Episode 20

Humberto Treviño on transforming Venture Capital with AI and Effective Team Building

TIM:Humberto I'd love to hear a bit more about what you guys are doing with venture analytics because it's not an area I know a lot about, and so it'd be great just to hear what you guys are doing and the kind of impact you've had so far.

HUMBERTO: Yeah, over the last couple of years there's been a radical increase in attention into how one can use data-driven initiatives in venture. It's funny because we call it the paradox of the venture. AI paradox causes a lot of the businesses and companies and organizations that in the past have developed or currently have AI capabilities to have been at some point received funding from venture and venture right now is out, by the way, all into AI. It's like the number one investment focus, and we see it; it's just incredible, yet very few venture funds actually have integrated AI and these new technologies into the way they operate, right? We see this going all over the place, going in all sorts of industries, and just changing the way everything is being done left and right. Yet venture continues to operate with essentially Excel, PowerPoint, and PDF, and that's it, right? So, yeah, we started like this road and tried to integrate AI into the way we operate, so more than four years ago, we had a couple of grants that allowed us to fund the initial work. One of them was aimed at learning precisely how AI could be integrated; the other one, since we're impact investors at Catapult BC, where impact investors, we wanted to also develop some AI model to understand how to quantify the potential impact that startups have. So yeah, today we developed quite strong AI capabilities in like the way we operate, and we're able to, for example, find startups using AI. We've essentially built an AI scout, but more than that, we screen the companies we were able to augment the amount of information that is gathered from, like the different deals that we're evaluating, and so on, and we've built these other models that allow us to understand the impact and the potential positive outcomes of investing in a specific company and how much or how close that aligns with our investment thesis and our impact mandate. So that's all the work that we've been doing over there. It's been quite exciting to be honest. Seeing this road, like, four years ago, we just had this big question: Is it even possible at all? And today we have very clear ambitions, and we know exactly where to continue developing these tools and what the next steps are. And as part of that, actually, we recently launched North Star Impact Analytics, which is a spinout of the work that we've been doing here and now making it its own organization, effectively a startup offering these services to other venture and impact investors as well as private equity. So yeah, pretty exciting times, and then really, yeah, quite exciting what we're doing here.

TIM: Yeah, it's exciting because it's innovative, and as you say, there's this weird ironic paradox of investing in all these AI companies but not doing it necessarily yourself or not typically for the AI scout, and then I'd love to know more about the before and after, like how was scouting for new startups traditionally done or probably is still traditionally done by most VCs? And now how has that changed with the AI scout that you've built?

HUMBERTO: I guess venture investors and venture funds essentially have different sources for getting access to deals. Right, of course, you may leverage on, I guess, like the number one that everybody thinks about is your brand and then how your brand is strong enough to pull natural applicants and startups that approach you upfront, but then some others may have as well their network that they leverage on. They also have scouts and internal teams that are actively reaching out to find companies, and that's essentially the area that we're targeting to make, in a way, increase its efficiency and how effective it is. Basically, what many do is like you get these junior profiles, and you onboard them for three, four, or five months. Some people just get some interns from students from the closer universities they got or something like that. You train them on the kind of companies that you want to find, and then you got this sprint where they just go out there and try to find as many companies as possible. and they may be going through databases or web pages that collect startups and this kind of stuff, just trying to populate, essentially, your CRM, and then you basically ask some basic, simple questions about whether this fits or not the type of companies they should be looking at. And when you do that, you're essentially limited to a couple of things: first of all, the human capability, or like the human time limitation, right? As a human, you just can't see so many companies in any specific given amount of time, and the attention span as well that you can have as a human. It decreases with time as you're going through it, and eventually you start getting tired of seeing things, and you just start missing out on great opportunities, so that's one And the other one is, I guess, like you also depend on the experience and the limited experience that more junior profiles may have when you're doing this kind of stuff, right? And it's quite easy as well to miss out on relevant deals or deals that I guess more experienced or more senior members in the team would not do. So what we do and the approach that we do now is essentially we overcome these two in specific ways, but basically we can see far more deals in a shorter period of time than what you could possibly have enough capacity in humans to go through. That's number one, and the second one is that we also learn from experience. And then our algorithms are set up in a way that they learn from experience so they become more and more accurate as they're interacting with more senior members in your team, and therefore you don't have this learning curve delay and then lose this capability when your intern leaves or when your scout leaves. Because this remains there, and you're just honing a more and more capable AI scout, so that's in a way how we're changing the scouting process per se with the approach that we've had today.

