The Great Data Interregnum: Morbid Symptoms of an Industry Between Eras

Speaker A: Hey, what's up? It's sho. And welcome to Freestyle Friday 4, October 30th. So earlier this year I posted an article called 2028 the Great Data Reckoning. And it was really a spoof on the Citrini research team's uh, publication of the 2028 global intelligence crisis. Um, both these articles came out in February 2026. And really what we laid out were more about the scenarios of uh, the impact of AI on uh, the economy. And last month I was chatting with uh, Tristan Handy, uh, from uh, co founders of dbt, and he called what's going on right now, uh, the interregnum. Uh, and I thought that was a perfect way of describing what we're going through. And it's actually um, Antonio, uh, Gramsci, Italian, um, philosopher from back in the uh, 1930s. And he has a very famous phrase, uh, which says the crisis consists precisely in the fact that the old is dying and the new cannot be born. In this interregnum, a great variety of morbid symptoms appear. This appeared in his prison notebooks. Uh, 1930. And I can see that the world is going through an interregnum right now. The old ways, the old order is dying. We're, we're not sure what comes next. And if we apply it to the data industry and maybe the technology more broadly, but we'll focus on the data industry since that's um, my area of focus. There's a lot happening right now. I would say that this quote again about the morbid symptoms, the interregnum. This perfectly captures where the data industry right now is at. So uh, the prior world was most encapsulated, um, most recently by the modern data stack. And uh, you know, the modern data stack as we understood it is coming apart. Right? So we've had, uh, fivetran and DBT are now one company. IBM just bought Confluent, SAP bought Dremio, Salesforce bought Informatica and a, ah, while back Tableau as well. Um, Snowflake and Databricks um, both bought Postgres companies. They're buying other types of companies as well. And um, basically every vendor, um, is suddenly discovering that its product um, might actually have been built for AI agents all along. Uh, I say this half jokingly and mostly true, but here's the thing. Um, the people working with data haven't suddenly stopped having the same problems. Uh, pipelines still break, um, numbers are debated and probably not reconciled. The business still can't agree on what a customer is. These are old tropes that we've dealt with as an industry for decades. Um, yet we're being sold a new world, uh, while we're still operating in the old one, the interregnum. And so today I want to talk about, um, you know, what might have gone away, at least for now, uh, what hasn't been born yet and what happens to everyone, you know, whether you're a vendor or a practitioner or a leader caught in between these two worlds right now. So I think, you know, a line I'll keep repeating through this, uh, podcasts is vendors are selling some version of the future which we haven't uh, yet determined. Um, but the practitioners are maintaining the present and the business is still waiting for results and often living in yesterday's world. So if we look at what actually ended with the modern data stack in particular, um, it wasn't a particular architecture that ended up, uh, I think what ended was the idea that customers could assemble a platform from independently purchased and specialized components. So if you'll recall, the modern data stack was very much encapsulated by this idea that we could um, have a modular data stack. Ah, very easy to assemble. Um, supposedly integration was actually quite hard in some cases. Um, but all this is available, uh, through the swipe of a credit card. And you know, in the modern uh, data stack heyday, which I think Tristan maybe says it's about three years ahead it's run, I think that's about accurate. You know, you saw, especially in the, um, late maybe 2019 to 2022, an explosion of companies, right? And you saw a massive explosion of companies after Covid and just uh, the zero interest rates, meaning anyone could get money, which meant there's a lot of money out there, which meant everyone basically built companies. Um, and you come to realize that those are actually just features, but that's a different article argument. Um, but this arrangement did create substantial innovation, but also I would say substantial integration costs. So again, we're talking about, um, the latest, uh, acquisition wave. Fast forward to today. Interest rates have risen, they keep rising and there's a lot of consolidation happening. So I'd say the modern data stack was very much, uh, emblematic of the unbundling phase. And now we're in the bundling phase. Right? So as I talk about, in the pendulums of data, things go through consolidation and um, deconsolidation and then back and forth. Right? And so we're back in consolidation. But there's a punchline I'll get to in a bit. Again, I mentioned five trend, uh, DBT so ingestion and transformation combine into a single infrastructure vendor. Right. So it closed around June of 2026. IBM and Confluent. So event streaming becomes part of IBM's enterprise AI and integration portfolio. This closed around March, um, 17th, 2026. I was actually at one of their events, uh, when that happened. Um, finally closed and it was a um, interesting day. I'll just leave it at that. SAP, uh, and Dremio. So Open Lake House Technology moves under a massive enterprise application vendor which is SAP Salesforce and Informatica. So