DataRobot Chief AI Evangelist
University of Utah Doctor of Philosophy Dropout, Chemical Engineering
Current Time 0:00
/
Duration Time -:-
Progress: NaN%

How did you get to where you are today? What is your story? What incidents and experiences shaped your career path?

Summarized By: Jeff Musk on Mon Sep 14 2020
um, e think for think For me a lot of it is luck making the right decision at the right time and for some of the right decisions. There was a lot of confusion where the decision was an obvious at the time. Eso for my background. I studied chemical engineering. I planned on going to medical school and I ended up getting a job. But until Micron and, um, I worked there for five years, and then I had an opportunity to go work for a hedge fund and that that really came from my my own personal motivation with high performance computing. So I was doing things on the side, not related to my job where I loved GPS. I thought they were really awesome. This is before deep learning in that that allowed me to get a job at a hedge fund. But then, working in the hedge fund allowed me to get a job for Sequoia Capital Company called Higher View, where I was their chief data scientist and build out their data science team. And that experience built up my confidence to go to a startup and make the leap three years ago and now I We were purchased by data Robot and now I work for data. Robot is their chief ai evangelist, which is a really, really fun job. But the path to get there was was not necessarily planned.

What responsibilities and decisions does one handle in a job like yours? Tell us about weekly work hours, including the time spent on work travel and working from home.

Summarized By: Jeff Musk on Mon Sep 14 2020
Yeah, So Data row. But they are. They started as a nought oml company. So auto mail stands for automated machine learning. So, back in 2012 the initial concept of data robot waas they could automate the process of building a competitive predictive model and back around 2012 into 2014 that was a controversial topic. People thought you needed expert data scientists to do that. And now a lot of other companies have followed two, and they've realized that not only can you do that, but you can do it better than most state of scientists. They don't have the experience, and since then did a robot has really evolved mawr into an unending end ai solution. So they go all the way from data capturing, did gathering, cleaning all the way through the model building process and then an entire part of model deployment that a lot of data scientists are not familiar with. So they take care of allowing models to get shipped into production, but also monitoring them if there are problems and usually for any model that matters, that's a really big deal. So, banking where you have compliance or in manufacturing, you can't have models that drift. And they could do it for lots of reasons. They could drift because features drift, meaning inputs, and then my roles and responsibilities. This is this is kind of a new area, this evangelism piece. And so my role is I I'm a co host for podcast called Mawr Intelligent tomorrow. So we have a lot of executives that come on that podcast, Um, CEOs and chief data officers we had Congressman Will Hurd on, um, looks like we're gonna have someone from the FBI coming on, um, executives from Google and IBM talking about AI strategy. So that's that's really fun to generate, thought leadership, but also get access to very important decision makers that are harder to get access to. And then the other part of my role is maybe anchored in storytelling. So if you're if you're going to go give a strategic talk or a keynote, can you do it in a way that your, um, the talk is memorable? It's inspiring, and it leads to new opportunities, and in doing that is something that is less common. If you go to a data science for AI conference, I would argue that the vast majority the talks were forgettable

What tools (software programs, frameworks, models, algorithms, languages) are typically used in a role like yours?