TIM: That's really interesting, and it is part of the focus to reduce the false positives or the false negatives more so. For example, are you more worried about the one-in-a-million chance of missing out on the next Facebook and making sure you have a chance at that, or are you more interested in not clogging up your VC's time wasting endless interview time in those kinds of first calls with the company? or is it a bit of both

HUMBERTO: We did this analysis some time ago, and we've done it with different funds and different portfolios, and we essentially were evaluating where the decisions were being made in the different stages in their screening process. and we found out that they were losing the biggest deals in the earlier stages when the more junior profiles were evaluating these deals, and this, of course, is quite infuriating, right? Because that's the business model; essentially, you're trying to find that needle in a haystack. And yeah, so that's where the most like the largest opportunity is, so having these false negatives and saying, No, this is not going to be a great deal because either whoever was screening it was not skilled enough, experienced enough, or just was too tired of looking at this company or did not pay enough attention or time to it. And then missing out on this great opportunity just because of these things that could have been in your control.

TIM: It's really interesting, and I think there's a great analogy here to hiring because, of course, it's a similar thing. You have a funnel; you have outbound; you have inbound; you want to have an effective screening process, but one thing I've always thought about companies is they don't have any sense of the false negatives. They never really think about the candidates they rejected and where they went because it's just a CV that someone clicked a button on. They would have no idea if they rejected the next Elon Musk or whoever, whereas for VC, that is, as you say, an unimaginable pain when you see a company that you actually have presented to you that you could have had a chat with that's gone on to be a unicorn. That's a nightmare, isn't it?

HUMBERTO: That is actually a fantastic point you bring up here. I've never heard of anybody actively chasing on or like trying to ask and wonder what happened with the people that you interviewed, and you decided not to continue to pursue ever after, right? I've been lucky to come across some of those who, in the past, by X or Y reason, ended up in another organization, and then we just cross paths again. And back then, when I remember, like these guys were, they were good; they were great. We just had some other profile that we were confident was just fantastic. Yeah, and actually now that I think a little bit more about it, we could have made a better decision back then, and it's great. We never do think about it, and just like we developed this methodology in-house to try to evaluate our decision-making, I don't see how that right now is fitting the HR processes into realizing where you're screwing it up in your process and you're not making the best decisions.

TIM: Yeah, it's completely missed and glossed over entirely. I wonder if in hiring it's partly because a lot of the decisions are human decisions. We're very confident in our abilities in making those kinds of intuitive gut-feel decisions. Even if someone then goes to another company and does really well, we might still think of ourselves as being bright and say that was just their environment at the time they interviewed. I still backed my decision that I made. And maybe it's not maybe also the error is not as obvious as with a company where it's they're now worth a billion dollars; oh my God, what have I done as opposed to, Oh, okay, they're now CEO; they could have had this inquiry here.

HUMBERTO: yeah I mean it's fair and as well you know There are so many other factors that go into hiring a person right unlike the context matters where they went to the kind of organization the resources they had you don't know that and you don't have any control over that right you don't know what would have happened if you had hired that person because of all of these different factors but you do get to see and I guess understand if you hire someone and you were betting on to this person and very shortly after you hired them they pretty much flopped and ended up either quitting or going to another organization or just not delivering the value that you were expecting them to deliver And that's like the easiest way to see things and where you could have thought, Okay, maybe these other profiles could have been a better solution, and when you see the other person, maybe it came back to your process again, or maybe you got engaged with another organization and in a similar role, maybe your competitor, and then they're just killing it. Yeah, it's not as easy, but nevertheless, it's something to do and definitely has some value in evaluating your decisions.

