Web2 Interview With John and John

This was a lot of fun. I got to share the stage with John Doerr, someone I've admired and respected for many years, and John Heilemann, who never ceases to entertain and educate me.

Here's the video. It's pretty long. Just shy of forty minutes.

#VC & Technology

Comments (Archived):

  1. David Semeria

    That was lucky. This appears to be the only session not available on the Web 2.0 Summit video page.

  2. NearMine

    thanks! very informative and enjoyable. Totally agree with the over valuations, being in the UK we do not have this yet, although they are not really encouraging innovation here either.

  3. Michael Langer

    40 minutes in middle of the day? Gotta be busy executing right now………maybe later

    1. Matt A. Myers

      I’m writing 4 emails simultaneously at this very moment — and writing this comment. I have 10 hands.

  4. Rick Bullotta

    Hi, Fred. I was in attendance and it was one of the best sessions at the summit thus far. I really enjoyed the way the two of you “jousted respectfully”. I must say that the silicon valley arrogance was on full display. Granted, they’ve earned the right to be a bit cocky, but it was a bit over the top given the clear trend of growth of innovation outside the valley (full disclosure: we’re based in Philadelphia, so fellow east coasters). You were actually quite successful in getting John Doerr to back off of his “the valley as the center of innovation” position throughout the course of the chat. I was also delighted to hear your comments on the proliferation of clones and copycats getting funded – agree completely. Again, nice job, and hope to have the opportunity to chat in person someday.

  5. Tom Labus

    Thanks, Fred. This is great.Reading Tim Wu’s Master Switch, The Rise and Fall of Information Empires. The First AT&T controlled radio in the 1920s. Their radios could only receive T frequencies and programming. This is a lot like APPL today.

  6. Steve Hallock

    Hi Fred,Thanks for the link. I was looking for a great excuse to take 40 min off from real work 😉 A few thoughts on the points brought up:First, boom vs bubble. I think it is clear to anyone paying attention that we are well before the peak in this new environment. Great companies are yet to be born, some good companies now will become great, and some good/great companies now will get crushed in the process. Accordingly, money is flowing in. No one wants to miss the next big thing and everyone knows that a few of them are just around the bend. So whether it is a boom or a bubble depends on one’s positioning. If your money didn’t get into one of the successful companies, it’s going to feel like a bubble. If it did, you correctly played the boom.On the NY vs Valley thing, I think you hit the nail on the head. One of the most powerful drivers of Web 2.0-3.0 is that straight tech is no longer the driving force. Now it is all about the intersection of tech and people’s lives. Silicon Valley is not exactly where you go to learn about problems faced in the real world. NY, LA and other world cities have talented people with experience in non-technical fields that can use these new platforms to create great companies for real people.Lastly, on the Google and Facebook not internally innovating point, I did not fully understand if you think this is a problem or not? I think the point you make is fair, but maybe that is just what these companies have become: major platforms whose core business is to innovate and improve just enough to remain dominant, only creating truly innovative products by luck or synergistic acquisition. I’m not so sure there is anything wrong with that.Lastly, if John is MJ, are you Durant or Lebron?Best,Steve

    1. Gorilla44

      Fred is Kobe. Marc Andreesen is LeBron.

      1. Guillermo Ramos Venturatis.com

        Fred Wilson Webb vs John Doerr Wilkins… http://www.youtube.com/watc…Thanks to both. Just amazing!

      2. Evan

        if I were Fred, I wouldn’t want to be called overrated like that.

    2. Guillermo Ramos Venturatis.com

      “on the Google and Facebook not internally innovating point” I think both cases are very different.Google internal innovation is huge and it is brilliantly mixed with selective M&A activity in fields were they think outside talent is leading. The balance is a good one.Facebook internal innovation & M&A activity is totally uncorrelated with their 600 MM users. The problem is that they have an incredible platform to build very profitable businesses on top of it, but should the platform is not strong enough, whatever is on top is at risk. The big question could be: is the platform stronger because of whatever is built on top of it? Up until now, I think that´s not the case.

