Evolving Workstation Needs with Larry O’Connor

This is Textron tv. Hey guys, thanks. VI throw.

We're here with Larry O'Connor, who's CEO for other world computing, and we're talking about workstations and how they're gonna evolve, particularly in the context of remote work. Hey, Larry, welcome to show. Hey, thanks for having me, Michael.

Appreciate it. We see everybody from, uh, content producers and creators to application developers, all requiring higher and higher end systems these days. And, um, in the age of ai, we're starting to see more usage and reliance on GPUs and, um, there's a forthcoming generation of platforms coming out known as AI PCs that have these NPUs in it.

I think that stands for a neural processing unit. Um, are we at some sort of inflection point here in terms of workstations and what, uh, people are gonna need or to go build the next generation of applications and content? I guess it all depends upon what you're starting with.

I mean, I would call out that on the Apple side of the fence, there's been and neuro processing capability in all the, the M series, you know, since actually the first M series introduction in 2020. So the capabilities, you know, for some of the AI assists, and that's the other, I think it's ai, I mean, it's an algorithmics, you know, support infrastructure for off record purposes. Very, very powerful, very useful.

But it's, it replaces, I guess, some of the mundane tasks with automation so that there's more time spent on, well, tests that actually require, uh, you know, some directed creativity. I mean, in all fairness, AI even helps creativity today. And with that in mind, so what does the, I, and I know there's no ideal, but the, how do you think that that workstation is gonna be configured?

Or what will it need in terms of memory storage and, um, I guess a lot of folks are wondering if the systems they have are about to become a little antiquated? You know, it all remains to be seen. I mean, a lot of the AI capabilities that creatives are taking advantage of are today cloud-based.

You know, a lot of the on, uh, prem AI capabilities, I should say, on device AI capabilities are really there to automate, you know, maybe studies keyboard or mouse and being able, they use voice activation or have, uh, I guess you can say preem preemptive pre predictability in terms of, you know, what somebody's going to do. But the primary, uh, workflow on a pc, you know, at least the first generation of these, you know, these, so-called AI uh, systems really doesn't, really, doesn't change a whole lot in terms of the capability. I mean, the applications that people depend on already have these AI plugins, so to speak.

And again, they, I mean, they're a little heavier duty than what, uh, is happening in, in a lot of cases on-prem. And that's, you know, that's being driven by the cloud. Mm-Hmm.

A lot of folks have it in their head when they hear workstation that this is, you know, something that is, uh, large and maybe heavy and feels more like a gaming rig than anything else. But, um, can I use mobile devices these days? Is those workstations, are they up to the task that we need?

Or is there still some sort of demarcation between mobiles and desktops? You know, it's pretty amazing what you can do on a mobile device today to talk about a tablet or a, or some of the latest phones. But I still would argue that a lot of the, the heavy lifting is still done with compute in terms of that, you know, laptops and today's laptops are incredibly powerful versus, you know, what we had even in, in loaded out PCs, you know, five, 10 years ago.

The other aspect, it really comes down to what you can connect to these devices and the external connectivity with Thunderbolt in particular, it gives super, super high capability for the ingest and the processing of data You used to have to have a, a desktop workstation to do now what you can, you know, you, you can achieve by connecting a device, uh, to, to one of, to A-U-S-B-C laptop port. And that's, that's a big change. And that actually helps keep things from obsolescence because when we're talking about ai, some of these AI capabilities which will be built into PCs are also gonna be, how do I say, addable to existing systems, you know, via these external, uh, via external thunderbolt.

Do you think the networks that we have, especially in our homes, are up to the challenge that we're about to encounter with all this stuff? Or is this an another area where there's maybe another bottleneck on the way? You know, a lot of things are proxied as long as you're doing the heavy lifting at home, the actual processes that AI and, and other services, other, uh, what I say off-prem capabilities or added via are done with relatively low bandwidth transmission.