Summarized By: Jeff Musk on Mon Sep 14 2020
um, it's a great question. You have tools that feel more academic and people will. They won't like this, but I would say our is maybe more for research and pythons for production. Um, when I was in school, I was taught Matt lab and I thought it was great. I became an expert in Matt Lab, but I noticed it was really holding me back until I wanted Python. So I use python when it comes to when it comes to machine learning algorithms. I think Python has done a lot to catch up because are used to lead our. If you wanted the latest random forests, you have to use our But now, in Python, I would argue that not only this python have a lot of your standard clustering and machine learning algorithms. It also has a lot more deep learning support. Deep learning. There is a lot of confusion in the market, and you see it by the number of framework. So Google has tensorflow. Mxnet has, um, Amazon has mxnet. Facebook has pytorch. Baidu has paddle paddle. You have all these different frameworks, and we see for people that are getting started, a lot of them use tensorflow because of the Google backing. It's popular but tensorflow for anyone that does things in production scale, it is not my first recommendation. So we used mxnet in our startup. But I think PYTORCH is starting to overtake mxnet. And so the recommendation for the students. It's really important to use a framework that you feel comfortable. Um, it's not necessarily hacking, but you feel comfortable customizing. So if you need to get under the hood, a good comparison would be if you wanted to be a core contributor and actually contribute code back, I would say in Pytorch in Mxnet, that would be, ah, lot more straightforward than, uh, Tensorflow. So I'm very outspoken critic of Tensorflow. Hey, I hate tensorflow with a passion. There's a core of question out there that's titled Why does Ben Taylor hate tensorflow? And I don't know if anyone's answered it yet, but a Yeah, so please use pytorch or mxnet and carries is a good place to learn, but I wouldn't recommend it for productionWell, I I think I have a bias because data robot is being for a lot of ai. The interesting thing. Here's the vast majority of a I actually never makes it to production. So I would say anywhere from 80 to 90% of AI initiatives inside very large companies where you've we know their brands who know their names. They are buried in notebooks, Jupiter notebooks, python scripts. They don't actually go into a formal production pipeline. But then you have customers that use data robot and they all the models go through our production pipeline, and they're they're managed. There's the scale. They run on the cloud we can support. Multi cloud. And so it's I don't have a good answer that question, because I don't have good examples. Toe Look at where people are shipping models into production unless you're fame company. So if you're if you're very large company, they've built things from scratch Internally, they have their own pipelines that work. Some of them are public. You can understand how the production of stuff at Facebook in Netflix, but for most companies that don't have a very strong AI research group, um, they when I was at higher view. We took advantage of a lot of the Amazon tools and wait for our our startup. We took a lot of advantage of also the Amazon tools using Ramdas using server less because it allowed for insane scaling. We could support AH, 100 million inferences per month, and we didn't have to. We don't have to worry about it. We could sleep at night and with server list that just runs. So if you look at the last 10 years or 20 years, huge improvements have been made when it comes to auto scaling deployment. Even looking at stuff like Docker in imagining how people used to deal with virtualization, it's it's amazing. Or the Twitter clone in five minutes, Um, which is really exciting for students because they can do stuff in a weekend where 10 years ago, we could not even imagine being able to do that, especially as a single individual

What are the challenges in a job like yours? What approaches are effective in dealing with these challenges? Discussing examples will help students learn better.

Summarized By: Jeff Musk on Mon Sep 14 2020
eso the challenges in my role. The when it comes to the research, most of stuff I work on, it's never been done. There's not a template, so there's not a template or a recipe to follow its new research there. There's no known conclusion if this will work or not. But that's kind of the excitement of it, because when you get to the other side, you get this high on you know this feeling of accomplishment. So last year I give a keynote on teaching I to play call of duty on the Xbox, and so figuring out how to take over the Xbox with a GP machine and making it work and control the controller and consuming the frames and react to them reacting to them was a lot of fun. So I think the important thing is to bring a sense of urgency. And so I when I was in graduate school, you're working really hard. But the time lines are long, and in a startup you're working really hard. But the timelines is short and the luxury of doing R and D that doesn't get delivered to go to value that you can't maps. Um attribution back to. That is not a luxury you have in a startup because, um, someone's paying payroll and it's not for free. And so that that's been really interesting, because when I managed to data signs seem I worried about their happiness because there it was easy for them to get poached to go work for another company. But after paying payroll and running a startup, you obsess about every dollar that's spent in payroll if you're not delivering in 30. But if you're not delivering value in 30 days, why are you working here? It really changes your mindset. So So the challenges are, too. Try to bring a sense of urgency. Uh, the danger of that is this concept of heroics. So a lot of times in the start of space, they talk about heroics. Last year I slept in the office during the week of Christmas, never leaving the office where we're pair programming and taking power naps every four hours for a week. And so some people in the startup community would see that is great. That's heroics, but normally when you're doing heroics, it's because something is broken down. You've made a mistake on a process or you've screwed up. You haven't been strategic because urgency could be a good thing, but can also be a really bad thing. So you kind of have this constant battle between urgency and strategy, urgent thinking and strategic making.

