Heard you talk about flow state. It's an area where a lot of our gifts emerge and unfortunately an area where we don't get that many opportunities, especially in the modern world. Do you feel like AI has been able to, for you, take you out of reactive mode? I think smartphones have made it worse, not better. Brings us to AI for executive search. What are humans doing wrong today when they go and make that hire? It's the way you use existing tool sets. If you're opening the funnel to get even more candidates using AI, how can AI help make the decision better? Adding goals, think, is a super interesting idea. We don't see that that often. A lot of times, we will see what you're gonna do, roles, responsibilities. If we're in two thousand and thirty, what is true of the world? One of our contrarian theses is that domain expertise and operating experience will become more important, not less. And the reason is Alright. Today we are here with Alex Bates, founder and CEO of Hello Sky. How you doing? Doing great. Fantastic to be with you. Great. Well, look, I wanna get in, right into it. I think like anybody that knows you knows deep background in AI, wrote many peer reviewed papers, have been founder of MTEL, which was trying to predict or actually, it was predicting when oil rigs, power plants were gonna fail, and now all the way to executive search, AI for executive search. Interested because I've been listening to a bunch of the the content you put out and and other podcasts you've done, and I find find that, like, it seems like a fascination with brain science has been kind of at the core of everything you've done. Absolutely. I think when I first of all got into artificial neural networks, you know, way back in the day, middle school, I think when I started studying how neural networks work, I was just mind boggled about this thing about automated self correction, error correction, learning how these neural networks could learn to compute useful functions in the real world without a teacher, just with a feedback signal. And so it led me down this fascination actually with the human brain and neuroscience and how our amazing capabilities work, as well as how AI works and can augment and supplement that. So it's definitely been a common theme and, you know, something I've I've been very passionate about. Heard you talk about flow state. Flow state, it's it's been profiled. There's a number of characteristics to it. It's it's a characteristic of you a feeling of timelessness where you are obviously in the zone. Sometimes people call it there's different versions of it. Some are more active, like runner's high. Some are just you might be taking a walk in nature, and you're sort of disconnected from the physical world around you, and you're primed for these, like, epiphany moments and eureka moments and just completely in the zone. Whatever you're thinking about could be artistic, could be more scientific, and it's an area where a lot of our gifts emerge and, unfortunately, an area where we don't get that many opportunities, especially in the modern world. Can we force it, or is it something that just kinda has to happen by, you know, random forces like whether you're in a certain mind state already, or can you get into it? It seems like it has to emerge. That's the hard part. There's been tons of studies on it. We obviously would all love all love to just flip a switch and be in flow state. But it seems like this complicated, thing involving our default mode network of our brain and other circuits that it just is an emergent state. And that's part of one of the hard parts. Like, we we know it's underlies a lot of our gifts and yet it's hard to manifest. So I've heard a well, I've read a lot in in books about, shower principle, things like where, hey, we're getting our best ideas when we are doing something that's a habit. Right? So I've, if I remember correctly, if you think about the brain and the outer layers of the brain, that kind of, you know, spaghetti looking things, and then you've got the basal ganglia, which is like a golf ball sized, you know, region in the center. And what I read was, what if you could look at the electricity of the brain firing, when you're thinking about things actively, that outside area is going nuts. But when you have something that's formed as a habit, all of that activity moves into the basal ganglia, which frees up the other areas. I remember them liking it to, taking a shower. You just know, you just naturally are doing this. You've done that over and over and over, and that's why you maybe get your best ideas in there. Same thing of like you're backing out of a curvy driveway with a a you know, it's got a flower here. You don't wanna run over that. It got a curb. It's got a basketball hoop. First couple times you do it, you are really paying attention. You can't think of anything else. Twentieth, thirtieth, fortieth time you're doing it, you're thinking about work, you're thinking about all the other things. Being able to move those habits into the basal ganglia in order to, like, get that area freed up, I'm wondering, like, can AI help us there? Can AI do the things that instead of having to make something a habit, can we make everything so easy that the decisions don't have to be made each time by our brain to enable us to be in somewhat of a flow state or shower principle area where we can think more creatively? I think it absolutely could. I mean, if you think about another term that's used as undirected mind wandering. If you're in some task positive things, something you haven't done before, your attention is completely focused on what you're doing. It has to be something where it's an ingrained habit, like you say, basal ganglia, it's a nature walk that you do all the time, stepping out of the shower. So it frees your mind into this undirected mind wandering state where you can now go into these deeper avenues that you normally don't have