20240613 AI Platforming

    10:41PM Jul 5, 2024

    Speakers:

    Rob Hirschfeld

    Rich Miller

    Keywords:

    data

    ai

    google

    oracle

    hubspot

    customer

    crm

    sap

    platform

    vendors

    aws

    cloud

    companies

    integrations

    question

    building

    agile

    big

    work

    product

    Hello. I'm Rob Hirschfeld, CEO and co founder of RackN and your host for the cloud 2030 podcast. In this episode, we discuss the impact AI and data sovereignty data protection will have on platforms, on consolidated management of your data, things like office by Microsoft or Google, on premises, systems, we didn't talk about as much, but similar, similar idea of you have a whole bunch of data that you will want to use to train AI models to improve your day to day operations. But how widely that data gets distributed to me and the framing for the question of the discussion is that you probably don't want a lot of vendors pulling that data apart and transiting it. So that was the framing of the question. You'll hear me reframe it a bit, and then we have a really fascinating discussion about how the market is impacting these forces

    when IBM is selling Watson X and and there's, there's a lot more Z getting purchased, I guess the same kind of question to that I posed to Rob is what, what seems to be changing about the way they're doing data center network?

    It's a no, it's a good question. There's it's funny, because they're building it out for quote, unquote, end to end. That's their, you know, new driver, whether it's in manufacturing or other industries. So on the data center side, they're they're dropping hints of what they're doing, particularly in terms of how they're going to be connected, but they're not actually telling you what's changing. It's kind of a they announced things. I think they dropped hints. They're dropping hints, but they're doing it by industry sector, as opposed to just overall. Here's the data center strategy, which I find interesting, but I think it's because of what kindred will offer in managed service as opposed to what IBM will sell in data center equipment. There's a relation, clearly, because the relationship, but I'm seeing it more as how do I put it, if you're going to refresh what already exists and you need to bump up, put a V in your data center on brown, don't you know? And maybe use IBM Cloud to an extent. But other than that, they're not really talking about the connectivity, whether it's mesh or, you know, like how things are going to work. I actually asked the question, a similar question, but not that question, I think, and I gotta stay tuned, as an answer, hmm,

    I know enough people at IBM that I was surprised by that

    there's an interesting um, actually, I'm gonna, if sorry, I would love to, I'm gonna step into the comfort the planned conversation, because We're 20 minutes in and Yeah, go for cool. All right, let me make sure I pull up. All right. So for today, this is part two of the conversation from last week, but I wanted to take it from the other direction, which is sparked by Google buying HubSpot, where you know this idea of platform convergence made even more urgent by the need to feed the AI engines for people's data. So the the idea, sort of behind the conversation starter is for these, assuming we're going to be training AI on on everybody's data. They want to do that like so, so I'll put rack ends name in it, just to keep things simple and specific. So RackN, we're a Google we're a Google customer. We're a HubSpot customer. Even easier and. We have a huge amount of data about ourselves, about our customers, about our product, that that we between those two sources. If there was an AI engine ingesting all that information, we could do a better job engaging with our customers, you know, and prospects, because of all that information I could send them more appropriate, you know, you name it, the whole system could be beautifully integrated with a whole bunch of AI data. So what I was seeing with this is a real drive from these platforms to want to pull in more and more data sources to create more vertical integrations, because on the other side of it, I don't want to send my data to a whole bunch of different companies to do redundant analysis. It's leaky. It's risky. I guess my sense on it is from my own business perspective, if HubSpot and Google got together and offered me, you know, a AI training system that ingested all our data but stayed within my Google accounts, that sounds actually really attractive to me.

    Yeah, when you said, Stay within your Google accounts. Is the is, does that? Is that predicated on the idea that you're getting use of all of your HubSpot data Google's AI improving your own business, but are you also generating parts of the data that actually go out to your customers, or using customer data and pulling it back, is there a is there a so an exchange, or a share of, at least a partial Share of data from

    this is where we're using rack as an example, is maybe not as helpful, because we we are pretty fastidious about not pulling customer data in. So this is stuff about like our notes and meetings and our positioning and stuff like that. So okay, yeah, so I'm not worried about leaking one of our customers, you know, internal information out that other companies, yeah,

    well, one side of it is concerned. The other side though of it is if, in fact, as a part of your added value, you are capable of taking. Well, scrubbed, appropriately protected data from various parts of your customer base, incorporating it into your knowledge base that's feeding your own, your own AI's well, your HubSpot, but also other parts of your and then using that as a means of improving, either optimizing, you know, benchmarking things like that, right? It may not be as relevant for what you are doing, but in other other layers of I can see that being a a very interesting aspect of the combination.

