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 Wed Oct 14 2020
I've got a long story on. Let's let's see if we get through it very quickly. I was born in Cape Town in South Africa. Uh, and I grew up in London. You could tell from my accent that from London. Although, um, my my people originated from India from Southeast Asia. Um, and at that time, the tip of Africa was a melting part of a lot of, uh, people, people from around the world. You might also recall that it was also a period off the off apartheid in South Africa on so our family decided to move to London. Um, I we're schooled in London with the traditional British education. Um, got a degree physics at King's College on, then proceeded with a A career in It started off in the electron ICS, but moved onto software engineering very quickly. I focused on computer video, digital video, digital media, digital audio. Let's see, I'm worked with a start up, and the term startup was not known then or commonly used. But it was a small group of developers that were creating a novel product that allowed you to edit a movie on a computer. The product was called light Works. That product sold very well in Hollywood. And on the strength of that success, I moved my family to Los Angeles where I am now. And so my career path proceeded along that line. I'm now cto off Rheal Networks after a number of startup tours of duty and experiences, um, 11 of which I was a startup that we sold to Google. Um, and, uh, eso Now, I'm, uh, the CTO off a public corporation. Riel networks. You probably know riel networks. Well, real notes is the company that invented streaming media many, many years ago, 25 years ago. So that's where I am. This is how I got here.

What responsibilities and decisions does one handle in a job like yours? What are the top three priorities? What are weekly work hours like?

Summarized By: Jeff Musk on Wed Oct 14 2020
well, briefly, um, had the CTO position in in a number of different companies, but I've come to realize that it's really very simple. The job of the CEO is to convert capital into technology on these air technologies that further the strategic aims off the off the company. Um so all of the decisions are typically around investment into technology. The CTO role is not the same as the VP of engineering role, so it's not managing the day to day engineering activities. But it's mawr off a strategic role. Looking at, um, technology trends, technology, investment opportunities, opportunities to modernize, opportunities to invest in a product roadmap. Um so the top three priorities one is to to align the technology roadmap with the strategic business aims of the company, um, the other is to is to form alliances on bake. Sure, that theologian mint between the technology, um, goals and ambitions align well with the, uh, the business. The business decisions that have been made s Oh, this is, uh, it zone important role, because typically the CEO eyes the big spender or his decisions on recommendations um, entail the big spending in the company and that sometimes runs at odds with the fiscal motives of the company. In order to extract maximum profit eso you have to form alliances and be able to maintain the level of credibility in order for you to be able to do your job going forward. What are the working weeks? What's the working week like? Well, it Z if you to do the job well, you're You're putting in a lot of hours, but it's not. It's not ours that are typically on the clock. Um, so you find yourself thinking about the problems, Um, and the opportunities, Uh, not outside of the 9 to 5 period on. I think all of those hours count towards your working week, but it's not a it's not a 9 to 5 job.

What are major challenges and pain points in a job like yours? What approaches are effective in overcoming them? Discussing examples will help students learn better.

