Deloitte’s David Linthicum on Cloud and AI

foreign speaks we're talking about what's driving the cloud computing market and we're also looking at an increasingly important aspect of the cloud Market which is how does cloud computing enable artificial intelligence very hot topic these days to discuss that I'm joined by uh David Linthicum Chief Cloud strategy officer at Deloitte Consulting Dave very good to have you with us today it's great to be here again James I appreciate you inviting me on absolutely uh so so what is driving the cloud Market here in 2023 it seems like there's a lot going on I hear basically that there's there's definitely a major desire for cost control uh and also a way to find the to control the monster called multi-cloud I mean companies started all these clouds somewhat parallel and now they realize that we've got to manage all this but at the same time all that's going on I sense that companies realize hey Cloud enables AI making it even more important uh your thoughts on all that yeah I mean uh everybody thought the you know there was going to be a downturn in Cloud spending and suddenly uh we see studies where it looks like uh people are back into the game uh not because they're necessarily looking to spend more money but they view uh generative Ai and using that as a force multiplier for their business is something that's that's non-voluntary they have to do it and so boards or directors are coming to I.T and saying how are we going to automate our supply chain increase our customer experience do all these things that they haven't done well in the past uh through the use of generative Ai and of course that runs in the cloud so there's an interesting cloud computing because there's an interest in generative AI there is still an interest in cost controls they don't want to overspend for it and finops is becoming more of a hotter topic within Enterprises that are leveraging this technology but it's definitely cueing up to be a larger spending spree I think in 2024 2025 around cloud computing because they're interested in getting to AI through cloud computing um well first of all thank you for using the phrase generative AI because that has to be used in every conversation at this point you get five watching It's All About It is I guess it's the funny thing about cloud is that is there any other topic except for AI really generative AI for all the years I've been covering cloud cloud has been like really dominant and suddenly as of last November when chat GPT came out it's like everyone's talking about AI it's really the only thing um it almost sounds like the the idea of controlling multi-cloud which used to be a really hot topic is a little bit quaint these days are companies still concerned about hey we need to control this thing called multi-cloud yeah from an operational layer you got to remember AI is about solving strategic problems in other words building core systems to take our business to the next layer if you look at multi-cloud it's really about operating infrastructure in more efficient ways and so that's always been a Hot Topic but it's it's a solvable problem we can move at it with you know different monies uh different money and you know siled based Solutions and looking at cloud computing as you know kind of separate uh you know Cloud brand name so you can run it inefficiently and still run it uh so people are looking to become more efficient at multi-cloud typically through the operational layer so in other words they're trying to do that better so that's not as interesting to talk about or even write about as kind of the generative AI stuff or do you get into you know the capabilities of technology which is approaching science fiction but you got to have both and I think people are still spending money and talking about multi-cloud it's certainly still a topic that I talk about a lot as well as the generative AI stuff but it it's just becoming supplanted because there's only so much writing and speaking we can do in the tech in the tech community and people seem to be focused on that but I get most of the uh things that I do on a daily basis is explaining how multi-cloud can be operated in a much more efficient way but in many instances they're doing so because now they're moving into a multi-cloud environment because they realize they needed more than one Cloud brand to leverage generative AI to kind of take things to the next level well all right speaking of that move to AI what is your sense of of the Enterprise approach to AI in terms of spending um I would think it'd be a massive land rush to invest or your desire to plan and research but I don't I also hear like oh there's a tentative thing we want to think about this a little bit more and we know it's how we have to get on board the train and yet we're not quite sure I mean certainly there's a surge in interest that's once in a generation you feel like the wallets are open correspondingly they are I I think the studies I'm seeing um is that the the spending is going to be a lot more uh generous in terms of leveraging AI technology because they do see the immediate feedback and Advantage uh the problem with that is in many instances I don't think businesses have a good understanding of the problems are looking to solve so it's one thing to say well we need gen III just like we needed Cloud 12 years ago what issues are we looking to solve what are we trying to improve uh within the business processing and that really kind of needs to be the Gap that that people need to narrow um so in doing that my concern is that everybody's ready to spend a lot