    It it's, I know our customers are worried about stuff like that happening. So it could be, it could be an interesting combination, but you'd want to do it very deliberately,

    obviously. Right? Sorry, is not meaning to get, you know, to cross a boundary line. But is HubSpot your content management platform, or is it your CRM?

    It's our CRM. It's not our content management. No, we this is the chat, right? It's a huge platform, and we could use it for content management, but it's really just a CRM for us, and frankly, it's it's isolated enough that it's a clunky CRM, because

    that's kind of why I asked the question, because I could see it as the content side. I couldn't necessarily see it as a CRM. Been there, done that. Wouldn't do it again.

    Sorry, that's what they but HubSpot sells on the on the premise that, you know, we're a one stop shop, you know, you get, get your CRM, get your content management, get all of these other good things, you know, nice quote, nicely integrated, the fact that their CRM sucks, okay,

    but, but hurts a

    little bit. But I agree. I don't think there's any great CRM, but maybe I'm maybe I'm wrong. I.

    It depends on how you use it and what you're using it for, but, but to your point about the integration with Google and using the AI, I could see it being used for,

    and I'm gonna the words will not sound really well good together, but call it predictive customer analytics. Oh yeah, right, if you wanted to use the AI capability to take the 28 steps of the digital buyer who now does 73% of their work without ever speaking to a human being in the buying process, and shorten that to something more rational, like 10 steps, as opposed to a 28 step customer journey to close a deal, I could see it being very valuable For those kinds of things. But to your point, I would look at it as being a seven layer cake, slash partnership, slash all these other components that would have to go with it to make it feasible. Because if you look at the way buying is going on in software in particular, marketplaces have taken off, and they're all about partnering and alliances. So if you were pulling feeds in, not directly from the customer, but the customer's other partners, then you'd have something really interesting.

    Oh, I see what you're saying. It's

    basically the the way you start to build out a kind of a multi sided, I'll call it a market, but it's more. It's a it is a flywheel. So going back to your original question, or the question you were posing, Rob, you want to just kind of restate it again. So,

    so what I'm, your point, what I'm what I'm seeing, is a walled garden supercharger in AI that that you know that the the interests of protecting your you're going to need a partner. You're going to need a vendor. But you don't, you know, we don't want to have 1000 vendors where all our data is getting spread everywhere. It might not even be feasible, because there's so much so the idea that I'm going to consolidate my, my, my, my data footprint, my interactions around a vendor, so that I can then consolidate it, AI, seems like it's going to drive that process even faster.

    And the data, the data you're talking about specifically, is platform based, back office.

    It's, it's a lot, it's back office. It's, it's documentation, it's research, it's glossies, it's, it's customer engagements through email, right? To join point right? None of our customers actually really want to get, you know, they're getting, actually a little better about wanting to get on the phone with us, but a lot of them want to get pretty far down the path without any human interaction. So if they felt like they were able to, you know, ask questions on a bot to get real, you know, real answers and find our you know, that would be amazing. It would be great. They're not going to get it out of chat. GPT is too specific,

    no, but so, so here's a thought, Rob, just just to drop this in there. I know of two other companies that are creating their bots around qualifying their prospects and even their customers by asking them the leading questions around their technology stacks to then be able to sell to the specific requirements. So it's a requirements, it's a set of prompts to get the requirements of their current situation, their problem and what they're trying to solve, and spit that back to say, we can make your journey that much faster by eliminating, you know, Step nine, Step 12 and step 15, right to get to get them in The door, but they're doing it in by using the data that they're using to train the model is all the existing content and marketing information that's on the web of that company, and they've written some out and some interesting algorithms to try and see how they can pinpoint, like, okay, they use x vendor for y function, right? They use HubSpot for CRM, or they use Salesforce for CRM, because you can tell based on the content and all the announce. Press releases and everything else. So they're running AI as an intelligence tool right, right to then be able to serve, quote, unquote, turn the prospect into a paying customer faster, and doing it all as this sort of notion of we're taking our own data and those of the primary partners that we have, or where we can digest or ingest the information from public domain, and say, Okay, so your stack basically kind of looks like this. And we know that. You know you're looking at us for X problem, but did you want to look at us for why problem as well. And here's where we bring more value add. And they're building that flywheel all from using chat, GTP or lemma.