data integration and governance become strategic assets for a major enterprise application. Governance. And then databricks buying Neon and so for buying crunchy, ah, data and so on and so forth. Right. So lots of consolidation still happening. OpenAI also buying uh, vendors as well. Simon Anthropic. And so you know the, the unbundling area era is giving way to a rebundling era, which again happens a lot. But you know, I think it's worth asking is this driven by AI in particular or is this driven by market maturity, maybe slowing growth or simply the economic advantages of consolidation? So maybe if we look at fivetran and dbt, that merger brought uh, to defining companies with modern data stack together. Uh, and so I think the original combination was um, described as approaching what, 600 million in ARR, uh, which um, more par to them, uh, I think it seems they're also developing um, a competing platform against the snowflakes and databricks of the world as well, which I think is a necessary thing. Uh, snowflakes and databricks are also encroaching on uh, these five trend DBT's territory. So that's how the uh, competitive dynamics go, I guess. It is what it is. Um, but I would say the capabilities of the modern data stack aren't changing. Right. So um, they're consolidating but capabilities aren't disappearing. The companies and uh, commercial boundaries around those capabilities are changing. Right. So again, consolidation for now. This brings me to um, the morbid symptoms that I see in vendor land right now. I spend a lot of time talking to vendors. Um, maybe there's four categories we can talk about here, uh, in terms of these morbid symptoms. Symptom one is positioning convergence. Every vendor, if you're talking about the modern data stack era, if you go to Matt Turk's um, MAD data landscape for example, that were just a lot of um, areas that vendors operated in, they kind of stuck to their lane, so to speak. You had observability, maybe had catalogs maybe had this, maybe had that. Everyone is now selling an AI ready data foundation of some sort, right? Or an agentic platform or an enterprise context layer, often all of the above. And the problem is, is that differentiation becomes harder to understand when everyone uses essentially the same language. So when everyone claims to be the foundation for AI, the word foundation stops telling us anything. And the same goes for context, the same goes for agentic platforms. All these words in my opinion, are utterly meaningless at this point, um, because everyone is competing again, often bundling these offerings, uh, together. And the result is it creates a lot of product confusion for the end buyer where they're not quite sure who's good at what, because everyone is literally using the same words to describe what they do. Symptom number two, I would say the buyer, the buyer is changing. So vendors increasingly want access to AI transformation budgets, not just engineering budgets. So what you're seeing is vendors are often going around data teams and going straight to executives, um, maybe chief marketing officers and so forth. And so this creates a tension between the person who purchases the technology and the people, the engineering teams, who need to make it work. So vendors are also often pitched that, well, we're just selling, uh, data teams in a box, we're selling engineering teams in a box. Um, but what often happens in reality is the engineering teams are the ones who usually have to implement this stuff. And it's not a good look when you tell the data engineering team or software engineering team or whatever that hey, um, love you guys but we want to get rid of you. You're going to implement this tool for us. Have a nice day. So we've seen parallels before, right, with self service, bi, no code, um, RPA and so forth. And each of these promised to reduce dependence on specialized technical teams. Um, each made certain activities easier for sure, uh, without eliminating the underlying need for expertise. I would say the difference today is that AI really can automate a meaningful portion of technical production work. But again, um, people who have the context, the people who have the understanding of the, of the systems and how you need to migrate those uh, processes and practices to keep your business running. That those, that is the engineering team, right? And so, uh, again, but it is interesting that the buyer is changing. Um, you know, it's the wet dream of every executive to basically just get rid of their engineering team. I don't think that that's a, uh, secret to anybody. Uh, you know, you want to control costs, you want to get rid of these pesky people that, um, you know, uh, you're not quite sure what they do, but you'll find out when, when uh, they stop working and your system starts uh, collapsing. Um, which it does happen these days. You know, you're hearing about people laying off their engineering team, software engineering, data teams, whatever, but then they um, then they come to find out that oh, those people are actually important. But they, you know, if you don't take the time to understand the work, um, I guess what you, what you uh, put in is what you're going to get out. Let's move to symptom number three. Right? So the agent is actually a customer. Now agents are customers. So vendors are beginning to design around agents as major consumers and producers of data. So you know the funny thing is these agents, they don't go to your, the user conferences, they don't complain