Summarized By: Jeff Musk on Wed Oct 14 2020
Yeah, it's It's easy. It's easy to make predictions about the future about what the future technology trends are going to be. Um, you can open up any tech magazine, and what is often correct is what technology is going to come in the future. What is often wrong and mawr often than not wrong, is the timing. And so it's easy for a tech journalists, Let's say I know in the 19 thirties to predict that, um, flying cars are going to be invented at some point. Yes, that's probably true, but their timing was way off. Um, and so one should one should, um, investors much time in figuring out the timing off technology trends as well as what those trends are. Um, it's easy to become reactive and look at where the thing is, where the in American sports parlance is whether hockey puck is now not where it's going. Um, but it's very easy for, um, uh, it's very easy to make Make predictions that are way off in the future, and you commit your company and your business on it on that course, and your timing is way off that the market is you end up actually achieving a technical advance. But it's way ahead of the market. On DSO. You end up not being the beneficiary off that success. You you've basically a philanthropist to the industry. You've created a technical achievement way in advance of the market to monetize it. So that's that is a real pain point. It's a thing I struggle with a lot. Um, there's another pain point that's very relevant for today, and especially in the area that I'm focusing on in a. I is to make sure that the the business decisions, um on the product decisions, the technical decisions align with your ethical principles. So this is Ah, um, is not a commonly held belief there, many in the industry who say no, keep ethics, ethical considerations out off business. And leave that to the politicians. The CEO of Coin Base famously says that if there's no place for for politics in our company, or there is no place of talking about ethics in our company, way were business people and the business motive would be first, um, I think that it's important that theme ethics come first, and here's why we, uh, creating technologies that have massive impact on society, and there's a potential for them to have a negative impact on society. And if we as thinking educated professionals do not anticipate what those negative consequences can be, um, nobody else is going to do it. There's a very good chance the business folks are not gonna turn revenue away if they are given an opportunity to sell a product I'm doing. And I think fire or a foreign unethical use case. I'm working in on technology in the area off facial recognition. A very scary a very skerry product, a scary technology if it's misused. So part off. The challenge on the pain point is to not only create the technology but create the ethical framework in which you run the program and make sure the technology is, um, secured in a way that you can police the ethical, use off the technology and be able to revoke access to the technology. If you detect any unethical use off the technology, these are important things, uh, stun times. Sometimes new technologies either get over hyped by the press or fear is invoked, depending on what the technology and depending on how many more clicks they could get or how many more impressions they could get from their article. Fear sells better than hype Uh, so often you'll get for brand new technologies like I'd knows, stem cells or crisper or even e commerce. You'll get the fear story, the story about why it is dangerous. Um, and you have to you have to blend these two. You have to look at the the hype as well as the fear and plot. Of course, in between on defined that that rational course, you can't say I'm just a nen gin ear. I'm just a technologist. I'm going thio, develop my technology and then let the business people let society deal with it. Afterwards, Um, there was a quote from Verner von Braun, Verner von Braun's the rocket scientist for the Nazis. Um, when? Later on, after the war, he was employed by the US government, he said, Well, it was just my job to make the Rockets go up where they landed with somebody else's problem. Well, that is, Ah, cowardly defense on. We don't support that on bond. It's We can do better than that. Ah, we can. We must do better than that now, more than ever. Oh, that's my response.