    Well, they're building agents. They're doing they're, they must be doing agents on top of chat. Yes, okay, yeah,

    yeah, yeah. They're not using chat. They're using open AI. They're not using chat, yes, that's what I meant to say, yeah, yeah. And

    or other models,

    exactly. They're using, they're using llama three, they're using Quinn, they're using anthropic and so forth. And the whole idea that you're positing here is as a both an intelligence and marketing and then sales, you walk in the door to a potential customer or an existing customer and say, This is what I know about you, or this is what I suspect you're doing based you know, and you're immediately, you know the customer is immediately impressed by how much you care about, know about their their situation, and that you've thought through and suggesting, you know ways of expanding the expanding view the for the sellers. Participation, in that. It estate, uh kind of land and expand in there, okay, yeah.

    Also take

    Yeah, take ownership meaning, I'm sorry when you say take ownership. Take

    ownership by being a more rather than a collaborator, more of a stakeholder. Ah, okay, in the process, so And because, you know, like, if you really want to have an eye opening experience, take a look at what Jay McBain has been discussing for the last six months on canalsis. He he's he covers channels and channel partnerships and the alliances and the marketplaces and whatever. And it's very interesting to hear what's going on, because he's likening things, not only, as I do, to the seven lira cake of if you're going to go to market and and look at how you're going to use AI for the task of driving your growth or driving your customer base, how they're doing it, and the methods that they're using to do it with and without AI is so Much different than it was even two years ago. It's all about your partnerships. It's all about your channel, how you're creating the channel, of channels, or the ecosystem, and who you're co selling with, and who you're driving with. So the Google HubSpot thing fits that model to a T in okay? It it's very interesting to hear what's driving revenue for everybody, and it's certainly not the old way of, you know, I have a big no pun intended Salesforce, and I'm going out, you know, covering massive segments of huge markets. Doesn't work like that. It's all about the 20 top top 20 GSIs and the ecosystems that each of them have built.

    And what Google wants to do is replicate what AWS has done,

    right? AWS has done a marketplace, okay? Well, AWS is has. You know, for years, AWS always seemed to be a step ahead or kind of ready to announce a product as soon as people started, you know, making noise. God, I wish, you know, AWS did x, and, you know, a month later, X shows up. You know, they couldn't have developed it in a month. So, yeah. Your point being that this is, this is kind of anticip This is the best way of doing anticipatory development. But also, as you said, owning, kind of owning the customer, yes,

    but what AWS does on, to some degree, Google as well, is they watch what their customers do with their platform, right, right? And you know that they observe the trends, and then if there's a gap there in their platform offering, they work towards filling that gap,

    billion, where we definitely were being

    that the Google like, once that trend goes away, they drop the product,

    right? Sorry,

    and yes. Where?

    Where Klaus? A good example from Google. What do you what do you point to at Google? Okay,

    it's a little bit more complicated than that do to be honest, right? But it has that endemic problem in that they value new products over maintenance of existing products. And you can see that like, there's a whole web page, like killed by Google. And these are products that, in many cases, have been replaced by other equivalent products. And Google actually promotes that kind of behavior, because, again, as an internal developer team at Google, you are more valued when you create a new product than if you take an existing product and maintain it. Yeah.

    And I ask the question, are they going to kill Oracle? One can only hope

    Oracle the whole platform or Well, yeah, did

    you not see the announcement? Two days ago? Oracle's infrastructure is being integrated into Google. What?

    Yeah, no,

    this of the day, oracles. But I was a bit late into the game. That's fine. I We just missed it. Oracle

    into Google Cloud. Like, are they? Yeah, buying Oracle, Oracle Cloud. It's an interoperability partnership.

    Oh, no, I,

    I

    if there's a multi cloud, I believe it is, then this is more related to Google's Cloud Interconnect.