TIM: And I think now again with AI It's AI is changing everything so profoundly that it's really good to sit down and think about some of these fundamental issues again that have been glossed over because it's like measuring the ones that got away for a company probably hasn't been in the top 20 problems because so many other things about the hiring process have been screwed but now maybe there's an opportunity that, hang on, we could be tapping into LinkedIn API or an Apollo API, so often understanding where the candidate is measuring if they've been successful or not, like there could be a lot more that's done that isn't currently being done.

HUMBERTO: But I do agree there are so many bigger pressing issues at the moment with where The status of the industries and what were the AI tools in place and how many roles in organizations are changing and the way we do work is changing because of these new tools coming into place And because of the same thing, I don't know if right now the attention or, like, the capacity for an organization to pay attention to that would be where they would be getting the most value at the moment. Right? There's so many other processes that you need to fine-tune.

TIM: Yeah, speaking of hiring, you have built some very substantial teams across a variety of industries. You've learned a lot, obviously, through trial and error, through success, and probably a little bit of failure along the way. What are your biggest lessons from hiring and team building as well?

HUMBERTO: one of the biggest things that I focus on when I'm trying to hire someone is rather than like the technical skills are a kicker right like you need to have it at least to be able to deliver whatever is like the immediate requirement for the role but I focus quite a lot as well on like the soft skills and like the capacity of it your new hire to deliver because as you're hiring people you're now delivering through people rather than being you yourself the one that's delivering Through your own individual contribution And the capacity that this person may have to successfully get involved with the rest of the team to contribute to becoming a good team member and then in the end how will his contribution guarantee that your team remains or becomes and remains a high-performing team, right? And in my opinion those soft skills are probably the hardest ones. and the most important ones when it comes to hiring someone right, because you can always teach technical skills, there can be a gap between a specific system, a specific language, a library, you name it, but the one thing that you cannot teach someone is like those soft skills that are intrinsic to the person. And to me that is just a skill thing to take into consideration whenever you're hiring someone to join your team.

TIM: And when you say soft skills, like what do you mean by that? Do you mean like communication? Do you mean more like their personality? What are you looking for?

HUMBERTO: Yeah, so there are things that, like I said, you can't teach, like drive or embracing teamwork, or the curiosity for me is just fundamental, and learning skills, the capacity to learn new things and unlearn other things, right, the desire to solve problems, that ethics, and just being a truthful, honest, transparent person, these kinds of things make a big difference in when you have someone already in your team and when you have someone that does not—that's not compatible with the values that the rest of the team has. Like I said, especially in data, the skills, like those hard technical skills, keep evolving, and they're so ephemeral; they're valid for a few months, and then they're just not valid anymore, so whatever remains is whatever you're actually carrying on as a human and whatever is part of your fundamental values. Another thing that I find very important, and I mentioned briefly or I said, is the drive that you have to what's moving you as a human, what makes you think, what do you want to achieve, and in the measure that you have someone that has the right drive that is compatible with whatever is the context in your organization. The more you can guarantee that you're going to have a successful team and a successful team member and someone that will be pushing in the right direction, because you can sometimes it's just money, and if it's just money, what's driving them? Anyone's going to show up with five more dollars, and then the guy's going to be gone. This kind of thing is super important when it comes to hiring the right people.

TIM: I think anyone listening to this would agree with you 100 percent, and so would I. I guess the challenge is how do you identify or measure these things in the hiring process as opposed to maybe in the first few months of them working there and observing how they behave, which is an indicator of their values? How have you approached that in the past?