about your UI because they don't really need to use it. Uh, they don't care about your community, Slack or the drama in there. Um, they can run up enormous compute bills. Uh, but the thing is the agents will always be there. They work 24 7, they don't take breaks, they don't complain. Typically maybe they do. Um, but if agents become major consumers of data, what happens to access control, observability, cost management and semantics and so forth, right? And so who's responsible when an agent makes an incorrect decision using technically valid data, for example? Um, I think this is an open ended question. Right now you'll notice that vendors are actually punting the responsibility of agent behavior onto uh, the customers who are footing the bill for the agents. And so this will be interesting. As we rely on agents more and more. I run them in my business every day. But I also know what I want. I'm also a solo company so it doesn't really matter if I, if the agent gets it wrong. So I can just course correct automatically. But when you're dealing with an organization of you know, thousands or tens of thousands or hundreds of thousands of people, uh, things are much more complicated, um, interesting, more, uh, I would say, uh, just complex, right? In that sense that it's, it's an almost unsolvable problem in some cases. But that's where we are. Um, I would say that number four is the disappearance of neutrality. And so in the modern dataset era you had a uh, plethora of independent uh, products, uh, and vendors. And now these are all becoming part of larger vendors. So customers need to reevaluate, um, roadmap independence for example and commercial leverage and interoperability. So um, you know, if you're stuck in an ecosystem, you know, say SAP, Snowflake, Salesforce, databricks, you know, one of the big clouds or whatever. Uh, is this a good thing or not? Right. Um, so I think the question is whether, uh, openness, um, persists in terms of your product investment and commercial incentives over time. And I would say, I wouldn't assume that every acquisition destroys neutrality, but I would definitely make, um, this, um, a big piece of how you evaluate your technology, um, purchases and your infrastructure going, uh, forward. But let's move to practitioners. I think these are the people caught between the two eras. The era of today and the era of the future. The job market right now is sending, I would say, very conflicting signals. So Data Camp, for example, they had some analysis. Roughly 2 million job postings. Um, and it seems like data and AI roles are up 80 point, uh, 2% year over year, with AI engineering roles up, uh, 255%. Um, but at the same time, uh, conventional data and analytics job postings are down according to, Indeed, um, I would say these aren't necessarily contradictory numbers, but they do, um, point, ah, to, I would say some interesting trends. Right. So does this prove that traditional data jobs are simply being renamed A.I. jobs? Uh, I don't know. That would require tracking. Um, I would say that the kind of tasks that we do, the job transitions and not just comparing job titles on the surface. So there's a few holes, these comparisons, but these are conflicting, uh, signals for sure. I think the more interesting question is whether the composition of data work is changing faster than maybe our job, um, taxonomies and job descriptions, um, can accommodate. I would distinguish the changing economics of data work into three areas. Um, there's sort of boilerplate work that we did for a long time that's writing SQL, writing scaffolding, writing basic documentation. This is increasingly becoming automated. So it's relatively easy to generate, um, boilerplate SQL and test it and so forth. Um, things like modeling semantics, evaluation systems, uh, design. This is potentially more valuable, but then it requires a lot of context, interpretation and judgment. And so the question is, do the people who are doing the boilerplate work, do they have sufficient context, interpretation and judgment to become valuable in an era where modeling, um, semantics and so forth will probably become more of the work. Right. But this also moves up the food stack to, um, the food chain, to things like accountability and stakeholder alignment and risk management. Um, I think this always remains essential and this requires more organizational authority. Right. So if you were in a data team where you were kind of operating at a Lower level as technology goes, you'll move up the, you know, the, to higher levels of abstraction. But this also means that you need to expand the purview of what you were working on. It's no longer just about the technical aspects, but it's about the business context and interpreting it and judging it, along with having the organizational authority, um, to operate with the business and make decisions. Um, because you're working with agents right now and you're working with people. So you've got to straddle both these worlds. Um, I would say that the important nuances that AI can assist with all three of these boilerplate sort of the um, contextual work and then the organizational authority work with. And none of these are categorical, categorically immune to automation either. Right. So I think what changes is the relative cost of producing work versus deciding what should be produced and verifying that that thing that's produced is correct. So this places a double burden on practitioners. Right? Practitioners must continue operating