    3. ShanaC

      I think you are largely right about this is the beginning and not the peak. Although i can’t hack it’s use (just another thing to do…)I admire greatly people like Mike Yavonditte at Hashable because they are at least insisting that our social gestures have much more meaning to them and therefore should be captured.I wish we had better understandings of how some of these social gestures work- it’s probably the new age of social media, rather than connecting us through “new platforms” and creating behaviors, it adapts to us and our cultural norms we exist in.

    4. fredwilson

      if those are my choices, give me kevini love the way he plays the game

      1. Evan

        Durant is awesome, and scarily, he just keeps improving.

        1. perpetuity

          … and we get to watch Durant play against the Celtics tonight. Go Celtics!

  7. theschnaz

    I really appreciate you publicly stating your views; Apple/Facebook in this interview and HotPotato/Drop.io in the NY Observer (http://www.observer.com/201…It’s rare for people to state their views, especially when they’re (to some extent) contrarian.Keep it up!

  8. Evan

    I wouldn’t be fulfilling my duty if I didn’t point out that John Doerr is a Rice grad.

    1. Donna Brewington White

      I’m not a Rice grad myself but I gave you a “like” for constancy. Some cool people out of Rice, it seems.

      1. Evan

        I love it! Unfortunately Rice still needs every bit of marketing that it can get. Maybe I should also use it to point out that Doerr got both a BS and an MS there? 🙂

        1. Dave Pinsen

          Rice U. should buy Rice to Riches and keep running it, sort of like how ING Direct ran its espresso bar.

  9. Darwinification

    I’m a big fan of John Doerr, he always has some really interesting / intelligent things to say: http://mutate.it/2468

  10. Alex Murphy

    Great interview, great content. Here are my favorite quotes:- Ideas are easy, Execution is Everything, John.- Zynga is the only successful thing built on top of the Facebook platform, Fred.- Innovation without execution is hallucination, Colin Powell (from John)- APIs are the new Biz Dev, Catarina Fake (from Fred)

  11. Jason White

    FredYou deserve thanks from the NYC tech community for your performance here. I idolize John Doerr and have nothing but respect for him and KPCB. But you displayed authority and critical thought here in a way that could only prop up NYC in the minds of investors and entrepreneurs who would question NYC’s legitimacy as a tech hub.Moreover, controversial or not, it’s refreshing (and a valuable academic exercise IMO) that you were willing to publicly challenge companies like Facebook, Google and Apple/Iphone at a time when few of the industry’s louder voices would dare take any kind of dissenting position on their longer term prospects for global domination.Anyway, thank you and here’s looking forward to seeing more public appearances,Jason

  12. Dave Pinsen

    Off-topic, but I thought some of the locals here might find this new site of interest: We Are NY Tech. The site profiles one member of the NY start-up community per day, and lets you search for local founders, developers, designers, and also females involved in the local tech community.

