Independent Analysts Recap of IBM Think 2023
Transcript
foreign [Music] you know what I'm going to say that if you haven't had this crew of independent analysts at your event is it really an event we are our schedules have been uh chocked full of events later we were just uh talking about this in a green room that there are so many events going on there's so much travel and we are going to try and cover as many of those events that's relevant to our audiences our Collective audiences over the next few weeks as possible
so we we have some some great content schedule we have Maribel Tim Larry welcome back to I think we're going to do this often enough that we probably need to name it any suggestions Tech talk time I don't know well we'll think about this right work on that and we'll come up with something so as we're thinking about names let's talk about the big conference from this past week there was a lot going on from nutanix next to uh Tim you went to Informatica World there
was just a ton of conferences but I think the biggest of the conferences were at least from a company perspective was IBM think IBM the cloud and AI company I think I was there and I want to say IBM the AI in uh some other stuff company the it was mainly about AI we'll get into all of the things that I'm excited about some of the things that you folks uh witnessed from afar and I'll try and fill some gaps on uh where we're at but
let's take a high level tour of the conference and what you folks thought let's start with uh Larry you're you're the most you seem like the most energetic right now you're fresh off a trip from India and you you you're bright at Brighton bushy tail what what did you think of the announcements what stook out to you what what fell short yeah see I I look at IBM as one of the leaders in AI back in 2008-2009 when they beat the Jeopardy champion what IBM missed
doing at that time was monetizing it in a fruitful way that the company can actually make a difference and they tried they tried in healthcare they purchased healthcare companies and sold it off but what I found here coming in there they're going after being an arms dealer for companies wanting to put their own models and provide the resources to train models you know what's an X what's an X data what's an X governance those three fill in the need for companies saying I'm not going to
depend on a public model and I'm going to do my own thing we're seeing that in law firms I mean I don't know if you have heard of Harvey and Harvey AI that have very specific models they're building for different law firms being used very large got a huge amount of funding so that vertical level of ability to train models and do something that is governed IBM is going to be a good Steward to help a company on that path the other thing that IBM has
which is you know now being promoted more and more is IBM Consulting that can come and take all this technology and make it relevant for companies having said that you know they have a partnership with Nvidia you know that they're going to start bringing in that you can train your models in so those are all the good things from IBM that I heard what I don't see is there are going to be startups that want to build on AI and will those startups build on IBM
uh and and you guys can correct or you know Keith you've been there why would a startup use IBM to build something with and will IBM be able to support a startup that is spending maybe only hundred dollars to train a model on some other platform and will they will IBM even have the interest in going after those startups who's starting at you know that amount of spend and come in there so for a large company needing advice all the tools to go to the AI
model when every company is saying I'm either in a weight of c and I'm you know maybe kicking the tires IBM may be a good company for them to go to I will uh respond with the with that inquiry with one kind of antidote which was the afternoon keynote on the second day was all about AI standing room only for afternoon keynote and they hit the founding of hugging face right for the majority of the time and that was an extremely well received in that IBM
friendly world of the conference area you know how we're in the conference bubble people are IBM customers and they love it I really would love to hear if one of you want to comment on how you perceive that session being afar and not in that uh romanticized bubble of a conference yeah yeah let's hear from you yeah just quick that's the only one I actually listened to from it was promoted on LinkedIn and I had the invitation I said okay let me listen to that but
but you're right Keith hugging face was a little bit of a differentiator you know in going towards startups yeah go ahead Maryville yes it's it's interesting like hugging face is definitely having a moment right now right so if you're not doing open AI you're doing uh hugging face and working with them on that and uh I did get a good chuckle out of him talking about like the Woodstock of Open Source event that they just kind of had which was pretty pretty funny but a few
things have happened right to think always was one of these more AI driven events uh two things that they brought to it that I thought were kind of interesting one was they brought the notion of hybrid cloud data to AI right so we we have a challenges organizations that people have data everywhere so one of the things that they did talk about that I thought was kind of interesting was people having data everywhere and how do you deal with that uh from accessing the data from