    It's a multicology partnership, yeah, the multi cloud production, the whole, the the big, the big deal is, as one might imagine, data, the tariff, the tariffs on moving data back and forth, and then the database itself, yeah, the data, data sources, Google

    is pushing down pretty heavily on those capabilities. So not not just Oracle, but really all cloud partners.

    That's a that's a reasonable strategy for Google's, for GCP to get into enterprise, which, yes, they continue to, you know, kind of futz around with, and never seem to actually execute on a you know, which is kind of strange, but, yeah, I can see how they would do that. Okay, so

    go ahead. Well, here's why it's interesting to me. Okay, it's interesting to me for three reasons. One, fintech. Everybody runs on Google, financials. Hey, sorry on Oracle Financials two, it's a next mover to SAP gets into bed with Google. Why? Because SAP uses Oracle databases. And the third reason is because, as they move towards SAP, they're moving into Oracle ERP on Google Cloud, which is a way better way of doing it than Oracle's ERP on Oracle Cloud.

    Because why do

    Oracle Cloud is, if you're if your supply chain manufacturing or industrial you will, it's a much longer and much more. Costly implementation. If you're SAP using Oracle, shoot yourself now.

    Using Oracle Cloud,

    yes, yeah, sorry, class, but I didn't hear what you said for Oracle,

    it might be a win as well, because they still get to sell their licenses, they just don't have to maintain infrastructure for it, correct the offshores that do Google,

    and it's also and it's an advantage to Google because as agile goes away from Oracle, because it's being sunsetted, leaving a huge gap in the market for anybody who does any kind of product development, like all the CPGs, all of the manufacturers, all of the electronics industry, they're all agile users. Oracle is no longer supporting agile. That means, if you have manufacturing or PLM and ERP plus supply chain stuff, you need a new home. Your normal course would be either go to SAP for all of it and struggle, or go to IBM. You can't really, you know, it's kind of a Google defends, it's will take in that that need without it going to Microsoft, because that's the only other play in town on a Cloud that would support that kind of capability

    well Azure, Azure has a foot in enterprising manufacturing, right? So? And even though AWS is making big strides, they're not 100% there Oracle people have been rebelling against for a long time because the overhead cost is huge.

    So, well, I don't follow that as closely as I probably should, but who's filling the gap for agile, other, PLM, support, other PLM,

    just so it's picking up a lot of yeah, just so is picking up a lot of accounts, although it's much more expensive. Believe it or not, the space is changing. So you have more niches being filled, like product development. Propel has done a very good job of filling that gap. PTC is kind of winding itself into oblivion, which is weird, because they were the leader for the longest time, but they haven't kept up. So there's a lot of fragmentation going on in that space. But it's germane to if you want to build anything, any kind of a physical product, you're really struggling to figure out where to go when agile goes away. And those kinds of products on cloud are really difficult, unless you have a really robust cloud infrastructure behind you, like SAP doesn't do a good job with its PLM in cloud. Not that SAP does a great job on rise. But irrespective there's, there's like so much change that's going on for PLM, mes Supply Chain Management, even. ERP, that's the next sort of big play, where entrance will be very interesting. There's a lot of challengers, small companies that are challengers, but they're coming from so many different perspectives that it's almost industry verticalized before it even starts.

    Do you get indications I'm not an Oracle watcher. Do you get indications that Oracle is kind of going to follow suit with what they've done with agile and start to sunset some of their other well, for example, logistics, you know, supply chain stuff. Are they? Are they clearly, you know, putting a bullet in some of those

    to for all intent and purpose. I would say there's two that I know of that I can't really speak about that. I would say that bullet is imminent within the next six months. And I would also say that Oracle wants to make a big play in AI in FinTech.

    So they're, they're insurance, putting their they're putting their chips on FinTech, financed and, and insurance, yeah, and, and moving out of the others, leaving it kind of as an open field for the small fry to to fight it out.

