make it more, you know, like I guess like reserve friendly and stuff like that. So like, how does people kind of like get the budget? Is that like each department has their own budget and then they invested to JPE slash like they contact you and your department to be like, hey, we need the support of X and then you kind of like negotiate a deal with the company since you are working with like, let's say, quote, unquote, 2,000 CEOs. Right?
00:35:59 · Kevin B. Boyd
[00:35:59] S1: So the the answer is it depends. There's a there's a certain level of support for research that comes from the central administration. And that... So like we have a research computing center that has high performance compute, you know, the sort of the supercomputer type resources. And those are made available to all the faculty across the the campus. We just, excuse me, announced in the last thirty days a very large philanthropic gift from the Schmidt Foundation. It will be several tens of millions of dollars in GPUs that are available to the campus. And so there's some level of resources that are just provided centrally and then we distribute them across the faculty. But the reality is that most research is funded by grants. And you know, the way that works is a researcher goes and they're going to write a grant proposal and that's going to go to the National Science Foundation or it's going to go to NIH or one of these large much of the research is funded by federal agencies, and obviously that's been impacted by this current administration. There have been some real challenges in higher ed research over the last couple of years. But so much of university research is grant funded, whether it's public dollars coming from the federal government or private dollars like this gift we just received from a foundation.
00:37:50 · Grace Gong, Kevin B. Boyd
[00:37:50] S0: When guess like when you're thinking about like the gift from like different foundations, do they invest into like the same or is that like, you know, particular domain that's like getting the funding? Or like, would it be like, you know, the whole school is going to be like benefiting from like a gift like this? [00:38:12] S1: So again, it depends. Like the very large gift that just came from the Schmidt Foundation, That one is intended for science. Thank you. And... But it's available. It's going to be available very broadly across the universe... Across the university. When a researcher goes and let's say they get several million dollars in grant funding from NSF, NIH, or one of these types of agencies, that is very specific to like whatever the project is. So basically what they will do is they'll say, I want to research XYZ, and here's what I need in order to do that. You know, this computer hardware and memory and storage and some money to administer it. That can really then only in most cases be used for what the the grant was specifically written for.
00:38:12 · Kevin B. Boyd, Grace Gong
[00:38:12] S1: So again, it depends. Like the very large gift that just came from the Schmidt Foundation, That one is intended for science. Thank you. And... But it's available. It's going to be available very broadly across the universe... Across the university. When a researcher goes and let's say they get several million dollars in grant funding from NSF, NIH, or one of these types of agencies, that is very specific to like whatever the project is. So basically what they will do is they'll say, I want to research XYZ, and here's what I need in order to do that. You know, this computer hardware and memory and storage and some money to administer it. That can really then only in most cases be used for what the the grant was specifically written for. [00:39:19] S0: Mhmm. So when... I guess, like, when you're thinking about, like, since, like, you know, you were... You since... You started working at U Chicago for... Since, like, 2012 compared to like from in the past like ten years or so like what would you say is like the core differences the school has changed and like you know nowadays like we're in like the AI era compared to like past like technical trends, whether it's like blockchain or like, you know, other different trends, like what would you say is like the core differences that in terms of like impact of the school? [00:39:57] S1: Yeah. Know, AI is different. I think in some ways AI is more similar to what we saw around ninety nine, two thousand with the internet becoming, you know, more widely available. [00:40:15] S0: And