HUMBERTO: Yeah, it's not easy; I agree, but at least, like every time that I've had, like, for the most relevant roles that I have within my teams, I typically have a longer process than what is normal when it comes to the typical process because, like, most people would just have one interview and then that's it, then you're hired, right? Some companies, of course, the more structured ones, may have some challenges or technical evaluations, and like, Here is this code; go through it and tell me what's the problem with it; solve it, blah blah blah, right? This kind of stuff I like to spend time with them. I like to talk with people. And even actually as part of my interviews, I actually do spend a good amount of time trying to learn more about who they are, about what they did on the weekend, what they do when they're stressed, and what happens. Tell me about, like, the worst situation you've had at the worst, the biggest challenge you've had at work or in personal life. In some cases I've even had a couple of conversations, and I invite them to have some coffee and maybe go for some lunch, like truly get to know the person and understand their underlying motivations for being involved with us. Of course, when you talk with references and so on, I actually do talk with previous employers. And just spending a good time just talking about how it was to have the person around and understanding, like, his engagement, rather than what he was delivering, which is what a lot of people just focus on—the heart deliverables—I want to know what it was like to have him around. And I think those are some ways to like that you can actually have an understanding, yes, of course, the more you spend time with a person, the more you will get to know it, and there is a limit to how much time you can actually have someone involved in a process until that person becomes tired, annoyed, or someone else finds him. But I figure that, yeah, the more time you dedicate to like developing these skills as a recruiter and just doing it over and over again, you start getting that sense and being able to sniff what you're looking for in a person or in, yeah, in a team member.

TIM: And with this approach you've taken, can you think back to the people who were either successful or unsuccessful in these kinds of in-depth conversations you had, and are there any patterns that have emerged? Why would you have said no to, let's say, the 90 percent versus the 10 percent, or was it just varied from Canada to Canada?

HUMBERTO: Yes, what's driving them? What is actually driving them, and how does that have a How does that match the current situation where the organization is right? Like people I don't know, for example, I don't want to be too specific because I don't want to, yeah, if someone ever hears this. But we had this case of one individual who happened to have the perfect technical skills and at least experience on paper for being in a similar role to the one that we needed, but there was something odd about him, and we just couldn't tell what it was. We were quite pressured in terms of time; we really needed to hire that person. And something didn't really match with the people that he said that he knew, so we let it go. We hired a person we liked. Like I said, we were a bit in a rush, and afterwards, one like this person was actually actively involved in the role. We figured out that, yeah, there was a bigger story behind how and why he had left where he had left and his drive to be where we were that we quickly realized that he just was not the right profile to have around. Yeah, he did not have the right reason and the interest to be with us. He didn't have the same drive; he didn't really have it and wanted to be part of the success story that we were trying to write. He was just there in a way just to use us as a stepping stone to a bigger goal that he had, and it was quite evident as soon as we got him on board that he was just not interested in that whatsoever right away. Yeah

TIM: Yeah, I'm sure those kinds of examples must burn in the memory, the ones that didn't work out.

HUMBERTO: It does burn in the memory, and you learn from them, like whenever you have this gut feeling and something's telling you something's not right. Yeah, more often than not, it happens when you just have to take a decision and you need to move forward because otherwise things are burning, but more often than not, especially when it's about people, that gut feeling kind of matters, and you end up realizing a little bit later that we've probably had the right hunch.

TIM: Yeah, it's interesting with the gut feeling because that's the other end of the spectrum of a data-driven approach but maybe it's just because we don't know enough about the history of humans to understand what exactly a gut feeling is, and maybe it's almost like an algorithm built on tens of thousands of years of our experience as a human. Maybe it's a lot more intelligent than we give it credit for.

HUMBERTO: It's just exactly what it is. We're essentially a model-based system that makes decisions based on our experiences and the culture and the context we grew up in, and based on that we're able to make decisions and split decisions, and some of those are taking the system one and system two, and some of those are taking the system one without us knowing exactly why. But there is a big reason behind these things, behind our experience, and we are not always able to articulate straight away what is going on there. Yeah, I guess gut feeling is more than just gut feeling; I guess it's not that much of a, like, and I know it is very strange to hear this, especially coming from me, but there is some reason behind why you're feeling what you're feeling. So I probably should pay attention to it as well.

TIM: Last time we chatted, you mentioned that there can't really be anything more important in your job as a leader than hiring the right people, and maybe people talk about doing that all people are our most important asset, blah blah blah, but it doesn't really connect to how they treat the hiring process and how much dedication they actually give to it. So it'd be interesting to hear your thoughts around, yeah, how do you actually focus on hiring? How do you elevate it to the priority that it actually should be?