    Um, to to a certain extent, for the small fighter to fight it out, yes, but where the interesting part is, is the SAP and disow side, because SAP lost a lot of share. They wanted back. They're building this business information network, which is their play. And there you're going to start seeing some really weird changes going on two tier. ERP,

    which I scratch my head over, yeah, well, this kind of gets back to the original question. I mean, the the point that Rob was making was the the change in platforms, or utilization and and adoption of platforms, I think, because of AI, because of repay, you know, kind of, I won't call it repatriation. It's, you know, re rebuilding, or, you know, renovating some, some stuff with, with new, new underpinnings. Now you're talking about major players literally deciding not to be a supermarket across, you know, with a with multiple lines of business. But rather, that's what I'm saying, focusing on a few and then allowing the customer to shop for platforms, elsewhere, to incorporate. And so, yeah, there's this is, yeah, this is multi factor. This is, this is, this is getting to be quite a, you know, this is, this is no longer a three body problem. This is a five body problem. Well,

    that's, I, it seems to me like it's, it's gonna pull in tight, like it's gonna pull in tighter, like, like, we're gonna, you know, we're gonna be more likely to, you know, look for, look for vendors that can do that, the full analysis for, for us, yes, because we don't want our data to traverse as many boundaries we or we can't afford our data, right? I mean, this is why we went down the whole Oracle path, right? It's, you know, all of a sudden you're like, oh, wait a second. My, you know, because of the Oracle Google partnership, my, my Oracle data doesn't have, you know, is, is adjacent to the Google Cloud and Google Analytics, and potentially, that makes it that much easier for me to do those integrations.

    Yeah, right. It also, it also, it also enables Google, if they've got the the interest and the money which they do, to kind of take on the job of being the ones you know, this is the way they they compete with AWS to become a one stop shop, but at a layer of abstraction that's that's a bit good deal higher than what AWS well,

    and this, this is potentially where they have an advantage, because we've talked for a while, right? Amazon does not have the office data suites that Google and Microsoft do, and so no, up until AI that was, yeah, that was really a deficit. But now they're hosting,

    they're hosting vendors in their marketplace, and that's, that's the extent of it, right,

    right? And raw, more raw infrastructure. But now, if you turn around and you're like, Okay, wait a second, I'm going to start doing AI analytics my, you know, and this is, this is part, part of what comes this conversation comes from is the casualness with which AI startups are asking for access to my Google Drive so that they can right? And this is, and this is, I was just reading about the Facebook analytical Cambridge, analytical Analytica, where they you think you're giving access to one thing and you're actually opening up your entire social graph to somebody. I yeah, I think that, you know, we're going to start getting a little more careful in the idea that, you know, I do want people analyzing all of my data, but I don't actually want them taking any of my data,

    right? Well, do you really want everybody to be able to analyze or do you want no selectively as a small group?

    I want to keep the group as small. As possible. But I also want, and this is where the marketplace idea gets interesting is, I would love to have a marketplace of things that could, that I could apply to analytics on my data,

    without losing, without the data, you know, without, without exfiltrated, right, correct,

    and even, even better of having some type of Third, you know, Google audit, or, you know, additional boundary applied to it. And,

    you know, it sounds like there's a market gap or data on the organization for the explicit purpose of sharing it with these kind of companies.

    Yeah, it's a it's a combination. It's not just anonymization or pseudonymization. It's it's different kinds of data well, it's probably a combination of data virtualization and obfuscation so that, you know, I can federate data sets policy based you, you know, you establish the relationships between you and your data And whatever vendor is doing the the analysis that that Rob's talking about, but it's, it's well, well managed,

    yeah, for, for the more pricey, conscious the people, on the other hand, there's Also the approach of, well, just moving towards local systems, like, yeah, like SlMs into llms, or portable analytic systems that you just run them on, on your own environment, that they're completely air gapped from the cloud,

    exactly which, given some of the performance people are getting out of the more open llms makes it, makes it an interesting that's that an interesting option, so long as you've got the skills, or you can hire the skills to do that in house.

    LMS, we're definitely getting to the point where those are becoming viable, because they're good enough, and there's enough development happening that it might become just a commodity in the future. Not seeing specialists

    that have that close are you seeing firms that have that kind of specialization or that, you know seem to be adept at that kind of that kind of development?