a chance to explore. And we have so many, like, things coming at us every day. We're at work, you've got at any point, I've got many unread emails and many more unread Slacks. So there's constant things coming at me that are forcing me to be in reactive mode. Do you feel like AI has been able to, for you, take you out of reactive mode? I think smartphones have made it worse, not better. So I mean, I think the challenge is it seems great to be able to respond in real time to any inbound ping from anyone in the world, but then the reality is that's zapping us. Because, like, that will knock you straight out of flow state. So I think there's an opportunity for it to figure out which things are actually an emergency I need to zap out of FlowState and respond to, and which things can it give maybe an auto reply and queue up information and tap me in later. I don't think that's been solved at all at scale. You know, I've experimented with assistance, virtual assistance, the whole gamut, and still when what is something I urgently need to reply to now, and what is something you can you can fend off? I haven't seen tech solve that yet, and yet that's come probably one of the biggest unlocks. We're also at Cadre. We're in an office. Most of our team members are are local in in San Diego, and, you know, you forget. I've I mean, I've been in been working remote for twelve years before this. Right? Even when I was CMO of private equity firms and traveling all around, like, it was mostly remote. And now I'm in an office, and I find that that's another input. I can't go get water without a whole bunch of needs popping up. It's either someone asking you for something or you're seeing something that reminds you of something you have to go do. And so getting into this zone of having the capacity to be able to think critically and, you know, get creative is, I think, getting harder. I think it is. And yet, you know, the serendipitous water cooler conversations can be extremely valuable. When you're hitting someone up with the twelfth message on Slack, it's unlikely to happen. And so what is the optimal middle ground where you get some face time, you also get some away time and time to, you know, lock in and go in the zone? I think that's where we need to go with the future of work. I think there's been discussion about it post COVID. What's the hybrid model? I think largely unsolved, and yet, I think that'll be super important in the next chapter. So AI can do as much as it possibly can to free up capacity for executives, but at some point, you need a general or two. You need those people under you that are gonna go, you know, tackle those those hard problems for you, ideally using AI to do it as well, but you always, I think, at least anytime in the near future, we are still going to need to hire executives. Brings us to AI for executive search. What are humans doing wrong today when they go and make that hire? Yeah. I mean, we've worked with a lot of different types of search firms from, you know, maybe staff level to the highest end of elite retained executive search, you know, multi six figure, sometimes seven figure commissions. And so we've seen the gamut. Some clear non best practices have emerged. I think some of the biggest firms, you've already assessed five hundred thousand plus people. You know them well. You've talked to them. There's obviously gonna be a slight bias to recycle people you've placed. Not only that you know them, you have a direct relationship with them. So that's one issue we can see. Another one is the way you use existing tool sets. There's a whole slate of invisible people. So you might, people that brag a lot in their bio are gonna get surfaced more when you're doing keyword search. So we try to do a model where we can literally construct your bio from every company you've been at, every growth, every outcome. We see some of the best people actually having no bio, and yet that doesn't impact our ability to find elite talent. So I think not over relying on, you know, bios and self reported things and being able to actually construct what they've accomplished, and then still look at the bio because that's useful for behavioral profiling and things like that. So I think there's a lot of opportunity. What we've seen is even at the elite end, there's a lot of people that get missed and overlooked, and everyone is struggling. It's like you're in six month of a search, you're getting to desperation mode. That's not good on either side. Yeah. Did I hear you right there that okay. So you've got what they say about what they've done, but we also know that they've been, at least from self report data on LinkedIn, they've been on this they've been at this company from this time period. Are you saying that you're able to actually go look at what that company was able to do if there's publicly report reported data, like how it grew, it shrank, it whatever, to be able to determine what their impact was there? Yeah. And we go deeper beyond even publicly reported. So companies like Oracle might acquire a company and eviscerate every trace of their online existence. They're gonna delete the website, delete the company LinkedIn page. All of a sudden, there's all these executives that went through a massive impressive exit. It's invisible on LinkedIn. And maybe they put it in the bio, maybe they don't. So we we have to construct point in time like the way back machine for companies. You know, what what was the scale story, the growth story, the outcome, even if it's completely eviscerated from the Internet, and give credit to all those peoples that went through that transition. So that's that's part of what we do because some of the super impressive stories are not clearly visible on on LinkedIn today. So, okay. So I've been at companies where we have engaged with firms that are where I remember it was like a hundred thousand dollar retainer to search, and I and it wasn't contingent