governing the data so I thought that was an interesting angle that they brought to it that they hadn't brought to it before it was also interesting to see like the real big return of Watson you know there have been a lot of discussions you know with the with some of the things that happen both around their Consulting business and around Watson Healthcare and other things that people are kind of like what's going on with Watson and now all of a sudden Watson's back right the third
thing I would say is um we've listened all of us have listened to a bunch of discussions about what's going on with generative AI from Microsoft from Google even from AWS uh on to Larry's Point marketplaces I think are going to be big right so AWS is definitely going to take it from the marketplace point of view because they don't have a lot of their own applications Microsoft and Google are trying to Anchor some of that around their applications since IBM isn't a real big application
company I think they really went deep in on the tooling and the tooling to do Ai and that kind of made sense from a positioning standpoint one of the things that I think we've all discussed maybe in another session or certainly together is how difficult this is to do and how people really need to build the equivalent of you know the software development life cycle but in Ai and in AI not just for machine learning which was ml Ops but also AI in terms of generative
AI so I think we got a lot of interesting toolkit discussion around that they talked a lot about governance you know we we have to kind of figure out you know not having been there I know Keith you had some discussion more direct discussions around that I was at another event at that time at the Intel Vision event which was very close actually one of the things that uh they've always been looking at is how do you do things like they did an AI 360 explainability
several years ago sort of an open source thing and now we're sort of see the evolution of that in terms of how do we think about trust and models so the biggest issues that we hear Enterprise leaders talking about right now is not understanding if they use the big large Foundation models if that is going to have either issues around copywritten material or trust of the data biases in the data lots of issues around the data so I think some of the tooling that IBM talked
about really spoke to those concerns I also think Microsoft had some really interesting discussion around governance as well we've heard less governance discussion from Google I think in their i o they talked a little bit more about that but in their original announcements it wasn't as strong of a play so the the long and the short from my perspective as an outsider looking in on what happened at think was that they really tried to address what they were hearing as real business concerns related to AI
so instead of it just being like Oh yeah it's amazing and you can do all these things it's like well let's talk about how we make the AI work and we can argue whether or not it we're really detailed at that but at least we're at the beginning of trying to address some of the real issues Tim yeah you know I think that that's a great um great kind of wrap on it uh from my perspective not being at think and I was at Informatica World
and also zendesk relate this week um but comparing and contrasting a little bit of what you were saying uh Maribel is that you know in the past think think has been this very infrastructure focused you know tools focused show and you're right IBM historically has not necessarily been um seen perceived as that application provider right it's been further down the value chain um still critical but when Watson when Watson gen 1 came out um it was an amazing set of tools that you could tap into
but the reason why I really struggled to get off the ground was that enterprises didn't have the education and experience and capability to really capitalize and and ingest it and take advantage of it only the biggest companies were able to to have the girth and and the expertise to be able to do that and so I think that you know that's great from the standpoint of um here's an amazing tool set that being Watson but it just wasn't easily consumed and I think that's that was
a that was a Miss for IBM in the past fast forward to this week and we're talking about Watson X and you talk about the governance you know there's the data component the AI component I think they've really learned from that mistake and really tried to make this more consumable for the average Enterprise and to me that's a huge huge Plus for IBM you know IBM has always had to think about how do we get beyond our base how do we bring in more customers into
the fold than just the existing IBM base of customers and I think when you start to take something as sophisticated and rich as Watson and you start to break it up into components that are more easily digestible by an Enterprise especially in this day and age where there are a number of tools out there that are looking to do this um it's it's really kind of a plus for IBM because they have that expertise they've been down this path for some period of time more so