What are the job titles of people who someone in your role routinely works with, within and outside of the organization? What approaches are effective in working with them?

Summarized By: Jeff Musk on Mon Sep 14 2020
s a lot of them are chief data officers, CEOs, the so the approaches with working with them. It's it's easier to work with them if you kind of approach them is appear rather than. And also if you could be authentic. So you're not. You're not trying to hurry and pitch them some rational argument on why you should buy their product. You're trying. And then I also talk a lot about the favor exchange. So if we have a CEO or CEO on our podcast, there's a favor exchange there. And the question has to be asked, is What favor are they getting from me? So what can I give them to me? Because their time is very, very valuable. Um, and so I find it's useful to have a favor exchange. So any time I'm speaking, I like to speak in a way that the audience can leave with something, whether or not they buy my product or whether or not they follow up. And then if I have a CEO or someone on our podcast, I always think of a way that I can help them, whether it's ah, networking introduction or its recruiting a lot of these individuals are recruiting and they need to go after a I talent. And so being able to throw a I talent their way, um helps. But sometimes it's nice to just share stories. So knowing that there are other people that know what it's like to deal with legal lose contract when a contract fire employees, um, deal with stuff like that can be, that could be helpful as well. So I have some virtual coffees that I'll have my calendar. Well, I'll just reach out to other people in my industry, and we'll just we'll just have an unstructured conversation because they're dealing with hard things. And sometimes it's nice to have someone who can listen because your employees aren't necessarily. Employees want you to be the captain of the ship, So whether you're the CEO of a company, they expect you to be in charge. They expect you to not show signs of weakness or doubt. And so sometimes that's nice for these types of individuals to talk confidentially outside of their organizations about seriously concerns and problems that they're having, that they maybe don't feel comfortable sharing publicly or internally

How would you describe your management style? How has it evolved over the years? Can you tell about experiences or books that influenced your management style?

Summarized By: Jeff Musk on Mon Sep 14 2020
So my management style. It probably started mawr from, um, an employee pleaser or a mentor, where I'm I want to mentor people I want them to develop and maybe going through the startup process. I'm I'm in a position now where as my team grows, I, um I am okay with mentorship if it's light touch. But the moment I detect spoon feeding, I get very, very irritated. So I we noticed that working with some capsule on projects with some other universities, it felt like we were spoon feeding the students. And the biggest thing you want when you're hiring is you want someone who is passionate ideally, someone who's obsessed where you don't have to tell them about a new AI breakthrough or some new algorithm. They just figure it out. And then the other thing we noticed, too, is for some of the projects we work on. You get blocked and you get blocked on things that would block anyone. You're trying to do this new. You're trying to apply this new software stack or you're trying to use this new application, and it's never been done before. You don't know how to do it or the Xbox example you're trying to take over next box. How you gonna figure it out? And we want eso. Now, I I would I really focus on hiring people that right, unblock themselves. They're just very obsessed, compassionate. And I don't I'm not actively hire hiring right now, but in the next year or two, um, if I could turn the budget to build out a team, I would love Thio announce, um, an interview with that would probably upset the data science community. So in the interview would essentially say it download the latest version of extra boost and make it run 20% faster. And a lot of people in today's times community would say that's not possible. And for people that say that's not possible, then I don't really wanna work with them right now. Um, so you're looking more for creative thinkers, But people experience brings a lot to the table because if you have a lot of experience, you've seen these problems before, you understand their possible that and the other thing I that is kind of hidden under that. That statement is there's a lot of low hanging fruit and popular algorithms, so you downloading this deep, deep learning framework or our extra boost or other things. People think that they're done, that they're polished, that there is the best they could be. And there's tons of low hanging fruit. Um and yeah, you don't have so many people working on these open source projects. It's not impossible to make them faster and make them better. Um, yeah. So my management style would be less of a meant. I I wouldn't be as good of a mentor today as I've been in the past. Um, and then books. Uh, books aren't really focused on management or focused more on startup stuff. So I love the lean startup play bigger. I'm obsessing lately about the concept of intelligence. I'm reading a lot of books about that. Um, SAPIENs and storytelling is something I'm super excited about. So So I guess to kind of step back a little bit, I would I would have very little appetite to mentor someone on a technical perspective. But I have a lot of appetite to mentor someone on a storytelling perspective. So if I know what, you're gonna go given important talk. Are you gonna go meet with some VIPs or executives. Then I would have appetite to mentor on. This is how you're gonna get your talk. This is the opener. This is why you do it. Um, but pair programming with someone to make them better. I have zero appetite to do that in my life right now.