on them finding someone. Was like, no, we're gonna pay them a hundred thousand dollars, go find this executive for us, which to me at the time felt wild. But, honestly, the quality of candidates was amazing that we were getting. They were liking it. They were saying the reason is it's because they've got this network, Right? It's that it's the people at the agency that have the network of of these high level executives that those executives aren't saying they're looking. They're just saying, I know this person. I'll I'll take a call. And so I wonder, like, where does AI fit into that equation of of you're always I would imagine you still need these networks in order for someone to be able to reach out. But how is it enabling a search firm to find those great candidates? Yeah. I mean, think, they'll do a lot of private assessments, and part of their sort of value is they know we've interviewed all these people, we know approximately the revenue story and the exit story. So what we do is we reconstruct those, and those notes can become out of date, and sometimes there's some bias in that too. So one is constructing the full slate of potentially qualified candidates and using AI to help stack rank that. On the outreach side, you know, premier firms at the top, like the columnist track, the big five exec search firms, they're known brands. So, yes, when they reach out, oh, I I got a call from someone, I might take that call. But on top of that, networks are important. We actually have another technology that helps trace third second first second plus degrees strong ties. So like, not just one of your eight thousand LinkedIn connections, but you were on a board with someone you're on executive leadership team or two degrees away that went through that. So we score those relationships so we can help firms harness their extended network that might not be in their CRM. And it's you know, when you get to the thousands, it is hard combinatorial even for humans to keep track of, so it can help us manage that. So both, you know, not over relying on your logo for outreach, but also being able to trace that full full set of ties. Okay. So it's helping us get more candidates in the pipeline. Now comes the the decision making process. Who should we hire? Right? Who's gonna be the best fit for this role? One thing that I find, and I I feel like I've been a a culprit of this, I've done this wrong in the past, is when I write a job description, a lot of times, I'm like, I don't know. Like, I know the title of the person I want. Have a general idea of what I need out of them, but, like, I've gone pretty boilerplate with the job description, when really, would imagine, like, that should be the most important thing that I would do. So but I just don't find that job descriptions help as much as they could. I've moved to a model of I'm gonna write out the first couple goals that I would want this person to achieve. And I don't look at it as I'm gonna try to figure out how they're gonna solve that problem. I put down of here are the things that are true if you come in and and just nail it in the first thirty days. If you nail in the first thirty days, these things are true. I can look at these and say, yes. I agree with all the that stuff. You are the one that's gonna figure out how to get there. I've been working on that model to give to candidates to say, is this what you can do? Like, can you accomplish this? And I guess I wonder, like, if you're opening the funnel to get even more candidates using AI, how can AI help make the decision better for the hiring manager on who's gonna be the best fit for the role? Yeah. That's fascinating. I mean, adding goals, I think, is a super interesting idea. We don't see that that often. A lot of times we will see what you're gonna do, roles, responsibilities, and then qualifications, where they think those are gonna contribute to your ability to implement those objectives. I think what one way AI can certainly help is when you look at structured qualifications, we need, you know, five hundred million to a billion revenue scale. Maybe we need these outcomes. Maybe it's private equity backed or venture backed or public. So it can certainly help manifest that and do it with correct point in time. When it when worst practice we'll see is we'll just get current match companies because that's all we've ever been able to find, and now you really whittle down your candidate pool, and you just well, this is what we got, so now let's loosen criteria. We can get the full slate of current and past match where they scaled through that window and maybe had an outcome, and company might be gone. But yet it's an amazing, you know, set of operators you don't wanna miss. And where we're careful to not over kind of rely on AI is when you get to behavioral, team fit, culture fit, and just like that sixth sense about, I know the operators at this company and the personality of the CEO and who's gonna fit well, that's where the human element is still supreme. And yet, when people are canvassing out to a lot of the wrong targets and phone screening, hey, did you scale past five hundred million? Now you run out of time. You can't go as deep. So I think the key unlock, if you target better with AI can help, then it unlocks the ability to go much deeper, I think, on the human personality profiling fit, culture fit, team fit, you know, those kinds of areas. At the large companies I've been with, we've always had a really solid HR team that knows whether it's a DISC assessment or some kind of culture survey. They know the outcomes. They're able to send someone through that and say, ah, they're high on this, low on this. That means they're gonna gel well with this. And, you know, I find that at the smaller or mid sized companies, you often don't have that expert that can analyze those assessments. I wonder if AI can help with that as well. Meaning like, if it knows what the assessments are