the existing infrastructure while simultaneously making it suitable for agents. And say this is going to be the work of the next few years is again operating existing infrastructure while simultaneously making it useful for agents. This isn't something I'm convinced that you can just bring the data team in a box and just get this done. I don't think that this is, this is how it's going to work. Um, for anything of sufficient complexity there's just, there's too much nuance and hidden context, interpretation and judgment and you need the authority to do this. I think this means reviewing um, more generated code which is going to be kind of paradoxically, um, hard because uh, as I wrote about in the post Literate Engineer essay, if human reading is going down while things that are written is going up, so that'll be interesting to see that dynamic. I think that reviewing um, more code is going to be done in an automated fashion with evals and so forth, but also managing non human identities, um, and entities, agents and evaluating agent outputs and monitoring cost and making business definitions and processes explicit to the agents. This means you need to understand how business processes work is why I wrote about it. In Mixed Model Arts there's a whole chapter on seeing the business. And this is I would say, one of the most underrated skills that is going to be very suddenly very um, necessary. So a question for you is how much of your work, you know, during a week is spent writing code versus reviewing code that somebody or something else wrote? Um, so. But this introduces, I would say, an uncomfortable identity crisis for people you Know if you like to write code. I still do. But what happens when that part of your job you most enjoy is no longer the part that the business values the most? Right. So I know a lot of people that love writing code, love building pipelines, love optimizing infrastructure, and there's nothing wrong with that. I think this is just, it's a lot of fun. Um, but um, this may not be that the business has been paying you for the entire time. Right? So it's a deep implementation. Expertise certainly remains valuable. But we as a profession cannot assume that the market will continue paying the same premium we've been getting for many years for every activity simply because it once required specialized technical knowledge. The um, alternative isn't necessarily becoming a product manager, a business analyst or people person or something, but it's more about becoming a complete engineer. And to me this is somebody who understands systems data, meaning failure modes and business objectives. You're kind of the complete package, the uh, mixed model artist, so to speak. Um, and it's gonna be interesting, right? I think this is the uncomfortable reality a lot of uh, practitioners are operating in and trying to figure out right now. Thankfully, you know, AI is not the interesting thing anymore. I think that it's now how, how are we using AI to um, and how are we opening our minds to the possibilities and the likelihood of just operating in a new future and what does that new future look like? And again to me it's about if you're a practitioner, if you're an engineer, it becomes becoming a more complete engineer. Again, understand systems data, meaning failure modes, business objectives and so forth. Which brings me to segment four of this whole thing. Right. So how do you operate within this interregnum that we're in? I would definitely uh, emphasize and this for practitioners in particular, but optionality over predictions, we have a hard time predicting the future, especially with AI. I think AI is a very polarizing topic for a lot of people for a lot of reasons. Um, and so predicting the future I would say is very difficult, especially now nobody has a good track record of it. I wouldn't say people that do, um, they're totally full of shit. Um, but optionality is the thing you can control. So avoid making irreversible architecture decisions based on vendor marketing. That's a big one. In Fundamentals of Data Engineering, Matt Housley and I call this, um, one way and two way doors. You always want to have the ability for two way doors where if you go into a door you can get out versus having the door Locked behind you, which would be a one way door. And this is what Jeff Bezos described as one way and two way doors, meaning reversible or irreversible decision making. You want um, to avoid making irreversible architectural decisions, especially that based on vendor marketing, which is obviously highly confusing at this point. And I think there's a lot of uh, charlatans out there, um, telling you that their product will make your uh, life is oh so great. Um, and this means you got to pay attention to the vendor landscape. So understand we talked about some acquisitions earlier, understand the acquisition risk and the switching costs. I would prefer portable interfaces where um, you can preserve flexibility. I don't think this necessarily equates to open formats 100% of the time. Um, but uh, this comes with trade offs as well. But I would say, um, I wouldn't say that keeping all business logic outside of proprietary tools is economically sensible either. I would say the relevant question is whether the value of portability exceeds the cost of maintaining it and whatever comes next. Things are moving at breakneck speed. I would say don't get tied down to a vendor too tightly, don't get tied down to a solution too tightly, but have optionality. For