What indicators are used to track performance in a job like yours? Think of the indicators such as key performance indicators (KPIs), objectives & key results (OKRs), or so on.

Summarized By: Jeff Musk on Mon Sep 14 2020
Yeah. So for me, um s o for this quarter. I'm tracked on the caliber of guests that we bring on. And I'm also tracked on the amount of engagement that comes off these research projects. So for these more exotic research projects that I'm working on, if they land flat and if they don't generate the leads from the projects or the engagement and that's a that's a sign of failure. So KPs were really attached to podcast guess quality. And then, for I'm gonna be Forbes contributor. I've got these other pieces they're rolling out. I'm tracked on the engagement that comes from those stories which can lead to Leeds and closing deals.

What are the recent developments in your domain? How significant are these improvements over past work? What are their implications for future research & industry applications, if any?

Summarized By: Jeff Musk on Mon Sep 14 2020
There's been a lot of developments, especially since 2014 when deep learning kind of took off and started making waves. We've seen rapid increases in accuracy. Model size keeps getting larger. Um, extra PT three. It's 175 billion neurons, the text generation version that came out of opening I, and that's great. I think what we're coming up against is going to be. I want to be surprised in the next 2 to 5 years. We see some significant rework on deep learning because deep learning has some fundamental flaws. So Hinton came out with Capsule networks. I have not seen anyone in production whose production eyes capsule networks, and the thing with deep learning right now is I feel like it's left the reservation as faras um, how the brain works. So if you look at how a child learns a language or how humans learn, it's very different than how deep learning learns today. And so I think there's I think there's going to be some big breakthroughs in the next couple of years. The one of the things I like to say is the work of a principal consultant today is the free intern tomorrow. And so for our startup, we were building deep learning models that could combine different data sets. Whether you want to combine video audio image if you want to predict the price of a house and combined text and images and structured data into a single model, we were taking care of ways to handle that, where before a data scientist would have to build it from scratch. And so I think I think the next 10 years you're going to see a lot of the model development, Um, kind of become less block and tackle where you're pounding everything out in code. You can even imagine a scenario where you're having conversations. So you're talking to an AI system? Um, and I think that's the thing that's so exciting about this Domain is we've grown up seeing all these science fiction movies, and the science fiction is coming faster than ever. And so 10 years from now, 20 years from now, 30 years from now you having a conversation in your home talking to your smart home about an AI project or curiosity or something you want to accomplish, I think you could do a lot with just your voice, where today everything I'm doing, it's all I have to program every line of it because it's new research.

What skills and qualities does your team look for while hiring? What kind of questions does your team typically ask from candidates?

