Goals vs Tactics with a16z's Marc Andreessen at Disrupt SF

ladies and gentlemen please welcome back your disrupt MC and TechCrunch senior writer Jordan crook fellow staff members and Sam I assume Sam here today my main man Sam he's over me uh you guys know that smell that hits the air just before it's about to rain you guys know what I'm talking about well if you're smelling something a little different right now it's because a tweet storm is a-brewin okay Marc Andreessen is in the house we're going to bring them out to talk with Matthew pans Reno please welcome andreessen horowitz is Marc Andreessen and Matthew Panzer you know nailed it all right welcome thank you happy to have you here we've got a packed house it's almost like people want to see you let's see what we got I remember when this conference was 14 people and my Carrington's dog my guess is the people are still here the dog might be in the back somewhere let's see let's see um so some some folks in audience might not know about the the a 16 D model model completely so you explain your brands workshop so this is very interesting to me so you bring companies in portfolio companies in and you bring fortune 500 companies in and you get them together what's that all about yeah so we the way our firm works so the philosophy is that we very much try to in as many cases as we can help founders develop into being fantastic CEOs and what we discovered is one of the as we sort of thought about how that goes we realized early on that one of the reasons that founders get replaced by professional CEOs early on a lot of companies is because professional CEOs you hire professional CEO he walks in the door he or she walks in the door they've been in the industry for 20-25 years in a lot of cases they have all these relationships they've got this giant network which is just really hard for a founder who spent you know a lot of their time probably in a cubicle writing code before they started the company to really catch up too and so a lot of what we do in our firm is to try to basically bridge that gap and we we think about it as giving the founder superpowers to help them be great CEOs a lot faster and so we've got a whole variety of things we do on that one of the big things we do is what is what you mentioned which is we call it market development and so think about it as sales and business development on behalf of all of our startups and so the the theory is that for every new startup whether you're doing self-driving cars or AI or whatever it is there are some set of big companies that are going to really matter right big companies big organizations big institutions by the way it might be the department defense as another example so not just companies but it's really difficult and complex to figure out how to go deal with the DoD for the first time how to go deal with Ford Motor Company for the first time or Bank of America for the first time please with labyrinthine yeah like Byzantine kind of pathways to the the person you need to talk to yeah these big companies these big companies the first deal you know set of an hour meeting with a big company and if you're there for the first time the first half hour is the 25 people from that company introducing themselves to each other right and then it may be the meeting will eventually get to the point we get to talk about your startup and so it's just there's a there's an impedance mismatch if you're an electrical engineer and it just doesn't work well but big companies big organizations I mean you just you just heard it from secretary Carter big organizations really want to understand what's happening in technology and I think actually now more than ever before and so we basically do you might think of it as basically highly organized speed-dating and so literally the way our firm works is we see two thousand startups a year upstairs right raising money we see two thousand big companies a year downstairs in our briefing center and then we basically do speed dating speed matching between the startups the big companies and then and then a basic I spend all day running upstairs and downstairs back and forth between those those two meetings which does bring up an interesting you know kind of question as the firm grows which obviously it's grown significantly over the past seven eight years how do you scale personally how do you scale your time how do you portion that so that you're giving you know companies the the most that you personally can give them without obviously destroying as oh yeah so a lot of that is actually in how our how our organization is scaling so we view our firm as a company even more actually than a partnership and so just the effort I talked about just in that part its toilets is on the order of I think now 25 full-time professionals just on that team so they do that all day every day they orchestrated all they run all the briefings they do all the follow-up they help the founders understand how to do all this stuff and so you know if you just to get you know gouge eneral partner like me it's just incredibly highly leveraged because I partner with them every day and then what our companies experience is you get all the benefits of the theoretical or real benefits of a GP but you also get this machine one of our founders Jeff will described it as being funded by us as like plugging into the matrix and this is what he was referring to which is there's this there's this engine that's just going and when you plug into it all of a sudden you have all these sales opportunities and business opportunities investment opportunities and PR opportunities and policy opportunity all these things that would normally take you a long time to develop they just start happening and then as GPS we work to help the founders develop in their own skill set and then we help make sure that they can for leverage what the firm is doing so recently The Wall Street Journal published an article about your firm's returns and compared them to other large firms which have been around much longer do you feel that the article was fair and evaluating how its performing how your firm is performing so first the main response to that article was people wanting to put more