HUMBERTO: If you're hiring someone, it means that you now deliver through these people, right? If you're responsible for a team and you're no longer delivering through your own contributions and your own actual work, you're not. Then these are the persons; this is the resource that we'll be delivering whatever will be used to measure you; therefore, you should absolutely guarantee that whoever is coming into your organization will be able to deliver and exceed the expectations that are bestowed upon you. Now if this is true, then why on earth would you not pay maximum attention to making sure that you get the absolute best candidate that you could possibly do to guarantee that this is the true case? the true case moving forward in the future, right? The one that's going to continue enabling your team, the one that's going to continue creating the right environment, the one that's going to have the same drive, the same focus, the same attention on the same things that matter for you. and that you will be able to deliver for these people eventually. I'm often surprised by how, and I guess I see it more often in first-time managers and individual contributors who also need to hire people, how little do they truly understand how critical it is to hire the right person. Or that they want to delegate this responsibility to the HR roles and the HR partner or whoever is helping them get the right talent or, even worse, like delegating it to a direct report who then has to hire his own peer. Yeah, I just, for me, just hiring the right person and building the right team is incredibly relevant incredibly important, and you need to dedicate as much time as necessary to guarantee that you're getting the right talent on board, yeah

TIM: Yeah, and if I think back to my own experience of hiring over the past 10 years or so, I feel like probably in retrospect my philosophy was a little bit offbeat. What I mean by that is I was of the view for several years that I would hire someone that I didn't have a reason to not hire if you see what I mean, like they went through each stage; they were fine. There wasn't a specific reason that I should not hire them for me. I was like, okay, fine, I'm going to hire them because they haven't failed any stage, but I now probably realize in retrospect, especially running a startup where each person has a profound impact, positive or negative, on the organization, is no The bar has to be higher. It should be no; they've excelled at each stage, and they are the right person as opposed to they're not the wrong person. Is that your experience as well?

HUMBERTO: Absolutely, and especially I'm glad you bring up being in a startup situation, right? Because in a way, also the biggest challenge when it comes to hiring someone changes as well, given the context where you are now. You were saying I was hiring for whoever didn't give me a reason not to hire him rather than just looking for that perfect piece that you absolutely need to make sure that whatever you're trying to achieve is guaranteed, and that's in a way part of one of the things that you really need to take into consideration when you're hiring people. So how do you envision your team in one year? How do you see it in two and three years? What will be the needs? What is the current context of your organization, and how will it be evolving? Where would you like it to evolve to, and then based on that, what will be the pressures and the kind of needs, and how will the roles and the requirements from your team and the people that report to you be evolving? and in that sense, then, are you hiring for that person that will fit into this current situation but also in the development over the next years, right? And when you look at it from that perspective, then you're going much deeper when you're evaluating people, right? And then not just hiring someone that just ticks the marks but actually really does fit that longer-term vision.

TIM: Yeah, it's interesting the way you frame it there because, again, if I reflect on the last five years of running a startup, I feel like at times you can get stuck in this almost like lean startup obsession where it's just, Let's get the next thing out. It's all about you doing the MVPs of whatever it is you're doing. If it's in the product or marketing or sales, it's almost anti-vision and strategy in a sense because you're not often thinking about the longer term, but yeah, having a vision over the person and how they will grow in the organization must force you to really think about whether this is the right person or not for the long run.

HUMBERTO: Exactly. I'm particularly in startups where the tech skills are quite important in delivering that immediate need that you have, but things will look very different in two, three, or six months, so do they have those soft skills to adapt, to evolve, to pivot, to solve the challenges that will come? More often than not, in startups you end up doing things that are outside of what the strict role description goes. And that will be true for the next year, so do you have someone that is capable of adapting to that, and that will evolve and solve and deliver and help contribute to the overall chances of success of your startup?

TIM: Thinking now about screening, so this is again an analogy here between hiring and the work you're doing with the AI scout, so I think this is especially interesting. CV screening is a typical first step in the hiring process. Someone is going to get someone's CV and look at it. someone in talent acquisition, a recruiter, a hiring manager, what have you Traditionally, that's just been done manually; now we're seeing the development of tools, AI, for example, to stack rank CVs or grade them, compare them to a job description, do some kind of automation. What are your thoughts on that? Do you think this is going to be of benefit? Do you see some issues with this approach, especially given you might have gone through a similar learning process with the AI scout system?