Summarized By: Jeff Musk on Mon Sep 14 2020
at higher view. I just joined in a robot, so I don't really have much of a team right now, But I'll be building a team, Um, in the next couple of years. So at higher view, one of the questions we like to ask Waas explained deep wanting to your grandmother. And so, um, what we find is in the AI machine learning space, you have people that they're very technical. But if you can explain something to an executive or to a lay audience, the argument is you actually don't understand it well enough and and one of the big things that we struggle with the data science is a lack of communication eso for people that I would be looking for. I'd be very excited about people that had previous speaking experience people. That s so this is a very special skill set. Do I do I feel comfortable sending you to a Fortune 50 company by yourself to present to 70 people about ai where they're gonna hammer you with questions. And so I'm looking for someone with experience. I'm also looking for someone Credibility. Um, we talk a lot about moving from normal to curious too passionate, too obsessed. And so the more people are obsessed, the better because they consume the content, they're gonna wake up early. They're gonna work on things on the weekend. Um, so yeah, so I think now things would be a little bit more objective. Would be looking for How many followers do you have? How many talks have you given? Can I review some of your talks? Can I watch your speaking style? I might ask questions about storytelling, But then on the research side, um, and this is actually a bigger topic because one of bigger topics this interviews are not a very good way to assess someone's talent. It's actually two. It's a it's a very short amount of time. And it's really not fair for a candidate. Because if I interview someone for 30 minutes about a I machine learning, what if I miss their core competencies? And what if I What if I over index on something that is a soft spot for their background? So in the past, I thought it would be really fun to ask someone, um, and open ended questions say, Tell me everything you know about least squares, regression impressed May. And and for some candidates that might be very overwhelming because for them to kind of get to the bottom of, you know, to know that's derived with matrix calculus and to know that it's an example of a convex analytical solution to know, like all the history, like like to show me how much they know about something that's so basic. What if What if they're an expert in something else and I ask you a question like that, That's not very fair to them to kind of give them a, you know, a black eye During the interview when I should have asked them about natural language processing, I should have asked them about something else. Um, so we talked about breadth over depth. It's better for you to have more breath. Um, full stack is always preferred. If you can show that you could build a doctor application. Um, yeah, it'll be really interesting to say if when I began to build up my team what that interview looks like because I'd love to just posted online that these air. These are the five questions, and if you can answer these questions, I want to talk to you and put time on my calendar to see if you're a good fit

Can you discuss career accomplishment(s) that you feel good about? Please discuss the problem context, your solution, and the impact you made.

Summarized By: Jeff Musk on Mon Sep 14 2020
higher view, so I they don't have any patents. When I joined, I can't remember the exact number. I think we delivered six or seven patents. One patent was approved within six months, going from submission to full approval. We also got a I integrated into a product. So if anyone that's done that, that is a rare thing. Um, we had to build, um, Emma lops things that could catch model drift. So models drifting in production. That's not something we knew we had to build. We had to react to it. The data science team that I built, a lot of them are still there. Um, I've I've been able to build a really good network with speaking and then selling a company. I think that's Ah, that's accomplishment that I'm very proud of to build a company that was getting so right before our acquisition. We were getting invites to go speak to companies like Red Bull, Goldman Sachs, three U. S. Government's Basics, where we have no paid advertisement, no paid advertising. But we're making a big enough of a splash that the those types of companies air reaching out for us, asking if we'll come present to them, um, with our deep learning. So, um, I think it's 2020 and 2018. I was number one in the world for a Iong Cora. Um, with some of these articles that went viral, I don't really participate in court anymore. Um, getting a model in production that outlives you is something that anyone should be proud of. Unfortunately, it's becoming easier because a lot of models die when someone leaves company.

What is a future career path for professionals in your role? How long does it typically take to advance through various roles? How easy are such promotions to come by?