money in our funds which is a challenge because we just raised our new fund in the spring and so we're closed and we can't take it and so my my big request is next time if they could run the article four months earlier that would actually be super helpful that's one part of the answer the serious part of the answer is it's the article actually gets this very interesting topic in in venture capital and Finance called Marx and it's the kind of thing that people don't usually talk about but the article surfaced a very interesting topic and so a mark is a value that you attribute to an asset that can't be bought or sold and so we learned in the financial crisis in 2008 that this is a dangerous concept because you may recall from that time a precipitator to the financial crisis was all these big investment banks had all these mortgage-backed securities and all these things in their balance sheets and they were valuing them by Marx and then it turned out that they weren't actually worth what people thought they were conversely in venture capital you see you do see some of that but you also see some of the opposite which is a couple years ago we had marked in our report Serapis we have marked slack all the way to zero right and I don't know what slack will ultimately be worth but I do know what to be worth zero do you think it has some upside off of that and of course that's because Stewart and why did you do that well because so so interesting quite so when we funded slack because startups are unpredictable and because amazing things happen with amazing people so we funded slack it was a slack it was a company called tiny speck and it was building a game and then the game didn't work and then this is where they mate you know the amazing thing happened is Stewart and his brilliant colleagues at the company reformed the company that same company they reformed it into into what became slack very similar by the way to how I've Williams you know ten years ago reformed OTO into into Twitter and so one of the things when things are right in venture at all is because you figure out how to get in business with somebody like Stuart that can do really amazing things like that and it's just like we do our damnedest to try to forecast try to pick which people can do that but it's an unpredictable process sometimes it works and sometimes it doesn't and so the way our investors view it is the marks they don't respect a lot of time we're required by law to report the marks but they don't put a lot time and focus on that they focus a lot more in the fundamentals so the companies were investing in the final thing I'd say which is I think really relevant to more broadly what's happening in the world right now is that when I entered the industry in the early 90s before the dot-com boom the one thing that was known for a fact by everybody including the press commentators analysts executives everybody politicians they all knew for a fact was the Japanese Japanese tech conglomerates were going to completely take over the tech industry and America was going to have no role to play in tech and it was literally going to be we were literally I thought in college I thought I was gonna have to learn Japanese to even work in the tech industry and people forget how dominant that theme was at that point but it was it was it was incredible at that time how deeply that was believed and the reason that was believed is because the theory was that Japanese companies because of how the Japanese system works could make long-term investments right they could make an investment for they could play out over five or ten or fifteen or twenty years in these really big ambitious projects and America in contrast was believed that like no no we run in this quarterly you know CEO mentality had that very the long-term carry to make those hat yeah exactly and American companies didn't then I view what we all do every day what we do in our job but also what all of us do every day is really trying to disprove that and really trying to prove the idea of building for the long-term building important things over a long period of time is alive and well in the world alive and well in the US and also alive and well in the world and I would argue that most of the world has actually gotten even more short-term than it was 25 years ago right stocks turnover even faster now than I used to investors in the public market are even more concerned about quarterly results so venture capital and startups and high tech are one of the very few areas in the world where you can make a bet for 10 years or 15 years or 20 years and that's what we're trying to do I mean that that what's interesting because it leads in this concept I talked to my writers about it sometimes and I'm wondering if this is something that comes up when you talk to your companies this concept of the fog of war right like you're you're in the trenches you're doing your thing you're looking at that next quarterly you know mark internally obviously it's a private company once you go public it's open everybody to see and the pressure gets even worse but how do you get beyond that fog of war how do you see what the long-term questions or long-term solutions you know to those questions are answers those questions are so when you're in advising companies how do you say like this duality of crank it and like do your job but also relax and look ahead that's right so how do you do that to a company yeah so a lot of this is actually you can actually borrow concepts actually from military strategy on this which is sort of the difference between sort of strategy and tactics right at a difference between goals and tactics so I think it's incredibly important to have a really vivid clear idea about where you want to get in the long run that you stick to and that you're very solid on and that everybody agrees to and then I think you want to be very very flexible in the tactics and I think the problem is you it's a dichotomy right it's a it's a contradiction so you have to you have to think about terms you have to think it from a long-term standpoint but also you have to think in terms of day-to-day tactics and this is where you know I've been very critical in the past to