HUMBERTO: similarities that I see between the process of like how having an AI scout for people and having an AI scout for startups I guess in a way, no matter how similar the startup is on paper or how adequate this startup is to whatever you're trying to look into or, you know, what you want in the end, it's all about the human element and, you know, in this case, how the founder is because those are the ones that will be driving the company. How will they drive success? The same happens with, you know, screening humans. You may, no matter how good this human looks on paper, it may tick all the boxes, and just like we were talking before, it ticks all the boxes, but does it have that human side of things? Because you're hiring a human; it's a human resource. And that is something that perhaps AI is not ready or not completely sure that AI is ready yet to truly tackle maybe in the future, but yeah, so in that sense I guess there is another I guess another challenge when it comes to using these new technologies is hiring. On one end, there is like this overhype of what AI can truly do at the moment. Those who don't understand the technical side or those who are not that involved with the technical side of AI, especially if you're like an HR partner, all of a sudden they tell you there's this new magical system, and it's a black box that can find whatever exactly you're trying to figure out really quickly, really fast. without you even having to pay any attention or putting any effort, I feel like sometimes things are sold beyond what they're truly capable of in a way, and like I said, like the users may have greater expectations of what the actual capabilities are, and they may relax and then just hope, like, Yeah, the system will do everything for me, and I don't need to, like, really pay attention to it and just trust it blindly. And then just not get involved, and then, in my opinion, this is a recipe for disaster because the human in the loop, in my opinion, is key in any AI-supported process, and you need to have the human looking at it and paying attention to what's happening, how it's happening, what it is that is rejecting, just like we were talking about—is it rejecting those false negatives? Why is it making these decisions? Which decisions have been made lately? And take just a general look at what it's being recommended to you or whatnot, right? Yeah, it definitely is still some sort of human in the loop, human involvement into making sure that things go in the right direction, and I guess these things are in a way inherently human, right? But yeah, I guess that is part of my concerns or challenges with that that I see with how this technology is being used today, probably.

TIM: Yeah, a hundred percent, and we're definitely not there quite yet. I wonder if, with hiring—at least this is my view anyway—the way it's been traditionally done is so fundamentally flawed. I feel like there's so much upside even if the AI is imperfect, even if it's going to have false positives and false negatives. we might overtrust it The current situation for most companies is that they would not have any clue why they rejected 10,000 candidates because it's just someone in a system in an ATS clicking a reject button. There's not really any evidence there's no transparency. There's obviously a lot of bias with a CV, and it's got someone's name, their ethnicity, their gender, their blah blah blah, so in theory, our system should be able to rank a CV created, give it a score out of a hundred, and give it reasons why it was rejected, give it reasons to the candidate why it was rejected like that, and with the current technology, it is possible. It might not be super accurate, but it's possible. and so I hope that we could be on a step change, but I could certainly see how there's going to be a lot of new stories coming out about so-and-so tool rejecting this kind of person or this type of person, which will probably create a lot of fear in the interim, I suspect.

HUMBERTO: Yeah, and I've been like, I've come across stories here and there on LinkedIn. This HR manager was really upset when he figured out that he himself was being screened out of the role that was exactly for the description, and I guess that is part of the reflection of humans not getting involved and not overseeing what's happening and understanding that this is not a magical thing yet. Like, it's not you in a computer; it's just a program that's making decisions based on things that can be adjusted and improved, and so you need to make sure that you are overseeing it and adjusting what could be done to guarantee that the system delivers what you're expecting. So not taking a hands-off approach

TIM: And a bit more broadly now about AI and hiring, so we've seen already candidates use AI to optimize their CVs to apply en masse to numerous roles, and companies are now using it at that CV screening stage. Some AI interviewers as well, it's like an AI bot doing an interview, either a video or a chat bot. any other thoughts about where AI will be in hiring in the next couple of years like any bold predictions that you might have