than their competitors and they can bring that to the table for the average Enterprise to be able to consume I think governance and policy is going to be a big um a big issue for Enterprises it is today and going forward it's going to get more complicated as you start to think about privacy and Regulatory and compliance requirements not just at a nation-state level but then looking within those nation states like just within the United States you've got regulations in New York and California et cetera
and so being able to automate that around the data in acceptable use of data is going to be incredibly important and you know Informatica really kind of jumped on this because they're they're all about data that's that is their their thing but we also have to look at how does this kind of start to move up the value chain so that the average person can start to use it and it's not just the super high-tech sophisticated tool that only the most sophisticated customers can use I
think IBM's made a great step in the right direction I'll be really interested to see kind of how this plays out and especially as you were saying about hybrid cloud how does that kind of layer in with this as we go forward yeah so I am I'm gonna if if you're if you work from IBM and you're in the analyst section about the analyst session about hybrid cloud and I stood up and asked the question I'm gonna you know I'm gonna borrow a little bit of
my Spiel when I ask my question which is IBM the hybrid cloud and AI company the you know it's kind of like when I'm listening to my friends the cube and they say the leader in uh coverage I think we'll we'll challenge them on that in a little bit though but I think one of the things that I found very interesting about the conference overall there was very little hybrid Cloud conversation directly that did not include AI this was a AI conference make no mistake about
it IBM is betting heavily on it while Arvin the uh CEO of IBM by the way take there was a outstanding AMA with him for analysts from start to finished there was no opening comments it was just questions from analysts and it is by far one of the best amas I've had with a CEO of a major tech company but I digress again this was a conference that was about uh uh this was a conference that was about Ai and how AI will impact the hybrid
infrastructure this was my first combined IBM and red hat event so I didn't know what to expect overall and I was fairly disappointed in the level of conversation around hybrid Cloud I've been covering IBM for almost a year now uh you know it's a well-kept seek their hybrid Cloud solution is a well-kept secret so much so that one of the most interesting parts of their portfolio is this IBM Cloud satellite offering think of uh as you're a stag or Outpost or uh anthos it is their
competitive offering to those Solutions and it on paper may be the best solution out there for for most use cases if you can think about taking the AWS control plane and putting it in Google cloud or Azure or in your data center without buying Hardware this is what this is you can take IBM's cloud and run it in your infrastructure whether that infrastructure is another public cloud provider or Colo or your private infrastructure is you're able to do that and I asked them the big question
which was well what's the minimum commitment and what's the uh stand up time if you're standing up in another Cloud the speed at which you can get your terraform to run is how long it takes to bring up the service and then once you bring it back down you stop being built for it it's truly a Cloud solution that's if IBM Cloud itself has all of the services that you want for it so overall it was a great show I ran into business users not just
technologists uh I think I sat at a table with a white man I told her that I was uh a infrastructure professional and she was oh you build manufacturing tools I'm like oh no no no no no no not that kind of infrastructure person I am a I.T infrastructure person and she didn't even know what that meant so that was that was a great indicator that they they had uh the Right audience there to talk about the outcomes Associated now I think I want to open
this up for General conversation and we just talk about uh questions you may have for me around whether it's Watson X I know what the X stands for uh and uh uh any of the AI announcements Etc I think I think I I I while not a data or AI guy absorbed way more about AI foundational models versus llms versus uh spatial Geo models Etc I I am now uh AI expert or talking head expert at least I love it so um you know one of
the things that I think is kind of interesting that I think Larry alluded to in his opening remarks um is this come is this concept of we've got the big foundational models that we talk about like the GPT 3-4 uh but there's also other models and you know IBM has actually been you know it talked about its data cards and you know the whole pipeline of doing things and I'm just wondering did you get a sense that they were really you know pushing some of their
models for the Enterprise to pick up no I think one of the they really highlighted if I hear the term hugging face one more time I will like go hooks because it was they really wanted to drive home whether you're buying a model from IBM building a model or bringing a model they want to be your home for uh uh either training or using inference on that model they want to be as hybrid as possible uh so much so that you know you look at Watson