this idea of like fail fast because like I always things like fail fast is like the key word in there is not fast it's fail and failure sucks and success is awesome and we should be trying to succeed not fail and I think the fail fast thing is people thinking because I think fail fast makes a lot of sense on tactics if the tactics Network doesn't work find a different tactic I think fail fast is catastrophic if it's applied to strategy and if it's applied to goals and I think a lot of founders frankly even still like talk themselves out of what are going to be good ideas in the long run because they're not getting immediate traction and so again it goes back to long term like we just ruled by data and what cited by their gut that's actually so that is actually a racing point so that is one of the things that happens which is in the old day in the old days when I when I was running short pants you just didn't hat you didn't have all the data it was much harder to get a sense of how well you were doing and so like on the one hand you you you felt less concrete connection to what you're doing but on the other hand you didn't have this cascade of data coming at you and now is you know like in all of these businesses today you have daily weekly if you want you have data to the minute to the second to the microsecond of how well or poorly you're doing and it's really easy to get distracted by that by that short term data and it's really easy to draw a long term Lucian's based on short-term data and you do see companies that have gotten in real trouble over that mm-hmm conversely you see companies you know that worked for a very long time on something that people think is just completely nutty and they get heavily criticized along the way and by the way sometimes those work and sometimes they don't but when they do work that is how you do something like for all this whole thing about how everything's supposed to be speeding up it still takes a decade or more to build something really significant I mean it you know in this world like it still really does so when it comes to like interesting products that seem to be sort of jerked around by their own data or were for a long time let's talk about Twitter so much Twitter oh okay I've heard of that yes so you tweet a lot i tweet I tweet a lot too but you get more guff for it so you recently tweeted never read the unfiltered tweets which I agree with never ever ever now and more people now are starting to get the filter yes right they didn't have a filter so you think Twitter has done enough to like curb this sort of barrage of like hate or just ancillary noise or abuse on their platform well so on the one hand no clearly not which I think a lot of people experience every day on the other hand I do think I think Twitter is a very interesting test case for a question that's going to get more and more important as the world evolves so the good news is you can see it the presidential election is here the good news is we're finding out what everybody thinks yeah right and so the dream I remember like you know Time magazine ten years ago the big cover was like you know the future is you like we're going to find out what you think we're going to see all your videos we're going to see all your great ideas all your memes one of our presidential candidates could be insane and we would never know but right now we have the chance to maybe know you know there's this concept there's an edge which I mean there's this concept so there's this argument actually of what's happening in the culture right now which I think has has a clay shirky talks about this a lot at NYU which is there's this concept in communication sociology called the Overton Window and it's basically this idea that in any given time there are a set of things that are allowed people are allowed to say out loud in public and then there are a set of things that people are just aren't allowed to say I mean they can say that if they want but they're going to look at ostracized I'll get laughed at they'll be they'll be score and they'll be consequences to saying these things and so you know one argument one argument like about this presidential cycle is that in the past you know people didn't have all these crazy you know opinion and have all these terrible beliefs and the internet has now changed people right and now people believe all these crazy things because of the internet the other argument is what the Internet has done and Twitter has been a big part of this is actually is broad in the Overton Window so Twitter has opened up the opportunity for people to be able to say things that they already believed that they said behind closed doors to their friends but that they could never say in public right and the they is any random individual voter it turns out the Veii also as any presidential candidate right right and so that that's a theory that is more less of Internet has changed us it's more that the Internet has actually really shown what people think and this gets to you know really fundamental philosophical more frightening option about human nature well that's that's one argument is that's more frightening you know I'm a little bit I'm you know I'm not like a strict libertarian I'm libertarian enough though to believe that I think we should get it on the table like I like I think there's a lot that's worth talking about what's happening right now but I think one of the good things is if people really think things I think we should get them on the table like I don't know that it's better to live in a world in which there's large numbers of people who feel disenfranchised because they don't they feel like they don't have a voice like I don't know that that was so great and by the way lots of bad things happen in the past so it's not like that things just started happening and so I you know like I'm not thrilled by what's happening right now but I think over time I think we're adjusting to a world in which we are really going to know what everybody thinks and I think you know if we're gonna make good decisions like we're gonna make good decisions in a company you want to know everybody thinks we're gonna make good decisions as a society we should probably start