HUMBERTO: There are already some quite clear use cases at the moment when we say transcribing interviews, summarizing these interviews, and then I don't know to what degree they are using this learning from, like, those interviews and transcribing it, like, making decision-making and, like, the actual screening and kicking out. Like I said, I'm not very familiar with the actual extent of where the current situation is. I am definitely, if I had to bet on something, if we're not there yet, it's definitely going in that direction because the entire process of hiring someone is just a deal flow; it's a pipeline, basically. right there where you're you have this funnel, and you have a lot, and you want to get to one, and that process as you're screening out in the different steps that you've set up, like the ambition, definitely will be there into having a system that makes complete, full decisions by itself until a point where I guess like you have the least amount of human intervention as possible I guess that will be the aim now; is that something that we want? I don't know if you're hiring humans, and those are the humans that will be involved with you. Like I said, how far will the technology actually be able to gauge and measure the things that I was saying and speaking about before, those soft skills, those drivers? Those things that you can see and that you can feel, and like how compatible a human may be with, like, your current culture, these kinds of things you cannot—you cannot for you to be able to build something like that. Not only do you need to be able to analyze a human in a much deeper way than what we currently are capable of doing, but on top of that, you also need to analyze your culture. What drives your culture? your context, your company, your situation, and in my opinion, we're very far from that, and two, do we really want it to be like that?

TIM: Perhaps it's a little glimpse into an Orwellian future, but that said, I was told about something this week, a really cool use of AI, which was in tracking animals and cattle. I think, like cattle, whales, and what have you, apparently the difference among a hundred cows faces mathematically is vastly different, vastly more different than among humans faces. obviously for us, human faces are much more different than cow faces are, but according to actual numbers, cows are more different. I wonder whether then there are differences among humans that we can't spot or see that an AI machine might end up being able to see, and it might almost be a quantification of this gut feel that we have or maybe something that's off a plane. We can't even understand that. That would be fascinating, wouldn't it? I wonder if we'll get to that point.

HUMBERTO: for sure, in the end, those differences are nothing other than parameters, and what is it that you're feeding the system, or what is it that you're evaluating? And in that regard, it can be anything from parameters in your DNA to physiological parameters and all these things that we just don't pay attention to that a system can actually pay attention to. Yeah, but when it comes to how relevant that would be in a hiring process, if it were able to get into those deeper motivations, those deeper analyses of who you are as a human, what drives you, what your values are, What are your values? What are you trying to achieve? what and in that sense that compatibility with your team Yeah, probably in that direction. Now there are already many models, by the way, on you have all these models when you feel some information that tells you about your personality and things like that, but I feel like those models which, by the way, has already been used in or has always been used in hiring and recruitment processes that don't truly surface all the other things that are embedded in your history, embedded in your context, embedding your ambitions, and all those other things. The only way to figure them out is by interacting with a human.

TIM: Yeah, we figured midway through scaling at our team that the key for us in the hiring process was to try to set it up. So that people had an opportunity to demonstrate with their behavior the values we were after as opposed to just talking about them, we felt that the interviews that were just like, Tell me about a time you did this; tell me about the time you did that, were fairly weak because if someone's a good storyteller, they had good examples already set for demonstrating a certain value. Then they would excel, but like talk is cheap, and so we tried to devise a way that within the hiring process a candidate could demonstrate something, so I'll give you an interesting example: for us, we basically figured out the most important thing for us was could these people learn new things quickly. Because, to your point, things change so quickly that if you don't have someone who is willing and able to learn new stuff all the time, then they're not going to do very well in a startup in a fast-paced environment, so then we came up with something that is cool for our software engineers: we asked them to do a short coding challenge in R, like the statistical language R, which no software engineers know. And so it was a very interesting test to see if they would say, I'm not interested in this. I don't care about R. It's a stupid language, or if they just couldn't do it. There'll be so many ways that I feel like they could fail that test, which would be a very good indicator for us that they probably weren't going to be a great fit. And so that was a way that we hacked the hiring process to force them to have an opportunity to demonstrate a behavior. Thinking about that now, have you seen anything similar, like any other approaches to hiring that are more based on the actions speak louder than words concept?