data leaders, the practical recommendations are uh, pretty straightforward. I would say budget for evaluation and review, not merely um, production, uh, uh, I would say make ownership and authority explicit uh, with your teams and with the people in your organization. Um, all too often everyone kind of brings their own AI, so to speak, right? And it's a very confusing mess. Um, the AI sprawl is very real right now, Agent Sprawl. And it makes microservices look like a dang cakewalk at this point. Um, I would say test Agentix systems against real organizational ambiguity rather than polished demos that all the vendors have. How does it work on your use case? Your organization's different than everyone else's. How is this going to work with you? I would say for vendors the challenge is maintaining credibility of practitioners even as sales conversations move towards executives and AI budgets. Um, the danger right now I see with a lot of vendors is they almost have a condescending tone towards the practitioners. And I think this is a big mistake. It's a big, big mistake. These practitioners might get promoted, for example. They might um, you know, be involved in your implementation. And I would say you need to treat them with respect. Uh, and so, uh, you can sell around the data team, uh, but you can't necessarily deploy around the data team or engineering teams. So you know what, So I guess Kind of bring it to a close. Like what ends this interregnum? I don't know. Um, I don't know that any of us knows because this new world hasn't appeared. And by the time the new world appears, uh, it will already have appeared and we'll sort of know it in retrospect. That's how history goes. You don't know what happened until you look backwards. Um, I don't think that things like a shared agent data protocol is a sole condition for the interregnum ending. I hear this from some people. I think technical standardization is one possible signal. But um, market structures typically stabilize around dominant buyers, purchasing models and operating practices. Right. So I probably watch for a few things. Um, whether agent to data interfaces and semantic contracts converge across competing platforms. Right. There's Aussie happening right now, the semantic um, interchange and we'll see if that actually um, becomes a standard or if it's just uh, you know, sort um, of semantic washing by the industry. I would look at whether enterprises develop repeatable ways to evaluate and govern agent, uh, driven data workloads. That's a big one. I would also look at whether data infrastructure spending shifts materially towards uh, consolidated platforms. That's going to be a big, big one, right? Or is it? In some cases what I'm seeing right now is people are actually ripping out um, you know, the big platforms and just using iceberg. That's also happening. Um, you know, so they said the consolidated platforms are at risk. Uh, in fact I was having this conversation with Tristan Handy yesterday where we're talking about this. You could be looking at right now. You're looking at you know, obviously bundling but you, you're looking at the early stages of, I'd say the next unbundling. And uh, he mentioned this where you're seeing just uh, the fact that people can use AI to generate apps, um, and point solutions for themselves means you're actually looking at the new era of unbundling. This is something I um, talked about another podcast too where you're sort of in this era of the great Unbundling once again. Um, but we just haven't called it that yet because you don't really see it. Right? It's not like you see the uh, there's not a bad mad data landscape for AI generated applications within an organization. But I was talking to a team the other day, they're ripping out tableau, uh, and replacing it with react apps. And that's the route they're going, right? It's a very Big company, uh, and come to find out that uh, within five minutes I talked to somebody else, they're doing exactly the same thing, ripping out tableau, um, and replacing it with their own uh, internal um, self generating AIs. And so this means that consolidated platforms might uh, be at risk as well from this kind of thing. Just move an iceberg and making it BI apps. I'd say the last thing I would look at is whether job descriptions and organizational structures settle around new responsibilities. And I think we're already starting to see that the number of companies requiring AI as part of the job description, uh, for anybody working as an engineer or in data or analytics, so forth, that's just table stakes at this point. I would say that, um, but what are the new responsibilities? I would say look at the job descriptions and maybe get an idea. But that's currently happening. And so I would say that maybe there's some things we should look at maybe by the end of 2027. The first thing is what's the distinction between analytics engineering and AI oriented data engineering? Uh, does this become less meaningful in organizations deploying agents or does it become more, what's the responsibility, these uh, roles and so forth. And I would say the second one is really what uh, happens with the semantic and context layers. This has been the big buzz of this year. But at the end of 2027, what's still commercially contested, um, is it clear where that whole thing has gone or is it um, still ambiguous? Do we have standardization or don't we? Who's adopted it, who hasn't? And so bringing it back, I'd say that the modern data stack, again