Summarized By: Jeff Musk on Mon Sep 14 2020
So you have your standard resume and we could have a whole talk about resumes, I think resumes air Interesting. There's a lot of college kids don't know how to ride resume. It's no fault of their own. They just haven't haven't had that experience. But they also haven't seen it from the hiring side. And they don't know the reality that your resume is gonna be looked at for 10 to 20 seconds, and there's a really good chance it's gonna go in the no pile and figuring out how to stand out. Eso you have your resume and you try to You want to show promotion? You want to know. You want to show that you're progressing, so because that will get attention from other employers. So are you going from a junior data science role to a senior to principle to achieve title eso? You have that kind of career progress. Some people would jump companies to get a title enhancement, So if you can't get it in your current company because of politics or timing, um, it can look bad if you jump companies a lot. So if you're jumping companies in less than a year, or less like, ideally, state a company for two years. If you can't get the title promotion that you want, um, find opportunities and other company, you will learn a lot more changing jobs into another company. And the thing I warn people about is, you don't wanna be the smartest person in the room. And if you're if you work with a company for 10 years, you may become that. And if you become the smartest person in the room at the company you've worked out for 10 years, then that's unfortunate for you. But if you are changing jobs every couple of years, you move into a new industry. So for me, I went from semiconductor to, um to the financial sector. Then I went to HR Um, and then I went into a I. And so changing careers forces you to learn to learn again. And that's the beauty of college college you're learning to learn, so students should not be upset if they take a class on. I think it's not applied their learning toe learn and and you're gonna spend the rest of your career learning. So for me, I think a lot more about my entrepreneur resume. So selling a company and paying investors back is a good thing. Um, And then being successful in the new role is a good thing. Um, so second time repeat founders could raise a lot more capital with better terms. And so for me in my career, I could stay a robot for a very, very long time. I'm I'm happy here. Um, there could be a scenario after my er now in the future in the next 3 to 5 years where maybe I go raise again and go raise much more. A lot more capital to g o try the next thing and and so I see a lot of value doing that. If you want to grow wealth, it's very scary to go to a startup. But the fastest way to accumulate wealth is to take risks and new startups

How did you set the scope for your minimal viable product? How did you get to product-market fit? How did your product evolve over time?