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

Summarized By: Jeff Musk on Wed Oct 14 2020
if you If you get to the position of CTO, you've probably experienced Um ah, full stack off development away from programming right down at the hardware level, in my case, writing micro code for processes assembler C C plus plus then higher level languages. On top of that, um, I I don't think anyone language is a must have, but you should have a full understanding off the entire stack all the way from the hard way. You should have some appreciation off it away up to the very, very sort of high level Web programming languages or even, um, but the constructs in the cloud like land, uh, frameworks for doing your cloud processing. But the the tools that have transformed our industry uh, there was a big transformation that happened just in this last decade. It's when Google published the tensorflow tool set, um, a I waas a mix off models and then tools and companies never really understood where their value add waas Some some thought their value add was creating better tools than others, and others focused on the model. Google just close that conversation down and said No, the value add is the model here are the tools free of charge. And now, um, many other competing frameworks. Uh, pytorch Cafe. But there's a whole list of them, um, you these the tools that we're currently using in the area of a I, um I have to say that they're not the not the end game for a i ai is in its infancy at the moment. Um, I wouldn't say it's really intelligent, I would say I mean, it's called artificial intelligence, but the stage we're at at the moment is we can build models that allow us toe ask questions off data based on historical experience. This is not like a human that's able thio come into a new situation on bond. Um, make decisions based on, uh, a completely novel new environment that they have not seen. The level of intuition on bond predictive, um, capability. Human brains have is still way in advance off where Karen, um, neural network models are the deep neural networks and eso eso. It's important to be aware off the state of the art off tools, but also understand where they are in the evolution cycle off off the industry. And this isn't the end point. These tools have a purpose sometimes, Um, because you're you've mastered a tool. Let's say you've mastered the hammer. Every problem looks like a nail to you, and it's not. That's not the case. Let me give you an example. So we're really good at facial recognition. I get point a camera at my face. It can recognize who I am because we were able to extract a unique signature from my face. And the compare that with a database of a million signatures. And the one with the shortest Euclidean distance from my signature is a match for me. Okay, so we're good at that. What else could we do by looking at a face? Can you look at the face on? Determine? There's Zodiac sign their astrological sign? No, that seems ridiculously unscientific. There's no brown truth for that, okay? And you look at her face and determine their credit score. Hmm, That's interesting. You know that there probably is a database you can train a model, you can have the faces, and you can have here of their credit scores on. Maybe you can create a predictive model stops and businesses would be interested in that But it's not a cool thing to do. It's not an ethical thing. Okay, what about looking at a person's face in determining their religion? Do you think that is possible? Thio do that? Um, maybe. Maybe so. The Chinese government commissioned such a program, and as a result, a million wigger wigger men are in reeducation camps in the northern part of China. Um, a completely flawed hypothesis that you could determine ah, person's religion by looking at their face. What about their criminality? Can you look at the face and determined criminality? Can you look at her face and determine their sexual orientation? So all of these things are, um, uh, ways of using tools, but using them in completely inappropriate ways. And be careful, be careful of not off thinking, you know, because you've mastered one tool. It is It's a solution to every problem. It's not. And so, uh, that's my caution. Yeah,

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 Wed Oct 14 2020
so management style in a technical software, um, context. Um, eat, I you may recall. Or maybe you're too young to recall the era off case tools. There was this, um, this period where the industry, the software industry, thought that computer aided software engineering waas a thing. There were such things as case tools, which were basically ways of drawing and diagramming software. Before you write a line of code and allowing you to do this collaboratively with teams, they communicate design architectures and make a reasonable high level, uh, design and allowed you to manage and delegate tasks. And there were methodologies called There was the Yordan methodology of data flow diagrams. And then there was a competing methodology that came along, which was the the UML diagrams, where and then eventually they went through a period of standardization and kind of merged the two concepts, um and then and that was ah, hope that you could, in a top down way, decompose a product from the top down so you could end up with the smallest components. And you could delegate that off to different team members and manage it like that on. There was typically a scheduling activity where there would be a massive Gant chart on the gang chart would be managed by one project manager who would, um, paste up this big a zero page on the wall. Or maybe he's taped together several sheets of paper to, um, toe layout this gang chart, which basically predicts when the end of the program is going to be and none of that worked. There's just a catalog of failures off that type of top down thinking. Andi. Anyone that peg their bonus on the endpoint of that gang chart was was often disappointed. They never met that bonus. They know he got that, um, and so that many. There's many, um, books that analyze why that type of management off technology projects fails and industry swung completely over to the to the other side and said, Well, we don't know when our projects going to finish, Um, but we know the direction we're heading in. Why don't we set sail in that direction and then well, when we reach the horizon, will look again and see what obstacles ahead of will will reset course. And so the agile methodology says we don't know when the project's going to end, we'll move it forward. Sprint by Sprint will make a sprint short on. Then we'll readjust. And that is not good for the business guys because they want to know when is the product gonna be ready to ship? And so the management style is Mawr. My management style has evolved around, I guess, in the same way that the industry's management style is evolved. It used to be very top down authoritarians, um, and prescriptive and managing a big Gant chart. And now it's become mawr or agile. I think, though the pendulum is going to swing some way into the middle, where, um, there's gonna be a balance between being able to make long term projections and saying, Yes, we are going to ship this product by this date by because there's gonna be the Christmas sales or the back to school event. We need to get the product for this date on bond that but still allow the flexibility off being able to make changes as we dio um, so the influential books and management style well, the classic books, the mythical Man month, uh, is A is A is a well known textbook. It's worth looking at that it talks about the failures. Off off prediction. Um, Steve McConnell, the guy who wrote Code Complete, also writes a Siris of books on management. And, uh, I've done agile training and there is a There's a whole bunch of books on agile. Uh, I don't have ah preference, but here's here's my recommendation is that you stay open minded. You understand that, um, these these management styles, project management styles or methodologies is a methodology du jour, or at least for at least for a generation, and understand where it is in the cycle off how the industry is moved on. Be critical of these things and you should be prepared to, um, make adaptations. This is not a religion. Agile is not a religion on DSO. Don't treat it like that. You should be able to make your own, uh, modifications to the methodology to make it work for yourself. Um, often religions are in place to impose the will off. Smart people on on dumb people on is to make the populace follow a set off rituals to improve society. I mean, that's the intent eso any of these methodologies. If you treat them like a religion. You become that dumb person on you abdicate your brain. You abdicate your your own thinking on do you should never do that. So, um, yeah, I've rattled on enough about management style, but I've I've answered it Maurin away about, uh, project management methodology, okay?