    I don't see it in in companies, whether startups or enterprises. I see it more in the open source ecosystem,

    I agree, yeah, yeah, but

    it's likely go like just amount of time before someone figures out how to essentially become a service provider for that kind of thing, like they don't need to make money from selling it. They need to make money from helping all customers use it

    because of the maintenance and the updating. Yeah. But what I'm seeing is some companies are looking at like, let's say it's net new buy on software. They're looking at the number one, the number of integrations and how those integrations are ranked. Number two, they're looking at, how flexible can I make this so that if I do run a slim against it. I'm not. It's not so much about data leakage. It's about the context that's being created. They're more interested in the context of the individual streams of data and how they play together, like think customer journey. Robin and

    I was going just on, you know, you're gonna, you're gonna need to every interaction that you have. You're gonna have to go back and scan through, find all the interactions to that customer, do sentiment analysis pieces like that. You You know, hopefully you're gonna start in tracking metadata or storing intermediate components for that, right? Because that would be useful,

    yeah, right, but it's the storage side that's a bite. You How much is too much?

    It's interestingly. This is, this is where I think it would get interesting. Is, if you know, and I think CRM is an especially interesting. Place, because the the ROI is so high, but right if, if, you know, HubSpot right now doesn't do any sentiment analysis for me, on, on customer interactions, and they don't really, you know, they I can build automation, but the opportunity to have my, you know, in, you know, CRM, which really ends up being email CRM, telling me, Hey, you're interacting with this customer. They have this sentiment. They're showing buy signals. Maybe you should send them this piece of collateral, you know, start actually pulling all those pieces together. And then what I would love to see is that would then end up in the audit logs for the customer interactions, which is really what, which is really what should be happening with that,

    I see that part of the automation actually starts to show up, you know, without, you know, kind of hands free, not touched by hands into the CRM,

    into The CRM. And then you would actually have a scan that would say, every night, I want you to email people who have, you know, positive sentiment, or something like that. And then, you know, I can even see doing, like, custom newsletters or, you know, I mean, they're the opportunities for this are amazing personalization, and then, and then, for me, observability of that personalization is really interesting, but it's it can't happen, you know, it's really hard to do today in HubSpot because it's so isolated, not because the tool is clunky. I mean, you know, let's give them the benefit of the doubt. Even without it being clunky, it's still very isolated. Yeah, so from doing that work, it's it, and it's hard to tie it into all the other ways that we interact with our customers. So

    does this imply a kind of a a more permeable enclosure to what has been data silos,

    I think that AI driven analysis and then prompting, you know, so completing the loop, so doing the analysis and then feeding it back. You know, if I if I could pull more data into that process and train it based on the information we already have, it feels like it would be much more valuable, much more practical without it being actually, frankly, a lot of extra UX. The ideal way to do it is it's going to be embedded into the UX I already have.

    So but then the question that I would have is sentiment analysis is not necessarily the same in some cases and not necessarily different in other cases than just the NLP, right? It's the context around each of those pieces that's really more important, and it's how you approach understanding that context of you know, there's a big difference between, sorry, I'm really tied up and I can't talk to you right now, but I'll get back to you. And sorry, I'm really tied up right now. Can't get back to you and I never will, no. But from that perspective, you'd have to be really, really careful about the linear regression that you're doing on the data to make sure that the newsletter you're going to fire out is not just going to tick someone off the

    and ideally that this is where you would start getting into the amount of analysis that we could be doing on these interactions is stunningly powerful, and in some ways everybody's benefit, because you know that you're gonna spend less time reading this those crap emails, right? You're gonna get, you know, you're gonna get things that are potentially better, sure,

    but including chatbots that actually understand what the hell you're saying exactly,

    but I but to me, the first thing to do that is, you know that the consolidate, the consolidation ends up being a powerful component to that.

    It still is highly dependent on, on the data governance and that that whole underpinning, because without that, you know, no one's going to trust it, and all you need is one or two failures and and everyone will, you know, just pull the rip cord, fire the exploding bolts, and you. You'll never see some of this stuff again.

    What does HubSpot integrate to?