Based on experience at: co-founder, Chief AI Officer, Zeff
Summarized By: Jeff Musk on Mon Sep 14 2020
So this is bringing up some fun memories because I remember talking to VCs and I think is a is the first time founder. You could bring a lot of arrogance. So you talk to VCs, and if they don't know the difference between you and Watson, you can leave with a bad tasting, real thinking. The VCs are incompetent or they don't understand a I. But the VCs deal with hundreds of startups that you know they'd either succeed or they fail and so VC to talk about product market fit. And I love this concept of product market fit because it sounds so simple. But it's the impossible problem that everyone fights with. And for a new founder, the the easy thing for me to say is your first product idea is wrong. So it doesn't matter how excited you are. Doesn't matter how secretive you are about your new product. It's wrong. And so the fat. Just a soon as you get feedback from your users, the better. And so some of the feedback we started get early on is we had engineers telling us that they would like to include structured data in our models and as data scientists. We would have never considered that. But this is this concept of getting credit. So if you know that the car is a BMW, and if you can't tell from the image, you need to include it. And so listening to the customer voice, um, I think there's eso building an M V P. There's a lot of tension on this topic. And so listen, reading the lane startup students might be really surprised, and you look out stories like Qualtrics and other other successful companies. There is a huge amount of value and not building a product early on. So if I'm gonna go meet with a potential buyer, I could show vaporware. I could show a slide deck. I could walk them through something that doesn't exist. I could even pay. I could even invested money into it into some designers to make it look really good. It's not a product, and really, what you're trying to do is get their buying. And technical startups are very dangerous because the founders can fall in love with the tech, give me a techno file, and so I I've talked to other people that are getting ready to start ups and their geek ing out about something that has no market. They're speaking out about something that's super technical and there's no market for it. And and and so for me, for so for us, we we built a p. I s initially we didn't have ah, gooey or anything for then to interact with. We got models into production. We did kind of play that role of consultant. Um, you start out as an AI consultant and then you try to roll it into a product, and then we ended it with it being a product with a P. I s where people could upload data that could build models, get insights, deploy those models. Um, but I love this topic. I think the next time I do a startup if if it is a SAS company, I would love to raise enough capital to not worry about burn for two years, So hire the right team and then based with a network that I have that's growing, I would want to essentially partner up with a lot of VPs of engineering and really understand their problems that they deal with and what the opportunities are but I wanna make sure I was solving a big problem and by a big problem. If you succeed, it's worth tens of millions of dollars to your customer, maybe even hundreds of millions of dollars to your customer. And there are problems like that. And if you could do that, the new charging 10% of that price is very reasonable. So to go after a million dollar year SAS license or a $10 million you're SAS license sounds insane. What it means. You're finding a problem with solving, and so a lot of a lot of companies jump into early. And so I think there's you talk about going from A to Z, building a product and delivering it to market. You should really go the other way around. You should find a problem. We're solving where people pay a lot of money and customers that we talked about customers rolling into customer. So if every customer looks different, I would be rude and say you're consulting in your entrepreneur and you're playing house. And in not only in the thing that's more upsetting about that is if you look at your hourly rate, you're consulting at a discount. So every time you pitch to a customer, if you have to make a new sales deck, you have a big problem. So find a verdict in NBC's Talk About This. Find one vertical and so saying you have a platform will scare a lot of VCs away, which is really interesting because data robots started with a platform pitch and they're delivering on the platform patch. But I'd love to see students try to make a platform pitch today, he sees. That's a very scary thing because the total market issues find one thing. Find three customers. You don't have to write a single line of code. Just go find three customers. One thing. Show them a power point. If you've done that, you're much further ahead than most startup founders.think about experience and life experience. So for the podcast, some of the people that I interview are really impressive. They're scary smart. They've sold three startups. Um, you know, they're they're on their third stint is the CEO, and some of their answers are very simple, but their profound there uncommon. And you have to unpack their answer to, say, 99% of people don't do that or they don't say it that way. And so um, yeah, I love the founder journey because first time founders are incredibly naive and myself included. They're incredibly naive in the market is a very hostile territory like the market. The market doesn't care if you have cancer. Market doesn't care if you're going through divorce market doesn't care what's going on with your kids. It will burn you to the ground if it if you're not finding value. But at the same time, it also doesn't care about your background. So if you're a college dropout or high school, drop out. If you have the right product market fit, that doesn't care. It'll it'll. You know it'll rocket you to the top, and so that's kind of this. It really is the American dream in my mind that, um, and the biggest thing I think people can gain from his find mentors. We had some fantastic advisers for startup people that had sold companies that raise a lot of capital. And, um, one of them we would call Call him our startup Jesus. So any time you're having a bad day, you could call him up and he'll just laugh atyou. And he's had it 10 times worse and he'll give you advice. That's profound. And one of the other things we talk about being a startup founder is things that are hard preparing for things that are harder. Yes, I remember the first time we were, um, you know, potentially losing a $200,000 contract. It was very, very hard, very, very stressful. But that prepares you in the future for things that are even more upsetting. And so if you look bigger companies, they're not that stressed about losing a six figure contract. They're worried about losing seven and eight figure contracts and their, um and so I think there's definitely something there, and and that is constant challenge for a founder. You have to protect your mental health because there's no there's no boss over you making sure that you go to dinner for the family or you don't work on the weekend and so that that's this constant nightmare Founders is you can screw everything else up so

Who were your early users? What marketing channels, approaches, and marketing tools did you use to contact users? What worked and what didn't?