it was never, it was for data engineering, right? It was never the purpose of data engineering, but it was one particular way of organizing the technology that the vendors, um, and the labor required to make um, data useful, right, that was the latest manifestation of it. But obviously things are changing, right? Organizations are changing, tools are becoming more integrated, um, through acquisitions and so forth. Interfaces are becoming more automated, I would say in a lot of ways disappearing. Um, and the economics of producing technical artifacts are just changing right now. But none of this makes the fundamental problems disappear. We still have to understand what the business wants and what the meaning is within the business of the things we're trying to work out. Uh, we still have to decide what's correct, we still have to deal with uncertainty, um, conflicting requirements and failing systems and so forth. Um, maybe the most important thing about this interregnum is that it's forcing us to separate the work that matters from the tools we've been using to perform our work. As I've always said, that's why I wrote a book on fundamentals. That's why I wrote a book on mixed model arts. These are technology agnostic. I don't think the tools are as important as we think they are. What's really important is always been the work like I just described, right? The things that, um, you know, the fundamental problems that aren't disappearing no matter what technology we throw at it. And so, you know, we'll see what happens, uh, you know, during this interregnum. But the thing I would urge you to again is just maintain optionality, um, you know, improve your skills, widen your skills, become, uh, more complete as a practitioner and a leader. If you're a vendor, um, you know, definitely find your place in the universe. Invent the future. But, but please stop with a lot of the horseshit. Um, you know, uh, I would say just, uh, you know, kind of the sea of sameness I see with every vendor where you're just doing the same thing everyone else is doing. You're not going to survive as a vendor, right. If you're doing exactly what every other vendor is doing, using exactly the same terminology, like what differentiates you from everybody else in the market? There's nothing. Right? Especially when people can just basically take your product and your marketing copy, put that in the cloud and build a pretty good version of that. So I'd say this is a very interesting time right now. Um, the things that, you know, that we, um, thought were safe or now unsafe, and that's just how the world is. If you kind of zoom out, that's how the entire world is right now. Geopolitically, economically, socially and so forth. I think the entire world is going through the great interregnum. And, uh, this is part of the cycles of history, uh, that we've gone through, um, since the dawn of our civilization and, and uh, you know, our industry is not immune from that. But, you know, I just wanted to address a lot of the concerns I see out there that I'm hearing, you know, from practitioners, leaders and vendors across the board. There's a sense of not knowing what the future holds, trying to hold on for dear life, pivoting. And that's the reality. Anyway, that's the rant for today. Kind of a long rant. He's 30 minutes. Um, anyway, uh, I think encapsulates a lot of what I've been seeing, uh, you know, on my travels this fall. And you know, I'm sure I'll be talking more about this. It's top of mind of everybody or these sorts of questions. And so I just wanted to at least give a perspective. As of today, maybe next week I'm going to do a, um, I'm thinking about doing a kind of the future of data engineering. That question a lot. I have a lot of opinions. It sort of dovetails off of this, but I think there's some nuance as well. Um, anyway, next, uh, week in San Francisco for Data Outposts. That's uh, Mother Ducks event. It's going to be great. Lots of great speakers. I would say show up. I got some discount codes. Check out the newsletter, uh, for discount codes. It's a screaming deal. If you like great deals, come check it out. And after that off to London for Connected Data London, giving a, um, keynote there on the 12th about the convergence of the knowledge and data worlds, uh, as told through various, uh, south park episodes. It's going to be pretty fun. Then Defore Data Conference in Paris, which is an awesome, awesome conference, and Austin for MLOps summit. Then I believe I'm in New York City the next week, um, after Thanksgiving for uh, AI Dev with Deep Learning AI. So I'll be emceeing that. I uh, think um, Andrew Ng is going to be interviewing Yann Lecun as well during that um, show. Uh, so definitely come check it out. It's gonna be awesome. And then, yeah, other than that, just uh, finishing up this course for Mixed Model Arts. Expect that sometime around mid December I think. So I'll be able to drop it just due to travel and some other things, but that's happening. And also starting to work on mixed, um, Model Arts Book two. And yeah, I just have a lot of travel coming up in 2027 as well, so. Busy year, but that's how it is. And obviously we'll be doing more events with the practical, um, Data Events, a new company I have, uh, we did cancel the summit. We're postponing that till February 2028. But I think that's a good thing. Uh, if you're looking forward to attending that, um, it'll be awesome. But uh, yeah, in the meantime, a lot going on. Have a great weekend. Take care. Bye. Bye.