How can one get better recognition of work from one's boss and higher management? What mistakes should one avoid? Stories or examples will be quite helpful.

Summarized By: Jeff Musk on Wed Oct 14 2020
ultimately, you have to deliver value. Um, so don't do things for show. Um, there has to be. There has to be value that you deliver. Ah, in everything that you do. So the many of these things are the invisible things go up on time. So be timely. Yeah, many of the students that they're gonna be there going into their first jobs. And so don't be overly needy. Onda, come over as needed to to your boss. What you have to do is quietly and consistently deliver value, Show up on time. Be professional. Be respectful. These are these air normal things that you have to do anyway. Get up in the morning, get dressed, make your bed. Have those rituals on your own even before you get to work. Do those things. If you if you're if you're slovenly and unprofessional in your personal life that's going to reflect. There's no magic switch that gets switched on by the when you walk in the door and you're Hey, I'm suddenly professional. So it's if if your goal is to be successful in a corporate culture, you have to or 11 successful recipe is too internalize all of the needs and the functions and the and the sort of culture off behavior in yourself. So you you have to you you're not a child anymore. Um, you're you're now moving from the sort of student culture, which is a bit relaxed and laissez faire. And now you're going into the professional world where you're being paid a salary every month, or or every week or whatever, and you value is expected from you. And if it's a career in technical development, uh, or any anything in the high tech area, there's a level of professionalism on and rigor that is expected. Of course, if you're an actor or if you're a creative person, maybe a different culture is expected. But in the area off the Koreas that result from stem sciences, there's a There's a level off regular that's expected. Don't do that thing first, be rigorous about your day, maybe the weekend you can give yourself, you know, a couple of days off, but monday to Friday, get up on time, set yourself a schedule, get up on time, make your bed dress professionally sharp to work on time and sharp prepared. Think about what your day is supposed to be or what you're supposed to get accomplished and prepare for it. Um, recognition from your boss should come. Your boss is looking, uh, because he's just spent money on hiring you. Um, so you're one short cut is to figure out what your bosses goals are. What has he promised his boss? Yeah, and that's a shortcut. So you figure out how you can help your boss look good to his boss. That is that that cuts through a with the other, um, recipes or strategies for success. Now that that might be a that might, you might not have, um, access to knowing what your bosses goals are, um, but subtly on And, um, in a professional way, you should You should ask that you should ask. Okay. What is it you want for the department to be successful? What have you promised? The your boss is how can we make you successful? How can we make you look good? And that is a very proactive way of doing it rather than hunting in the dark and trying out different behaviors to gain attention. Attention seeking is a very, uh, but very distasteful trait. in in the employees? Uh, yeah. So don't be. Don't be asking for feedback every five minutes. That's you could be a great achiever, but if you come over is overly needy than that is, uh, not good. Um, what other? Um, Miss takes, um id, right? Trying to write everything down. You're not gonna remember everything that you're told, and you probably don't have a lot of coverage time with your with your superiors. So you should write it down and digest and review and reflect on it on and, um, get a cross reference with peers eso that you can understand what's needed of you and how to how to succeed.