of the leader that that person's gonna report to, the team that they're gonna manage, I wonder if it could potentially, you know, call out things that could go right or wrong. A hundred percent. I mean, psych you know, these profile tools are super interesting. You got DISC. You got Ocean. You have Hogan, which is a more business centric one that's pretty popular. What you find we always ask search firms and VC private equity clients, do you do this? And what they'll say is we can't ask a hundred people, hey, subject yourself to a four hour, you know, psychology survey, and maybe you'll get hired. So what happens is maybe the top five candidates might opt in, they might not. If you have a high profile person, they're gonna be like, look, I got five offers. I'm not sitting down for four hours and doing your battery of questions. Like, do I like red or blue better? You know? So I think it's an interesting opportunity. I think AI can help us maybe pre assess, which is something we look at. And then and then you can really try to drill down on the edge cases and get a get a feel for it. But I think building up those datasets, and we talked to one private equity firm where they have every executive, because they have the leverage. Like, they're the investor. They can sort of force whether it's an exec search firm or an in house search. You're gonna do the test, and they have a pretty good data set on that. But, it it they have some contrarian perspectives, know, like the score is not as important, but we look for these criteria. So I think it's an interesting realm, and yet, if everyone would do it, it would certainly be incredibly Yeah. Like, you know, think about LinkedIn, right? It's this, if you go back before it came about and be like, Hey, we're gonna create this website and you're just gonna go and put all your information in, instead of just a resume, we're gonna make you do it, resume in another place. And I in theory, you're like, why? Why would I go do that? But it has become so ingrained in hiring and and networking that people wanna do it, and they are adding more to their bio on LinkedIn as things happen. If there was only a place, you know, and maybe there's an idea for someone smarter than me to come up with, but, like, if there was only a place where everybody could get in and also put in, whether it's, you know maybe it doesn't have to be a four hour assessment, but, like, something that'll that would allow a company to know if that would be a good fit based on who they'd be reporting to and managing. Because I think as much as people want jobs, when we're at this executive search level, they're most of the time, it's not like a desperate need for a job. They wanna go somewhere that they're gonna be able to succeed and fit. And if there was any way that AI can help make that match better, then I think everybody would be open to it. A hundred percent. Yeah. And I think they're, like, if you look at the arc of someone's career, there's obvious inferences about their degree of risk taking and certain things that you can pre assess, and then not force them to take four hundred questions. Like, let's really drill drill down the finer parts. I think that's an opportunity. And then, of course, we know that, you know, the the search engine companies probably, they always joke, know more about us than we know about ourselves. Not really accessible except to consumer advertising. You know, could we tap into some of that on the on the talent search side? But of course, there's issues with opt in, and I think everyone would be concerned, you know, to not be overprofiled. But I think in general, the wave of the future is some kind of combination of opt in and and people more people doing some of this profiling where it benefits them. Everyone wants to be placed with a team with culture fit. It's a lose lose if you go in and immediate clash, and a month later, you're, you know, you're back on the streets. Yeah. I've been on the kinda operating partner, portfolio operation side of a couple different private equity firms, and one of the things that I've seen as a miss when it comes to executive hiring, especially at the the CEO level, is maybe a misalignment between the the skill set of the CEO when it comes to either a turnaround build situation versus a the the business is running fine, we need a people leader and a culture leader. And when there's a mismatch there, it is it very often goes poorly. And, you know, if you're coming in as a private equity firm with a investment thesis that we're gonna three or four x our multiple invested capital over three to five years here, and you believe that the team that got them from a to b is not the team that's gonna get them from b to c, you're investing that time and and you're not gonna go change a CEO in the first six months. You're gonna give him at least six him or her at least six to eight months to get it going. That's a big setback if you didn't place the right person. I wonder again, like, what inputs could we potentially leverage AI for to, you know, get to the get that answer better up front? I mean, it's such a hard problem. You need every possible input and analytics insight as you could possibly get because the first hired gun or a turnaround CEO or this executive, those are the hardest to get right. And so I think clear opportunity to get more data driven. What we see again is, like, they they might know a small network of people that have done a turnaround, and they're gonna over index on that set and not have access to the full slate of turnaround execs in your target sector. And then how important is the sector versus the business model versus, who you're selling to. But having a full slate of candidates I think gives the opportunity to now, okay, bigger data set to work with, let's look at outcomes. We've seen some higher end firms partly using our data to honestly run, like, to back test some of this. Like, let's look