So as long as you're maintaining know highest performance data storage locally in particular, you're doing your editing work on-prem, you know, what's required of the network is relatively light. If you're pushing up you raw data up into the cloud, that's gonna be a problem in, in both directions. And interestingly, you, when we talk to folks about these cloud workflows, I mean, there's so, I mean, there's a lot of convenience and, and some of the opportunities you have definitely have to do everything up into the cloud, but the cost of the time, you know, quickly outweighs the benefit.

You know, one company in particular, this is a small group, you know, they found their entire budget was consumed when they moved from, they had a great on-prem set up, you know, the cloud gave 'em some, gave 'em some benefits, say, Hey, also some drawbacks, but the biggest drawback was their entire one year budget was burned and the first month that they were a hundred percent cloud based. You know, you've gotta, uh, look at what works locally, you know, what needs to be shared, and, and ultimately, you know, what, you know, pick the services and pick the, uh, capabilities the cloud is best there to, to serve for. How connected are these remote workstations from your perspective?

Are they centrally managed by some IT organization, or are they pretty much operating in islands of little pockets of, um, productivity that exists out there, but essentially they're, um, disconnected? You know, and entirely depends on the organization and how large the organization is. Obviously we've got independence everywhere, but in terms of organization that's spread across, uh, the world versus being an on-prem and local, you know, the, the, the same tools that we've had for same tools we've had for on-prem that continue to work, you know, wherever somebody's located, I mean, our team in particular, we've got almost 300 people spread around the world.

Now, granted, we have concentrations, but it's all single platform that's used from an IT standpoint. They manage all of those, uh, laptops, workstations, even mobile devices. So if you wanna be, whether you're, you're local or not, the management is, is actually, uh, still, still pretty straightforward.

And who typically owns that workstation, especially that remote workstation? Is it the developer or the content creator goes out and buys one themselves and then does their work on it? Or is this something that is still kind of provisioned by a centralized IT team?

Once again, I think it depends on the organization and how concerned they're with the security and the ownership of, you know, what's on, uh, a particular, uh, pc. It can be either way. I mean, there's a lot of bring your own, uh, device, how they say opportunities today with different companies.

And I think people prefer that because they also like to use the same device for things that aren't just for work. And again, with the right management. Now, that's really not a, uh, not a big problem, but it comes down to the organization and I even potentially the, the employee in the case of a, a hired employee, that employees preference, you know, and at o WC for, for all purposes, you know, we try to provide the equipment, but we also let people upbringing our devices on occasion as well.

Mm-Hmm. What is the life cycle of the, of a workstation these days? When I do buy one?

I mean, a lot of times, you know, historically I went way back. People were trying to make those things last for five years before they would replace them. Is that cycle gotten faster or what is the level of upgrade that we normally see these days?

You know, today, I, I think, you know, a lot of the capability, I mean, depending upon what you're doing, you know, systems that are years old are still more than adequate. I mean, especially 4K and below in terms of, uh, content creation, you know, we're, we're way overpowered of us tripping today. You know, whether you're going into six K or eight k, I mean, okay, the latest and the greatest, they're gonna make a difference.

But ultimately these systems are super capable and the places you upgrade to improve your workflow are really your connectivity and your storage. You know, we've had 40 gigabit, uh, thunderbolt mouth, you know, since 2016, and we're about to go to Thunderbolt five, which is gonna double battery rates and all, but you pass, I mean, with, you know, 3000 megs a second across stores, you're past real time on, certainly on 4K, and even you, the process six K and, and run six k, even eight K workflows. So it's these systems, you know, absolutely are able to last.