What skills and qualities do you look for while hiring? What kind of questions do you typically ask from candidates?

Summarized By: Jeff Musk on Wed Oct 14 2020
So when I'm hiring, I'm typically hiring for the layer below me. You know, I very rarely look at their education. Um, I've often hired people with no college degree. I know. In some cases, that is, Ah, um, that is not a a common practice in industry. Often when when often the resumes that reach me have already been filtered by the HR department. So I don't get to see everybody the way these thes things happens different in different companies. But typically, a job description is created on defy. If I am the hiring manager, I would create that job description. I typically take a template on guy massage. It, um, the HR department manages the search on surfaces up candidates, and we decided which of them will make make make their way through an interview. Okay, so let's assume we've got some good candidates that come for the interviews. UmI'm I'm looking for a candidate that's gonna ask me questions as well. I typically ask them to give me an account of their career. Um, why am I doing that? I'm looking for relevant experience for the job that I expected them. Um, I probe into, um, the issues and the difficulties they faced in positions in order to see if they talk negatively about their former employers. That is Ah, that's a big red flag to me. Um, when a candidate says, Yeah, but they were a bunch of idiots, and they didn't listen to me. And so, uh eh, So I left. That is, that is not a good trait, Andi. It's a sign that this candidate would not be a good fit and not a positive contribution. More likely a, um ah, hub for discontent on bond. It's not not the sort of culture that you want to promote, um, often for technical roles, for for software. All you you you set software problems on dso. Um, these problems are not, uh, ones that, you know, you'd score out of 10 and then Okay, off you go. I set these problems, Um, so that we can have a discussion about, uh, the computer science issues around that, or the issues of a computer language or a processor or, uh, ways off doing things better. So, um, that I think many candidates in the area of technology should should expect toe face a technical, a set of technical questions. So I think that's common. And often it helps engage a technical candidate in a conversation, but also helps um, figuring out whether they can think on their feet. Um, I don't like the ones where you you just get the questions. And let's say the questions were run by the HR department. And then you get given a score of this candidate scored seven out of 10. That's not that's not useful to me. Um, I like actually conducting the questions live maybe on a whiteboard on bond. Get to understand their thinking process

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 Wed Oct 14 2020
Okay, Let us let us tackle two of them. The first one I mentioned early on. So the the product the product was called light works. This is the, uh this is the the edit controller. Um, that used to be connected. Vyron old nine pin serial port to a PC. Anyway, on the PC ran the software that allowed you to edit a movie. Um, it was it was one of the first systems that allowed you to edit a feature length movie on a computer. Um uh, dated back to the early nineties. Um, it was difficult to do it at the time, because in those days, the processors were maybe until 46 processor running at 33 megahertz. Um, the operating system, this was Windows had just been brought out. But we do these applications still on the dots operating system with an extended. So any graphical user interface you had to dual yourself. So you have to create a window manager in addition to thio running. In addition to your application, eso all of these things were big. Uh, it was a big problem in the early nineties to create a system that, um needed to digitized video digitized audio, store it on a computer file system, um, in such a way that you could stream it in real time and make editing decisions, the graphical user interface and then output and edit decision list eso that the original film could be cut. To do that in the early nineties was a huge leap forward on I Worked With a a great set of engineers that are Eljero asses in their own right and the product one. A technical Oscar on a technical Emmy Award. And in fact, it's still being used to edit movies today. The the movie The Irishman was edited on light Works on Light Works is still going concerns. I'm very proud of that. And it, uh, that success launched my career. Uh, from a fairly mediocre, uh, progression to something that took a jump. It it enabled me Thio Thio, leave the UK come to the US and take a position at another corporation here. But it was on the strength of the experience and light works, So I feel proud of that. Andi. Uh yep,

What college programs did you attend and what were their best parts? How did each of your college programs prepare you for your career?