    I mean, they have a ton of integrations. The simplest one is just back into Gmail. So, you know, a lot of my emails are, you know, my outbound emails are filtered through, through that, and then anything in the thread that I use gets filtered like it. I mean, the challenge is it just slurps up every everything I'm doing, and it's, it's actually a bit problematic from that perspective of how much you know, how much actually is going through that system, right? You know, I'm not sure how much I trust that. I'd rather, you know, you got to trust people at some point, but adding more vendors to, you know, slurp through my data makes me nervous, even down to like I stopped, I stopped using the frigging Google Drive plugins to do different charts and things like that, because you're granting them access to your drive right to do the work. And I just got, I got increasingly frustrated with or nervous about, you know, oh, this could look at any documents just the way I'm opening things up. That's

    okay. So it's, it's so the way they've done the integrations, they the access is, is kind of all or nothing, or you the access

    that the access with most of the integrations I've seen are, you're granting access to Google Drive, which is very open ended,

    yeah. I mean, it's not, it's not selective. No selective at all. Oof, Oh God, I know I wouldn't, I

    wouldn't. Well, yeah, actually, though I wrote, sorry. I wrote a rule in my Gmail for the Google Drive to not allow it to suck up or replicate certain documents and only keep them local smart, if it was an attachment to an email, or if it was an email I was sending to myself with a document, like something that I was thinking about or whatever, because I do that fairly often. If I get a brainwave, I just, like on my phone, just send it to myself so you can sort of start to restrict it that way. And on the permissions from the corporate side in the administration, there are a few little, you know, workarounds to it. But overall, I agree with you, yeah,

    it sounds leaky.

    That's what makes me nervous about using any of these chat bots or things like that. I'm just like, you know, I too easy for you to do it, and it's going to make me nervous either way, right? If, if I, if Google does an AI scan all that stuff and gives me a thing then, now, now, depending on the security of that, somebody could be asking that, questions, to that, and getting all sorts of information. And it's even, you know, they don't have to find anything out. They can say, you can ask consolidated, you know, a question to consolidate a certain amount of data. But I would rather get that from Google than from, you know, startup, I

    don't know, yeah, but by the same token, Thomson, Reuters, LexisNexis, all the big services are already doing this. They are then. What do you mean, tho? Thompson, Reuters started writing its its AI to add itself as a service to anything and have anything be added to it as a service more than a year ago. They see they put their stake in the ground that, you know, generative AI is going to be their play, regardless of what model or which company or anything like that. The AI aspect will will get them where they are trying to go in digital media, where they couldn't go without it, making that capability. It's like, Why did IBM by the weather channel for exactly the same reason, access to the data. Yeah,

    all right. Everybody gotta run. This is good, yeah, I gotta go to next week. Is electrical power as the governor for growth. Electrical power as the governor. Electrical power as the governor for growth.

    Okay, by the way, just as an aside, John didn't confirm the July 8 date yet, so we'll see. Okay.

    I know you'll know. Yeah, okay, got it. July 8. Cool. July 8. Do. I'll look at the, let me look at the I was out of the office all day, and I've been swamped, so

    I just randomly picked a date that was a Thursday that was giving a month if we were going to do a little, you know, promo thing or whatever. Definitely,

    I'm just confused because I think you mean August 8. And I'm now upset July is not a third. It's not a Thursday. Okay, then I

    have the date wrong. Okay, I'm tired. No

    worries. Let me look at it. I'll get back to you. I'm sure it'll I'm sure it'll be fine.

    Okay, I'll talk to take care. Have a good one, you guys.

    Wow. What a great conversation. There is so much going on in AI and the way in which we manage, govern and share our data and learn to do better jobs, automating, building more workflows and AI capabilities into our daily existence. I think we're only at the beginning of this, but I do know that we will need to be, and this was a good point for Joanne on closing, we will need to be very careful about who's who's getting it, who has access, who owns our our own data, our own information, and how they're using it all going to be critical pieces and things that we will continue to talk about on the 2030 discussions. Please feel free to join us. We meet every Thursday morning, and we cover this and many other topics, including our upcoming book club discussion about the two butt rule. So please read that book and join us. Can Find out more at 2030 dot cloud, I'll see you there. Thank you for listening to the cloud 2030 podcast. It is sponsored by RackN, where we are really working to build a community of people who are using and thinking about infrastructure differently, because that's what RackN does. We write software that helps put operators back in control of distributed infrastructure, really thinking about how things should be run and building software that makes that possible. If this is interesting to you, please try out the software. We would love to get your opinion and hear how you think this could transform infrastructure more broadly, or just keep enjoying the podcast and coming to the discussions and you know, laying out your thoughts and how you see the future unfolding. It's all part of building a better infrastructure operations community. Thank you.