And then it's, it comes down to I buy something new, what's the improvement for my workflow, including my downtime and, you know, the transition and going from one system to another, you know, what's the, what's the benefit, you know, versus the cost. And in a lot of cases, there are other upgrades or other things you can do with an existing workflow or other enhancements outside of the, the workstation, other services and such that are gonna make more sense than buying a brand new laptop and less, or workstation or tablet or, uh, or phone for that matter, unless there's a specific feature, specific capability of what you have now, now is not able to perform or a performance a benefit, that's something new specifically addresses. But from a lifecycle point of view, you know, most creators, uh, really wanna have a workflow that's fast, stable, and, and reliable as opposed to, uh, being on the bleeding edge.

The bleeding edge can give you great, you know, performance benefits in certain areas, but you know, sometimes you fall off that edge and have to climb back up in terms of working out the bugs and the tweaks with, you know, new OS versions and, and new hardware. Mm-Hmm. Speaking of new hardware, there's everybody who has a workstation and is probably, uh, wants a GPU, but how available are these things these days?

You keep hearing about the shortage and they're pricey and, um, there seems to be for those that have them, they're trying to figure out how to maximize the utilization rates. But what is the current, current state of the GPU space? You know, the high-end GPUs used for AI are definitely, uh, I put in the hard to get, uh, category.

They have lead times, they're expensive, but GPUs for how to say creative purposes and even the, the technology, uh, direction, we're going to bring AI on prem, uh, we can actually provide pretty comparable, uh, performance to the really high-end GPUs with a different kind of configuration that uses lower cost GPUs. There's a great I abundance might be, uh, you know, overstating it, but there's, there's really not the supply constraint that we had a few years ago, a few years ago. Crypto, uh, in particular Ethereum and other, uh, types of GPU based mining really were taking a bite out of the availability.

And since Ethereum has gone to staking and, you know, the, the crypto in general in terms of GPU based, uh, crypto, uh, mining technologies have kind of shifted. There's been a much, uh, and that's freed up availability about the support, the big shift to, to AI and also for general graphics purposes for having a high power VP for doing your renderings and such. You know, those GPUs are, are pretty, pretty available today.

Mm-Hmm. So what's that one thing that you see organizations doing in terms of how they manage infrastructure and laptops and workstations and everything that goes with that? That still just makes you shake your head and say, folks, we need to be a little smarter than that.

You know, skimping on internal storage, skimping on memory, you know, not, you know, just, I'd say deploying a large display or even two displays, you know, you have these fixed costs and then you have your ongoing costs and personnel and it's, you know, while you can run the numbers and I mean, it's, and look at budget, you know, the productivity you get from just a few little improvements, you know, you know, better dock. I mean, it's, it's interesting. You know, we look at, you know, someplace that takes 'em a couple years to, to come around, but, you know, you save a dollar today and if you have to replace it in six months or a year, you know, whether it's under warranty or not.

I mean, the disruption in the cost of productivity, you know, that apparently falls into a whole different, uh, category or, or, or place of measurement as opposed to the original budget. But it's making good choices on the front end that really provide benefit and productivity to the, the team that is depending on that technology for the long term. Mm-Hmm.

And the people who use, especially workstations, they're in high demand. These are some of our most expensive, uh, employees, and we're always trying to retain them, and other people are always trying to steal 'em. Do you think that all things being equal, if, uh, salaries and benefits are roughly the same, it might, the decision as to where they're gonna work might just come down to the equipment that they're gonna be provided?

Absolutely. Yeah. That's a, a very, uh, a very valid point.

And I, I would think pretty true. Yeah, all things are equal and you wanna be someplace where you're gonna be able to, you know, create the best and, and have the most, I mean, a lot of these people enjoy what they're doing, and it's, it's a lot more fun to be on a system that you're not being held back by. All right folks, you heard it here.

You know, the hardware still matters. We've been overlooking it for a few years now, but ultimately, um, that drives a lot of the work experience. And if I am frustrated with that, I'll go maybe work somewhere else.

Hopefully not, but at least we know what the issues are. Hey, Larry, thanks for being on the show. Hey, likewise saying thanks for having me, Michael.

All right. And back to you guys in the studio.