Summarized By: Jeff Musk on Wed Oct 14 2020
in in the UK I was fortunate that in those days the government paid for your, um, paid your entire tuition and even gave you a stipend. Yes, if if you if you were good enough in high school to on your high school grades in those days, it was own levels and a levels in the UK If you've got high enough, uh, grades, then you've got accepted into college. You know, I didn't apply. I didn't apply to do physics. I actually Well, the truth is, my mother wanted me to be a doctor that they all. And I got an interview at King's College to study medicine. But the worst thing you can do in an interview, if you if you are shooing candidates, you know, if I just answered their questions, normally I probably would have, uh, got into study medicine. Um, they asked me. So tell me the final question. Why tell us why you want to be a doctor. And I paused for a little too long. I said to them, Well, you know, actually, I can't stand sick people. And that is apparently a big no, no. Well, someone who wants to be a doctor on. I talk to them about all the things that I thought were exciting about medicine, about the pathology, about biology, about medical physics and so on. And then they said, Hmm. Perhaps you want Thio apply for something else anyway, So I ended up doing physics Kings instead, Um, and never became a doctor. And I think if I had a doctor would have been a very unhappy doctor or specializing to something else. Um, computer science was a big part of the physics degree. Many of the simulation models, um, for thermodynamics and nuclear physics and so on were, um we're, uh, sort of legacy programs that are written in FORTRAN. That every generation of students every year of students would come along that would would would work and would tweak event and make improvements on. So, um, in those days, the the computer science, um, department on the physics department would would hold joint lectures on bond. We had access to the mainframe computers as they were, but but there were also microcomputers. The pet lab off your pet was a early microcomputer make a personal computer. So the course was specialized in physics. There were no, um, separate miners, the particular course. And because the government's footing the bill you had to do your bachelor's degree in three years, there was no opportunity to do General Ed in the US, bachelors degrees are typically four years long because you spend your first year doing General Ed deciding what you're gonna major in now. It was, you know, business. It was very business, like governments putting the bill. If you if your if your grades dropped in any year, you get kicked out, and then that's it. So there was a pressure to, um uh huh. Go through all the courses and Onda take away the units in a very compressed time. Um, what was it like? It was It was not a It was not a pleasant experience. It was more like work. It was commuting in with with ALS, the London commuter crowd, with the proletariat on public transport going into the center of London Kickings college. At the end of the day, you didn't spend time with your professors and have tea with the, you know, the chancellor or whatever everybody needed to get home and get on the tube train going home. So it was a very much a 9 to 5. Um, schedule. Um, but I I studied in the department where a generation before Rosalind Franklin had done als the x ray crystallography work on, um, on DNA. And so Rosalind Franklin never got the Nobel prize because the crystallography data was, um, was shown to Watson and Crick. And over a weekend, they guessed the structure off that DNA. They didn't dual the hard work anyway, so I came to a physics department that was kind of still licking their wounds from not having got the Nobel Prize for for DNA a generation later. But I got to be, uh, my professors were some of the renowned professors from that era. One of them was Alex Stokes. Eso I still remember his lectures very well, Andi. Even though my career did not pursue, I did not pursue your career in physics. I still ponder on the problems off physics that were. We were taught in in that college in that BSC course bachelors of science Course. It's important, though, that, um you when when you it is important for me is to continue the work after leaving college even though you're doing your, um, your earning a living, a career that's now in a in a different space I found that I needed to continue maintaining my, my, my, my studies and my thinking in the area of physics. So it stayed with me. It wasn't wasted.