of money but not necessarily are going to end up solving a lot of the problems they need to solve because they're misapplying generative AI in certain domains where it doesn't belong so for example there's no need to use it to build transactional Sales Systems you may want to mine your sales data to figure out new patterns of understanding and how you approach your customers from a marketing perspective or the ability to leverage cnai as far as recommendation engines and dealing with you know digital transformation and how you interact with with clients that's going to have some benefits but in many instances I'm hearing about businesses that are going to be pushing this in certain directions where it really doesn't belong so spendings open the wallets are open uh everybody's looking to uh Nest it's going to be net new spending as well they're not not moving or diverting their I.T budgets and everybody's for that including the boards of directors the core thing is are we going to be making mistakes very similar to the way in which she uses technology in the past and by the way we've seen AI waves in the past you know you know my first job out of college was an AI analyst because I was work AI was a big deal at the time doing list based development things like that which is very expensive so here we are again with much better technology and cheaper technology commodity technology but we could end up making the same mistakes I noticed that a lot of people are around here great uh made the initial mistakes or you know retired and out of the business in my work and CERN is that we're going to move after this in a lesson strategic Direction and make lots of tactical tactical mistakes they're going to derailing the the value of generative AI yeah on this same note I have a feeling that many of the executives and managers they really can't fully predict what's going to happen we're generative AI or even its role in their business and so they they know they need to move ahead but are they able to really predict it um because it really is not like previous Technologies there's client server there's data center I mean it sort of moved to a certain orderly movement even Cloud it's it was a step forward but it wasn't it seems like this is really a hockey stick moment I'm not sure companies really know what they're getting themselves into no and I think you hit you hit the nail on the head it is going to be a hockey stick moment in terms of interest and which is going to lead to spending but is it going to be a hockey stick moment in terms of spending aligned directly with the value that's coming back to the business and that remains to be determined uh I've seen some very good uses of generative AI that are really going to take business to the next level and we'll see businesses in the next few years that'll be built upon this technology in other words they'll be around their technology they're going to do something better than their competitors whether it's Pharmacy or Pharmaceuticals or Transportation or whatever they're going to provide a better experience and therefore weaponize the technology to take their business to the to a level where they're able to disrupt their market and capture a larger share of the market and we saw a lot of this in the past years with you know the you know the uh ride sharing business and things like that so that's going to be applied in different Industries where they're just going to be able to take over but I think a lot of businesses are going to misspend the technology and not necessarily look at the efficiencies of it and not put it to good use and we're also going to see leveraging of data in better ways we have the potential doing that and so in many instances they're going to put gnai in place but not necessarily get their data Assets in a mature place and so Jin AI is not is not a uh is is not going to be useful to you unless you're able to leverage a certain amount of data in certain amount of ways it's garbage in garbage out so the other concern would be they're not investing in in modernizing their databases to really apply to gen AI to allow gni to do what it does interpret data in certain more intelligent ways to drive business processes better to drive better customer experience better product development better Innovation which is really going to increase the value in the businesses so we've talked about this somewhat about how AIS is you know supported heavily by the cloud and no doubt you know it enables data storage and Analytics how can decision makers use this knowledge to optimize their AI deployment in other words they're really aware like hey cloud is where it's at in terms of growing our AI deployment how does that help them how does that help them navigate yeah there's great Tools in Cloud but probably better than on-premise now we have better security we have better governance we have better uh uh uh management technology uh that are networking systems believe it or not and so it's leveraging that technology in a certain way so we can monitor what those things are doing how much spending is occurring and also dynamically how much value is coming back that's very difficult to do on premise so the nice thing about Cloud we have our ji systems in the cloud our data in the cloud uh and we have our Core Business processes in the cloud I can monitor every aspect of that because it's on the same platform it's homogeneous it's on I know a particular Cloud brand and make those things happen so leverage the tools of that cloud brand to find out what's going on what's Happening and ultimately what value is