Based on experience at: co-founder, Chief AI Officer, Zeff
Summarized By: Jeff Musk on Mon Sep 14 2020
early users. They were data scientists, people that we knew early on. Um, when it comes to finding early users, we talk about at bats. So how many at bats can you get? So if you can talkto 10 people and one is interested, can you talk to 100 people where seven or interested? Can you Can you work the numbers? And, um, I I got a lot of value through my network. So seven years ago, I would have said Lincoln was not that big of a deal, But we had a lot of leads that came in through LinkedIn where I could send them an invite. We we also got really good leads with presenting, So I'd go present at a large company and this gets back to the evangelism piece. So if you did a really good job with evangelism peace and had a lot of people in the room, it was an opportunity to show a level of competency where then they wanted to talk to you about their strategic ai problems. Um, one of the issues with a eyes. We had a lot of inbound requests where they haven't ai problem. We don't support that was very common. So I'd say for every inbound requests, probably 44 out of the five times we were not a good fit. We didn't support that use case. We never did paid advertising. Um, we has more. Just about hustling in our network. What else do we We did some, um we had some executive leads through block posts, things like that. Um, but that could be really tricky because you can look online and see founders that air writing medium articles or Forbes contributors. I I love the idea of just having if you could grow a really big network, then you can just you can essentially sell to your friends. You can say, Look, this is the start up. I want to dio This is the value we're gonna provide. I'm not going to charge you anything and let's see if we could get to value and and that's a great place to be because it's low risk. But for early founders, they don't have the reputation. They don't have the network they can't do. That s early users. There's a lot to be said about a visionary so working in big companies we would talk to one department and then we go fight, talk to another department and someone who's a VP of innovation. They're much more likely to deal with warts on your product. They're much more likely to geek out about the potential of your product. They're more likely to use your product. They're more likely to go against the grain and potentially buy your product. And in a company, you talk about this idea of an advocate and even a super advocate. So when you're selling a product, you need to find an advocate. An advocate is someone they're gonna go to bat for you. They're gonna be honest with you. A super advocate is your friend so super advocate will actually text you and say, This deal is going off the rails. You guys aren't gonna close the deal and they're not being a dick. They're actually just being really, really nice. So super advocate will do everything they can to help you land the deal. And so that's that goes back to having good emotional quote like good emotional intelligence. You The sooner you can establish advocates and super advocates, the sooner you can kind of face the mirror and understand that your product sucks or, you know they're not gonna buy for these reasons. Thea Other thing I became very sensitive to. It's so good to blow up pricing during the first meeting. And so some people get scared of that and they say, Well, I can't go to this company and say our product costs $200,000 to start because we can't That's not long enough to give them the value Prop. And one of the issues with a I is so many projects are not worth doing. Um, you can have a buyer where there's a casual curiosity. And so this definitely came up with our startup, where we spent 567 meetings with very, very big names, some very big, impressive vanity names. And then we realized towards the end that this was not worth hundreds of thousands of dollars to them. It was if you put in a basket, it was if you took the PSC and times it by 10. And so I think there's you really have to find a problem with solving. Do a POC, um, reduce the time. If you're doing a 30 to 60 day POC paid or Peter free. You want that one poc to justify your entire contract of value. And if that's easy to find in the first meeting, If you can't find the first meeting, then don't have a second meeting. Um, but I think sometimes the startup you're chasing squirrels and in in vanity brands can be dangerous. Vanity brands can kill you. So you trying to go after a Fortune 100 company as a small company? If you don't have, you know, x number employees. That's a very dangerous thing that can kill your company because they're gonna demand too much stuff you're gonna be. You'll be too distracted by the possibility that you could have this vanity brand on your website your first year and that z the wrong approach. They take longer to close. Um, yeah, kind of a scattered response. But most of the contacts where I do a lot of speaking globally and in a lot of our leads came from people that saw me speak, or people that I went out to dinner with. We had some leads come from in video and dell, where they would they would introduce us to these leads because these hardware companies, they don't have software to sell. And so if you have compelling software, there's opportunities for you to ride along and go co pitch with in video and l. So we did that as well.