by knowing what everybody thinks for for better for worse we're finding out yeah that's for sure what is your block list look like what's my what your block list oh I don't I'm I'm I'm a very was at like trigger finger fast chicken feather feather trigger different yeah okay well I mean that's fair I mean I think that I think almost everybody has decided or wanted to at least want it to block somebody well and it goes back to you again Twitter's representative of okay so then so then you're Pro free speech right and of course everybody is Pro free speech in theory up until whatever they just decide is bad speech and so Twitter is interesting because they had declared themselves at one point right the free speech wing of the free speech party and you know that brings with the consequences you know people have I think equally reasonable critiques of platforms that are maybe to lock down and to closed right and so you know is there a healthy middle maybe I don't know I think in the next 10 years I think what a lot of these services are going to have to try to figure out is what is the right point in the middle yeah yeah I mean I think we're seeing this with Facebook is trying to figure it out you know with this they obviously just issued an apology and about this you know very famous photograph that they block from the platform because it had it involved nudity I've argued on on Twitter about Facebook which everybody does that they need an editor-in-chief that they need somebody to own at least the philosophy of a media company I know Mark has said it's not a media company but what do you think about Facebook needing something at least a filter that involves humans instead of just algorithms or interesting on that so I can't I can't comment on a director of Facebook and so let's write I have to know freedom is face when it comes to Facebook yeah you got it let's broaden it to humans versus outcomes right yeah so any company facing a similar issue um that says hey we're gonna trust in the algorithm we believe that we can tune it and tune into it until it gets right do you think that we're ever going to get to that point or as a human gonna need to be involved to make that judgment call so I think judgment you know judgment is one of those things like you know with wheat when we get the algorithm that can exercise judgment I'm all in favor I hope I get to fund it you know judgment is a is a human is a very fundamental human activity in a lot of ways for a lot of things I think the same by the way the same question you're posing is true or the other side which is organizations that have relied entirely on people right are increasingly grappling to try to figure out how to deal with data right now to deal with software and so I do think there's a there's sort of a mirror image set of if you know if the internet services have one challenge over here for example the traditional kind of classic human driven media on the other side has that challenge coming from the other direction and I just think like one of the sort of as an observer I just wanted things I think is like these issues are only going to intensify and get larger and more important in the years ahead like I we're gonna look back in 20 years and say like these are the early rounds that abate like this stuff is going to get much more overall all these topics are getting much more controversial more angsty yes tenner because you know the media sphere is what we all live in like we're all we're all becoming interconnected and that that has magical aspects to it but it does mean that we are all you know for is that great line in the in the in the Watchmen when Rorschach gets sent to prison and they're all there the criminals are all excited because I going to get a beat up the superhero and he's like you're completely wrong it's like I'm not stuck here with you you're stuck in here with me yeah and there's a little you know that's one of the big one of the big things that I think a lot of a lot of people are gonna have to really think through in the next 10 or 20 years so that brings I mean that brings us to sort of when you're handing off sort of components of your company to systems we talk about AI right and so you've mentioned obviously a is at the center of a lot of your investments now or most if not all so what do you do you think there are opportunities for thin I thin AI companies still that work with small teams small datasets on vertical problems or is it own completely by the Giants now or will be yeah so if we were sitting here two years ago I would have said I was very depressed on this topic because it appeared that there were four or five big tech companies that had basically hoovered up all the AI talent and there was only a small a couple thousand people total in the world who really understood how to build AI and the big company said Hoover them all up we're all paying them a ton of money right it's a ton of actual current cash income which makes it very hard to recruit them out to startups and then we're making these big long-term investments in AI resulting in things like Alexa and echo which you know then you know just miracle amazing products that Alexa and echo are the result Jeff Jeff says I think it's 1,500 people over four years like I mean that's a scale of effort that I kind of can't match and then the other thing that would have depressed me is I would have said the big companies have a data advantage because so much of what makes a I work now when it works is you get all this data and so what makes the self-driving car work is you able to train you're able to gather all this data of all these cars in the road at scale and then you're able to train AI on millions and millions of miles driven and the theory went that if you were a start-up you'd never be able to match even if you could get the people you couldn't get the data today I think we've seen an almost 180-degree reversal in all those all those topics and we're now seeing a wave of really high-quality AI startups and I think several things have happened one is a lot of the sharp people at a bunch of those big companies have decided like people who came before them in other areas of tech that they can actually