coming back to the business the ability to kind of set up these Square metrics so you can't do that in a widely distributed heterogeneous world just because the tools are very difficult to integrate it's much more easier to do in the cloud the difficulty there is I'm not trying to limit everybody to a single cloud provider so moving forward they need to consider um many different solutions on many different providers and by doing that they may end up in a heterogeneous environment we have the same problem we just talked about in multi-cloud in other words the complexity of the heterogeneity is causing a latency and leveraging the stuff so we have to find kind of a a good trade-off between leveraging this technology on the platform where it's most optimized and it's going to be most efficient and do so in such a way we're able to determine and monitor the metrics and the value that comes back and I think that's thinking that needs to be occurring today we're probably not doing enough of that yet and I think that we need to kind of reflect the fact it really kind of converges the problem if you think about this as a multi-cloud problem it's an operations problem but it's applicable in terms of our movement into gen AI systems you know it's funny you know to hear you use the phrase heterogeneous and I've had this theory for years that I I think at some point we might start to collapse under the weight of our own complexity so you're talking about a multi-clouding world it's distributed world and yet there's a new technology that's more complicated than ever before it looks like a bit of a train wreck on the way like are people going are people really going to be able to know how to handle this well it we're seeing complexity as a as a key inhibitor to people leveraging multi-cloud operationalize it and in many instances they're spending two two and a half times that that they spent on the same systems running on premise and so they do get the advantage of leveraging AI and more faster moving technology agility the ability to have uh efficient access to that technology things like that but it's really uh a wall that anybody to get around right now in fact the industry right now the AI Ops business is calling it a complexity wall so how do we get around that operational limitation get by the added expense uh that multi-cloud is going to bring is going to be the big thing right now and of course we're talking about super cloud and meta cloud and lots of people are inventing you know different kind of conceptual ways to deal with the problem the reality is we don't have one Fair stack and one set of best practices that are applicable in every domain so if you're going to in essence remove a lot of the complexity that means removing a lot of the redundancy by finding cross-cloud services to deal with security operate stations even Ai and data using a single logical layer that you only have to deal with as a human being therefore it reduces the complexity but getting to that single logical layer and that single layer of automation is something that people have made some tactical inroads and making that happen but we don't have one strategic killer application or product that's able to provide these services for us so people who are solving this complexity problem is going to be a bespoke solution with whatever stack that makes sense for that particular problem domain and quite frankly we're not good at doing that we like things that we're able to get out of a box and apply and you know throw tools of the problem it's going to take a while before we figure out how to reduce the amount of complexity and I think it's going to be a lot of battles that that went that we need to win to win the war we need to figure out security we need to figure out governance we need to figure out operations need to figure out all these things that are in essence moving piece parts to start eliminating the redundancy but some businesses to their credit are moving in those directions they're experimenting this technology apology you know they're engaging Us and other people who are experts in you know dealing with Cloud complexity mediation and getting the ideas and how they're going to have the path to make that better and I think if they're able to get to that point really reduce a lot of complexity then it allows open it opens themselves up for the ability to leverage other Technologies fairly easily including generative AI just snaps in and we're able to manage it using the same infrastructure basically creating versus creating another Silo of technology that we have to manage on its own terms instead we're managing that technology on our terms that's the difference what about the issue of large language models I hear Executives talking about um you know we need to we need to make sure that large language models are really optimized some people are using more proprietary ones some people have larger large language models uh it seems like it's sort of the the under the under the headlines issue of competitiveness and AI how how powerful is your llm do you hear people talking about that I do but and it's really if you think about it it's everything you know ultimately becomes the value in the brain of what generative AI any system can do you remember without the data and without the without the uh uh the llms that are built using that data it's it's really nothing that's core to everything that it does and so there's nothing magical about that the more knowledge that's in those systems and the more ways in which we're able to get it out the more value that generative AR other