HUMBERTO: Yeah, I guess my latest or most recent recruitments were something in a very similar fashion. Actually, we were hiring people who were completely unfamiliar with venture, and then we would have all these discussions about where do you invest, why do you invest. and like trying to think how, first of all, trying to evaluate how fast they were able back then to understand how venture made decisions and then the type of tasks that he had at hand, we gave them a big list of companies, and they had to run models and explain what's going on and think not from a data engineer but as if you were in the role of an investor, which was not what I'm hiring you to, but I'm trying to figure out if you're able to understand beyond what you're doing. and if you are able to learn quickly about what we're having about the discussion we're having and so on, also another type of exercise that I was involved in multiple times was in this. What was the name of it? It was not a case study, but it was a—I'm trying to remember the actual name. It was like this, some sort of workshop, right? And we would invite all the candidates. This was a position for high-performing young profiles that Yes, thank you very much. An assessment center, so like the problem with assessment centers is that you have to require a lot of time and some senior profiles, ideally, to be involved in there and time from the candidates as well. but those are great examples of how you can actually gauge this and measure these other things right. You put them in a situation, and it doesn't matter what they are trying, like the answer, the actual answer they're coming to, how they were thinking, how they overcame this challenge. You actively put them under pressure and try to figure out, okay, So are they able to handle this? Are they able to, even with the people that they're competing with for the role or the position? Can they actually work together and overcome the fact that they are technically adversaries in this situation, but they need to achieve a role, a goal, so can he overcome this kind of thing? and this is just an example of, like, different ways that you can, yeah, you can test these soft skills, but what you did, and I figured this is like the most relevant thing rather than what and how you do it, is like understanding that you need to figure out these specific soft skills that are important for succeeding in the role. And you can do it in very complex ways, this assessment center, so you can do it in easier ways with small cases, one hour, and a couple of questions, trying to think quickly and evaluate this specific target.

TIM: I'm wondering if you had any general advice for, let's say, new leaders who maybe are going to try to build their first team. You've got a lot of experience here. Like, what should they be doing if they're building their kind of team from scratch? What should they be thinking about?

HUMBERTO: I guess the first thing that you need to understand is that now you're going to be delivering through people; now you're going to be delivering through someone else, and the process of hiring these people is incredibly important for you to be able to succeed at your new role. The first thing that you need to understand is that your new role, or your role, is now different. You will no longer be measured by whatever good things you're able to build yourself but with how you're able to put a team that gets the right motivation and the right clarity to achieve the new goals, and in that measure, you need to guarantee that whoever joins your team is going to help you achieve your future goals. Therefore, recruiting becomes absolutely critical, and you need to hire someone who will be able to evolve with your role and the requirements of your context over the next few years. If I give you a few pointers, will you then understand your context? What will these challenges be in the next one to three years? I'll draft this plan of what your ideal team should look like if you want to, and how would you like to evolve your team, by the way? But if you'd like to guarantee that you've achieved this plan for the next one, two, or three years, understand as well that your level of involvement with your team members will change, right? A lot of the time, new leaders struggle with understanding how to be a leader and understand that, I guess, like the initial approach is just to delegate something without realizing that the human on the other side will have different levels of needs and the situation will have a different need. and the way that you engage with your team needs to vary as well depending on the team, the context, and the situation you're in every moment, and by the way, hint: if you're trying to figure out if you or, I guess, like an easy model to start learning how to be a good or effective leader, there's like this very—it's quite old, but it's very good for, like, new or first-time leaders, like this idea or, like, the situational leadership theory. which, in my opinion, is a great first step to becoming an effective leader and understanding how to engage with your team, and then I would say put as much care into tech savviness for, like, delivering the short-term requirements as you will in soft skills, and then, yes, finally remember that tech gaps can always be bridged, but drive and intrinsic values from humans, those things stay with humans.

TIM: Awesome! Great advice from a lot of experience, I'm sure.

HUMBERTO: Hopefully, let's see if someone finds it useful, and I guess those are the things that I learned from messing it up multiple times and getting it right many other times as well. Few things are as great teachers as experience, and very few people actually learn the heads from others. But if anyone listens to at least one of these points, I think they will be better off.

TIM: awesome