at how important it was to have scaled a previous exit or a previous outcome. How many first timers came in and did well? It's just like but I think we've we're only now just coming on the data to actually be able to assess those, you know, from that data driven perspective. And even probably doing some, like, regressive analysis or anything to to say, hey, the people that did have a successful exit, are there any commonalities of their background that could be a predictor of whether they would be a good first time CEO? Because if you get that first time CEO that's gonna try as hard as they possibly can and knock it out of the park, that's gonna be I mean, it's gonna be better from a probably an equity standpoint for the private equity firm and others, but also, you know, for the, you know, team in general. I mean, there's terms like Riverman or River People for these execs that PE firms tend to come in. They tend to do succession planning. They're gonna bring in their slate of operators. And there's some that come in for shorter term turnaround periods, and they tend to execute pretty well. But for that that pool of people, like what is the full slate, and how do we look at getting more data driven? I mean, you can bring over the one you've brought in three times in a row, and it may or may not work out. But I think in general, everyone's looking to get more data driven about this, and I think that, you know, that could be a major unlock. I find that my the biggest issue I have with hiring when I'm bringing an executive on is my own urgency bias. Right? I I often try to solve the problems as much as possible, know, before I'm at a point where, like, it's at a critical mass of I gotta bring someone in to take some of this off my plate or I'm gonna fail. And I think I've already done myself a disservice at that point because urgency bias can be there, I might be tended to make tend to make a hire that I wouldn't have made otherwise if I had a little bit more time. Do you find that the executive search firms have a similar urgency bias as the clients? You know, me in this case would be would be asking them, alright, when am I gonna get this person? Or do they experience it to where they're just gonna start throwing anybody at, you know, the wall because they don't wanna look like they don't have a good candidate pool coming their way? We see a lot of pressure on them. You know, they'll start off a search, and you start with your list of criteria. By the time it gets to month six, you're in crisis urgency mode. And it's like, now it's everything's a trade off. Waiting another three months, massive downside issues, and yet rushing someone in. So I think one issue is we're trying to help them prequalify earlier in the process, look at the full slate, calibrate, get feedback. If you if you're still doing that on month six, you're you're already in trouble territory. And in this I think in this economic environment, there's so much pressure in bottom line, just pressure in general that, you know, not bringing someone in, you're there's obviously gonna be negative impact on that company, and they're operating now without that role or with the previous person you've already decided isn't a fit. So just tons of trade offs. So I think huge pressure to solve this problem well. How are you guys at Hello Sky using AI in your own business? Oh, you mean like dog fitting internally? Or even just AI in general to make it more efficient. I mean, any kind of startup is is obviously leaning into AI to to slow down the amount of hires they need to make in order to but still reach their growth goals. A hundred percent. I mean, obviously, we use it on the talent side for sourcing people for our team. We're we're heavy adopters. You know, we use it quite a bit, obviously, on the development side, like like everyone, but for operations, for marketing, I I'd say we still see a need to have a domain expert go in because there's surface level tools that get, you know, really popularized on social media, but then there's the actual best practices. If you use it the way everyone else is using it, like, it's popular right now, it's you're not gonna get those outsized results. So we, you know, we see the best fit is combining a domain expert with the latest tools. And and then there's just the part of risk taking. Like, it's like they always say, if you do the same tricks that worked five years ago, you're not gonna get the same output today. So part of it is just innovating in in how you apply those tools as well. Coming back to that, like, the eureka moments and all of that, I think what what you just mentioned about obviously using it for development, it's pretty wild what has been happening recently with Claude Code and others like that's just giving you this ability to build way faster than we've ever been able to. As a founder with, you know, the ideas for what this platform could potentially do, do you find that that you're able to I don't know if it's if you wanna call it vibe coding or just be able to give the orders to someone, but are you able to get the ideas out of your head and into the product at a much more rapid pace these days? I think there's there's been some massive unlocks and productivity gains. I think there's still this trade off, and a big debate right now is why is why are companies like Meta no longer hiring out of college, right? Is do you need a mid to senior level engineer to use these tools well? That that is kind of a career ladder problem that no one's really addressed at scale right now. We still hire people out of college, but I will say mid to senior people can work better because the tools, you can vibe code amazing, you know, MVP and demo quality code. You introduce side effects, and cleaning up the side effects are actually much harder and require a much more senior skill set. And so I think the opportunity in the next chapter will be what's the optimal playbook for because