x dot AI that is where you bring those other models from to adopt I think what they wanted to drive home is that the best but you know that Better Together story when you use all IBM solution you're going to get the breadth of uh Watson dot AI watson.data Watson x dot a DOT washingtonx.data on Watson x dot governance you get the best of all worlds but any one of those portions of the platform can be separated out used as Standalone you don't have to have
all IBM solution see this is what I think is really different between some of the positioning that we've heard from say Microsoft and Google where they talk about a lot about you know what they're creating from a model perspective and what developers could have access to and what types of intelligence will be available so it's actually interesting and it's also very I think Amazon also has a very different take on it so this is really an interesting time in AI to see how each of the
different hyperscalers are trying to play to one of their strengths and see you know what they can push forward with their customer base yeah I you know one of the things that kind of um stuck out in my head is when you talk about hybrid cloud and you think about it from a customer standpoint for a minute there are a number of companies that are touting they have the best hybrid Cloud solution on the market right so now IBM's there um we could say the same
about hpe hpe has been talking about hybrid Cloud for some period of time they've even been talking about AI I think it's a different data set that they're they're looking at versus what we're talking about here but then you look at the hyperscalers you know the Google Microsoft Amazon um and they have they all have the infrastructure piece of it but Keith one of the pieces that stood out uh here for um for coming out of think was the AI pieces sat on top of it
for these other companies but it seems like there's more of an integration between those two conversations coming out of IBM more so than some of the other Alternatives is that part of what you were taking uh taking away on the ground yeah and I wanted to tease this out because when I heard the umbrella term Watson X especially the wax the next data I really had a adverse reaction when I first heard about what I thought it was I watched next data that data when they
introduced it at the keynote stage they introduced it as a open data store for data and I'm thinking oh as a I.T infrastructure person that's exactly what I needed is yet another data store to solve my data store uh sprawl issue exactly the store for data store yeah datastore for my data store and they call that metadata by the way that is that's very that's very my time is done here so what uh what IBM took pains in doing and when I went and talked to
their Chief scientists around Ai and the owner of the Watson X brand Watson X the X Stand stands for scale so walks and times whatever the the thing you're going to do dot AI which is the models Etc dot data which is the data Lake and then uh dot governance which is the governance around the data itself each one of these components whether how you consume It ultimately with DOT AI and openshift is something that you don't need to start with IBM to get the advantage
of which I thought was really really smart uh the other smart thing is that they're sub uh uh they're deprecating the Watson Studio which to your point Tim was the big learning curve in watson.ai is replacing that so they're resetting this Well I this idea that using Watson is hard or Watson X is hard they're focusing on a single way to do this across platforms whatever you choose from the underlay and I don't think this is something that I've heard from other Cloud providers that you're
able to separate their AI tooling from their infrastructure itself so when you're looking at what AWS is doing what Google is doing it's a holistic story and not necessarily a story around uh Lego over here was IBM can learn uh internally of what they can do with AI and then start using it for providing value to customers because when you finally look at what our customers looking for from AI they want to get productivity gains when arvind about a week or two ago said we can
not hired for 8K jobs or 8 000 jobs it made some news and shook something saying hey what are you going to lay off 800 8 000 people and he never he said that I will you're going to lose people through iteration but we're not going to replace them and we are going to now be in a 27 000 or 30 000 organization be able to eliminate 8K jobs over the next three to five years through attrition everybody missed that through attrition point and just said
you'll go to lay of 8 000 people and use it but that's not the point the point is now they're going to learn from internally what can they do for back office processes to automate that with AI and then make that available to customers that you know you can improve your productivity especially a large amount of value that can be multiplied just like you said scale when you think about a retail or a fast food company now there are more people talking about fast food companies
you know replacing that drive-through with AI that little bit of efficiency you gain can can have a significant value on the bottom line so I think that's something that you know IBM might be able to provide value in the future I don't know what about what you guys think about arvind's comments and whether that's something that can be used so he expanded on those those comments uh give credit where credits to Dan Newman followed up on the this comment from Uber or doing his Bloomberg interview