start their own company and so we're seeing people spin out a lot of those companies start their own startup also the technology has become more widely available and so for example Google diameter but this afternoon Google's Google released tensorflow is open source is a significant kind of opening of the component of how this stuff works and put it in the hands of a lot more people we objection sexual interest we have a company a Udacity that's here they actually well now you can go online and you can learn google caliber AI literally as an online course and so the the education the technology has made it possible for a lot more people to come up to speed on this tech and then it turns out the data also it turns out the data there are so some cases we need big data but startups either finding very clever ways to get the data or there are new AI techniques that can actually do really interesting things on smaller data sets and so one example is you can now train like if you're doing it self-driving car if you're doing a self piloting drone you can train them in the real world or you can actually train them in simulation environments and there was just a presser for somebody's now training self-driving cars AI and in Grand Theft Auto which you know you probably want to watch that one a little bit more closely it makes some of the others it might to that dataset it might develop some bad habits running people over on the sidewalk but there is this idea and actually turns out the video game engines companies like unity it turns out I think are going to play a pretty big role here because you can create a video game engine and then train the AI basically in the video game yeah and so we're seeing a bunch of really interesting companies in areas ranging from health care right life sciences we're seeing you know we're seeing lots of maybe if a FinTech we're seeing lots of areas where people are not bringing in AI transportation we're seeing a wave of transportation startups out you know auto was one recent ones that spun out of Google and else immediately got bought by uber but you know there will be more of those companies exactly like that you got George you know you got George George is one of our companies George Hotz with his company comma will be on stage later and literally George built himself is a self-driving car in his garage I kind of typifies that example of the other tools are there brilliant founder but not like really no all right brilliant one person right with the toolset and the technology and the ability to get the data so speaking of IDEO games I hear you're big gamer I love it what are you playing now what do you playing right now you know so I try to take about two weeks over the holidays and I try to get all caught up so I'm behind so I'm looking forward to overwatch I'm looking forward to was the sky game okay no man sky yeah yeah yeah no infinite were infant warfare not yet you should I got a gonna be that nothing ya know I'm behind no okay oh yeah I usually get whatever is the latest Call of Duty all right let me know if you want to raid in destiny we'll hook up and decorate all right so one last question and that is if you were going to revise your software is eating the world quote for 2016 now that we've had a little bit of a gap obviously as you mentioned tons of shifts how would you do that what would you revise that so hopefully it's going to bear up pretty well I think it I think it is I basically made three I made three claims in there I said busy every every product will become a software product and so everything that's not been a software product is going to become a software product and I like for example I didn't anticipate I think in five years ago I didn't anticipate that would be true as much as it is for example now if the car right where the car is but you know self-driving cars primarily it's like 80 or 90 percent software what makes that thing work and so every product becomes to suffer product because of that every company becomes a software company and I think there's a lot more evidence of that now right you've got these big companies you know GE and Ford these really big important American companies now building big presences in the Bay Area and doing lots of acquisitions excuse me all around software and then third the the sort of provocative thesis at the end was in the long run the best software company will win each market right and that that the best software company it might be a new it might be a new upstart it might be an incumbent right but it'll be whoever in the long run makes the best software right that's still the most provocative statement of all and there's a lot of big establish encompass a lot of industries that would get very mad if I said that out loud when they're in the room but then they would also say how much they're doing to become great software companies right and so whatever ask you to if you've got a start-up handy for example it may be it may be that we have a start-up if they could buy or partner with and so maybe we can help so that's still the most provocative part of the thesis that I think will take probably multiple decades to play out but I like you just take transportation like the revolution and transportation is happening right now and the fact that it's happening so much of it is centered here in the Bay Area and here in California and so much of it is centered around chips and software and AI and data that it's become you know transportation became something that other people far away built that was big and complicated and mechanical and has become something that is now tractable for a Silicon Valley startup right a fundamental country that seed jobs quote about look at the cabinet and saying hey I know how to I can build this cabinet in the world around you you can build it yeah and so automotive you know I'm building a whole self-driving car seems unfathomable until you have the tools all the Senate is yep great thanks mark good thanks everybody uh-oh everything okay that's some Karma for going five minutes over

Channel: TechCrunch

Views: 42,968

Rating: 4.8189654 out of 5

Keywords:

Id: GfZ41h086N4

Length: 25min 35sec (1535 seconds)

Published: Tue Sep 13 2016