what they the main issue with these LLM driven coding is that they have kind of like the memory of a goldfish. So, like, they can see in their context window. They can't see the what a senior engineer has encyclopedic knowledge of system architecture and regressions from five years ago. They don't have that yet, so they can churn out code amazingly well, accidentally introduce side effects. I think that'll be the next wave is having them learn more about the history of the company, the code base, the regressions, so they don't make those same mistakes. But no question, the the degree of output, especially for new product development, is incredibly massive. For maintaining, you know, five million plus lines of code, I think still iterating on that. Speaking of next waves, where do you think it's like, executive search of AI takes over as whatever the vision is of Hello Sky for how AI is gonna help. If we're in twenty thirty, what is true of the world? One of our contrarian thesis is that domain expertise and operating experience will become more important, not less. And the reason is LMs are trained on the Internet, and some of the most important decisions are not made with publicly visible on the Internet. So when you bring in an operator that has to make incredibly complex, sometimes heart ranging trade off decisions about what not to do and what to do. None of that has been publicly visible, and the LMs don't have that. And so operators that have the wisdom have been through crunch time and know how to make those hard decisions and when to wait and when to pull the trigger, I think that'll become much more important, not less. So profiling people incredibly well. So we we talk about unlocking human capital. I think there's tons of invisible talent that's not visible today through LinkedIn and through conventional methods that we're trying to help unlock. And in the future, I mean, there's debate about are we gonna have conventional corporate structures, or are they gonna be more of these, you know, distributed autonomous entities or something in between. But I think if you could map domain experts in any given area, let's say like emerging sectors, not historical industries, who are the top ten or top hundred or thousand domain experts in each category? How do we get through to them? How do we move quicker? I think now you tend to get introduced to people. You miss out on five experts. You miss out on their perspectives. If you had a full map, things would move much quicker. You'd make less mistakes. You'd innovate faster. You'd partner earlier instead of forming two competitors that are now duking it out post funding. So I I think we'll head towards, like, more complete, like, full spectrum visibility into the talent map and domain expert map, and I just think that would be incredibly massive unlock for humanity. Do you think because because really what you what I think a lot of that boils down to is like, more inputs for the LLMs. Because right now, like you said, trained on on Internet. Right? What's out there? Well, I've I've always been the type of person that if I go to a website, especially with my background in marketing, I go to a website and it says, cookies, I reject them. Right? I'm I'm like, I'm not interested in getting any kind of advertising, you knowing, like, you know and I know that there's it's still out there and it's gonna I'm okay connect to LiveRamp and tell, you know, tons of things about me and my search history. But now, like, that I reason I think I do that is because it's never been valuable to me. Those ads or anything I'd get served, they're like, I don't really care. They're not valuable. But the LLMs have been so valuable to me when I give it the proper context. And I'm almost at a point now where I'm like, ****, if there was a wearable that just recorded my whole day and every micro decision I made at work and it had all of that context, then that's a lot of inputs that it could help me remember the action items that came up in a meeting. It could you know, and I wouldn't have to wait for a meeting re transcript or to come to come on and then go somewhere else and access it. Like, it would just be all available. I wonder if we're gonna get to a point by two thousand thirty where everybody is just offering up their lives to train the LLMs. And if we did that, there'd be no such thing as a DISC assessment or a four hour survey. It's based on the inputs that it's getting from just being with you. A hundred percent. I mean, I personally would take some risks there for the opportunity to be so huge. I'd be like, okay, here's all my Evernote notes. Here's this, here's that. Just make me more efficient and correct my blind spots. I remember I don't feed too much personal information in, but I asked at one time this thing, what are my top five blind spots? And it was like a huge, staggering moment. I'm like, how did you know this much about me? So, you know, massive opportunity there. I think memory is still kind of unsolved. You know, you have the Yan Lakun camp that LMs will not get us there. And memory's part of the issue. Like, they have this context window, and they sort of use rag because I gotta combine stuff I'm not trained on with stuff I was trained on and try to, like, fit it into this window to make a correct output. So, like, if they could expand the context window to terabytes or even greater than that, I think that would be a major step forward, you know, technically, I don't know how long will take to get there. But when it if we could fit in more of our life log of of activity and output and we all know when, like, when we're hitting the wall and burnt out, if it could, like, come in and do more for us and know what needs to be done, I mean, I would I would opt into that in a heartbeat. Yeah. Think I would too. Alright. Well, Alex Bates, thank you so much for coming by. Really appreciate the time. Thank you. Thanks so much. Alright.