and Arvin gave us the rest of the line that Bloomberg left out you know their Discretions leave it out or not which is he believes that IBM will double in size as a direct Resort so yeah the jobs that are destroyed will be replaced with better jobs you know there's a a lot of just social and uh politics around what that means like what type of jobs they'll be replaced with and who's has access to those jobs but I think it is undeniable that when you
increase productivity historically you increase opportunity the when you abstract away the complexity you allow people to do more things one of the ton-in-cheek things that happened was one of the executives shared a story about how a Auto shop is using generative AI to reduce or increase their productivity by 25 percent and how the husband that was using the generative AI helped his chemistry professors teacher use it to create tests exams for her classroom and how this you know non-intuitive direction of the kind of blue collar
workers taking hold of this technology and teaching the white collar workers to use it was pretty quick and I know we're coming up on our usual cat we might end up going over about a couple of minutes the one the one thing I did want to stitch together which I don't know if I I think intuitively I believe it but I haven't really fought this through two is through is how AI will drive a hybrid cloud and they gave a lot of examples of data being
collected at the edge at the data center not in the cloud and bringing the entire tool set to the edge so when you think about satellite which will bring me managed openshift which brings me openshift Ai and all of the tooling I need to process the data at the edge IBM has a compelling story there so final let's do some final thoughts uh let's go on reverse order Tim let's start with you I you know I just I'll just build on on what you just said
I do think that as Enterprises really kind of start to to move up the value chain move up that stack right and move closer to trying to gain those insights as quickly as possible they will they will want Automation and want a closer tie into the infrastructure and the underpinnings of the technology so I just have a question to ask I don't care how it how the system Auto magically Works to to make that answer come about um and the data May reside all over the
place and so we have to think about the integration pieces but I think the underlying infrastructure there too those are going to have to come closer together than what we're seeing and so I'm encouraged by what I'm hearing coming out of think and I think there that we're seeing this um as I mentioned you know informatica's really going uh gangbusters in this space but there are a few others that are also trying to connect the dots there so I think that's the real opportunity for these
companies is to start to consolidate and tie up some of these functions to make them more um more consistent in the way that they operate and more approachable as a tool for Enterprises uh all right as I mentioned earlier I really like that IBM is actually pulling together a nice set of tool kits that meet with you know what Enterprise issues are but the rubber doesn't meet the road until you actually connect that to data so what I'd really like to see is you know the
kinds of Partnerships that IBM is going to have with different applications or other things where you know most the reality is we talk a lot about ai ai strategies are really difficult to deploy and the ones that I think will get deployed first are when you can anchor them to some application or data source and try to do something different like what is that generative AI work on so until we see a little more of that out of IBM I think it will be a really
nice interesting set of offerings but you know won't get traction until we see it actually connect with the ecosystem and layering yeah I think the lack of skills in Enterprises is still a real thing for making AI work and IBM with their Consulting and their Partnerships with other public Cloud providers like Amazon and others can help them give a customer a you know story or a narrative that can be actually implemented and you can gain value out of so I think this would be a an
interesting thing to see where IBM and can play a role yeah and I'm still processing My overall thoughts from the conference I I'm unlikely to write about a single vendor and I like to expand out kind of these industry wide views as well as u3r as well what does IBM think announcement certain approach mean for the overall industry where are the challenges how are customers trying to solve this because not every customer's IBM customer and uh I think it helps to understand why folks on IBM
customers and why maybe some should or shouldn't be you want to learn about our individual Brands uh you can visit each one of our websites which will be in the description below all of us are fairly active on social media our Twitter handles and Linkedin profiles are both in the description of this video and in the show notes if you're listening to it as a podcast talk to you next episode of insert Title of shows or whatever we're gonna call this whatever we're gonna do whatever
we're gonna call it excellent all right folks thanks a lot bye thanks man thanks ciao