AI systems for that matter is going to have so the race is on the ability to build these things who's going to have the best large language model and who's going to be able to elaborate the most unique and innovative ways and you know we're using the stuff online we find it very amazing I can write a thank you note and you know even write your bio or write a resume and things like that but when we get to the point we're going to have certain specialized areas industry specific areas you know for example Finance modeling and Retail modeling and things like that where it's not only a big brain that provides general information but a big big brain that looks at particular specificities around particular Industries and able to do things very well Investing For example that's going to be huge your ability to invest using large language models that are able to you know take a lot of that stuff to the next level you know leveraging current and historical data and that's kind of where you get into where the race is right now is who's able to build those things in a certain level of quality a certain level of Dimension that's going to be applicable for particular businesses we've already done the general purpose stuff like chat GTP things like that what about Finance what about travel what about retail and what about specialized industry specific solutions for making that happen all right well I think the big question is that the future of cloud and AI is a combined entity and we've talked about this a bit in the past but I I think it's a strong possibility they would they would combine so that AI eventually becomes almost an invisible omnipresent part of every cloud-based tool instead of a standalone tool to purchase uh agree disagree how do you see that evolving no I completely agree I noticed it's sneaking into developments uh with the you know AI code based generators which are incredibly useful tools for people who are developing and and building models and sneaking into devops it's certainly sneaking into operations we've had AI Ops around for years the ability to apply Ai and security and do security better um you know we always joke about AI washing that is kind of part of every conversation now because you want to you know be you know be the Cool Kids on the Block with the AI capabilities but the reality is it's useful and applicable in lots of different domains now because the cloud is made it ubiquitous and also made it cheap so and the core thing with AI was always the fact that it was just it's too expensive to deploy you know even back when I was building you know AI systems when I was much younger um you know they were going to be 50 million dollars as the buy-in price just to get the base functionality that's not the case right now it's it's you know practically free you know even the cloud providers are charging for it but the amount of value that you get back from it is certainly discernible and certainly something and everybody understands even if you're not a technician you see the fact that this stuff is cool it's able to do stuff that normally I would uh I wouldn't be able to have a technology do and so and it's applicable in any number domain number of domains that we're seeing that right now interesting so I mean before we wrap up I'd like to get your your personal sense you know it's it's funny since chat GPT came out last November I I see more of a sense of Doom around AI than I mean it's certainly an excitement about it but there's a vague unease that before then I it seemed like it was only optimism about AI but now there's a feeling like whether you're Doom there's job loss there's there's uh shape-shifting we don't know where it's going do you hear that feeling of anxiety or fear or not necessarily well I I I heard it all my career um the reality is every time you have a merge of Technology we had co-generators and we had ai's been around ai's been around since the 50s and conceptually it's had kind of the same understanding of the value that's going to be able to bring and the reality is we deal with technological changes all the time it's a constant with business and what we do and certainly we had automation of uh you know farming systems and all these things that kind of come into play and we adapt around it and normally it has a positive impact on society it's able to in essence automate things that we may not want to do as human beings and we can be left to what we do we're creative Innovative Innovative beings that are providing and creating this technology so what we're going to see that the business is going to adapt around the utilization of this technology there's going to be more opportunities in the environment certainly for technicians and the ability to kind of Ed of this stuff to add value to our Human Being Human living is is where the value is going to bring and so I see mostly upsides there's always going to be some downsides certainly Industries are going to be disrupted and jobs are going to be disrupted with this technology but I think most of it it's going to create new opportunities for those same people where they're going to be able to have a better life make more money you know just kind of enjoy life better so I think it's a net positive and where we're looking to go with the technology I like I like your optimism on that uh and on that note uh thanks so much for talking with us today David I enjoyed it learned a ton as always and please come back and talk with us again sometime anytime James thank you

This transcript was generated automatically from the video's captions and may contain errors.