  13. sigmaalgebra

    It’s fun to listen to Doerr, hear him say again “it’s not a bubble; it’s a boom”, and see him wave his hands and do a soft sell. And it’s good he’s no longer talking about his daughter being afraid the sky is falling due to evil humans and filthy CO2. Then when pay attention to some of what he is actually saying, he’s less fun to listen to.In the interview, Doerr made some huge mistakes. One of his worst mistakes was his”Ideas are easy. Execution is everything.”I don’t want to get too close to particular work or our work in information technology entrepreneurship so will step back to other evidence and examples.Okay, the Clay Mathematics Institute is offering a prize of $1 million to anyone who can settle the question P versus NP. All that is needed is an “idea”, and Doerr says that “Ideas are easy.”Showing that P = NP is true would likely be valuable in information technology entrepreneurship, and showing that it is false would make a lot of people feel better about a lot in information security. For more, there are several serious diseases without cures — cancer, heart disease, arthritis, and much more. Since “ideas are easy”, Doerr should do all of us a favor and publish his “easy” “ideas” for solutions.Of course, given a good idea, commonly people protect it as a trade secret, with the patent system, and, maybe, with a stamp of “Top Secret”. So, not everyone agrees that “ideas are easy”.The “ideas” of Turing at Bletchley Park? He helped break a German cryptographic code, and one result was much of Rommel’s supplies on the bottom of the Mediterranean and Rommel’s loss. The “ideas” of Commander Rochefort? He helped break a Japanese code, and one result was four Japanese aircraft carriers on the bottom of the Pacific near Midway and the turning point of the war in the Pacific. But to Doerr, such “ideas” were “easy” and ‘plentiful’.Also from Doerr we learn that the work of Szilard, Wigner, Fermi, and others on “ideas” was easy and “everything” was the “execution” of flying the Enola Gay. Money? Roughly each $3000 in the Manhattan Project saved the life of one soldier for a good ‘ROI’.More “easy” ideas are Kerberos, RSA public key encryption, Reed-Solomon error correcting codes (crucial to CDs and more), heap sort (meets the Gleason bound and guaranteed (n)ln(n)), AVL trees, balanced multi-way branched trees, extendible hashing, and much more.During WWII the US discovered the importance of good ideas and since then has moved to have Congress generously fund US research universities as a source. The funding does result in good ideas, but they are usually NOT easy. The funding helps the US do well in the Nobel prize awards in science. If the “ideas” were “easy”, then many other countries with much less funding would do much better than they have been.Instead, quite broadly, the US is just awash in excellent ability at execution: Review how quickly the Empire State Building was constructed. Likely we could do it faster now.Given a pill that provides a safe and effective cure for a major disease, Doerr is saying that part of “everything” that is difficult is the “execution” of putting the pills in the bottles and putting the labels on the bottles.As we’ve seen over and over in the best of information, biomedical, and military technology:Good ideas are difficult and rare. Given a good idea, execution is routine.And the US is awash in ability to do routine execution well.Of course, given a bad idea, “execution” is likely from difficult down to foolish and hopeless. From “execution is everything”, it appears that Doerr’s experience has been with not very good ideas.It is clear that Doerr doesn’t ‘get it’, that is, doesn’t see the difference between good ideas and ones much less good.We have to conclude that to Doerr an “idea” is not some new, powerful, valuable work, difficult to duplicate or equal, that becomes the crucial, core ‘secret sauce’ of a successful company. He just doesn’t think about such “ideas”. The NSF, NIH, DoD, Ph.D. committees, and editors of peer-reviewed journals do think of such ideas; biomedical venture capital may; Doerr and, yes, nearly all of information technology venture capital do not.In a biomedical analogy, to Doerr an “idea” is a catchy new label on bottles of old snake oil instead of a new, powerful anti-biotic. More specifically just now, to him an “idea” is some way to let young women use the Internet for gossip and finding dates.Of course, denigrating “ideas” might be to his short term advantage: He can preemptively stake out a ‘negotiating position’ that lets him tell any entrepreneur that (A) his work so far on his “idea” is not worth much because “ideas are easy” and there are “plenty” of them and that (B) the “execution is everything” and mostly hasn’t happened yet.Well, an entrepreneur can respond:Quite generally, the ‘problems’ of gossip, looking for dates, and ‘social chit-chat’ are of diminishing importance; we need to attack more important problems. One hundred years ago, no one got the lower 40 acres plowed standing at the fence chewing a cud, and the ‘productivity’ issue remains. For means of solving problems, we need more than routine software based only on intuitive and heuristic techniques; we need good ideas essentially as the US DoD has long understood.More specifically, why does an entrepreneur want someone who doesn’t know the difference between a good idea and a bad one on their Board and there able to make destructive decisions from ignorance about ideas? His money once in a bank looks like any other money, and his understanding of “ideas” and “execution” are misinformed, uninformed, just plain wrong, foolish, and dangerous.Yes, we could say that a ‘good idea’ was one that led to a successful business, and now the ideas of interest are ‘social’ uses of the Internet to get ‘engaged users’ and, thus, ‘eyeballs’ for ad revenue. The ‘technology’ is just routine software. The ‘defensibility’ is not in ‘secret sauce’ or even in the server farm but just that the users are ‘engaged’ and, maybe, create a ‘network effect’ — everyone wants to be on Facebook and use Twitter because everyone else is and does. Okay, if it leads to money in the bank. But the problem being solved is not very important, that is, the solution is not a “must have”. The ‘idea’ looks a lot like a fashion frock fad that can both come and go quickly. Evaluating ‘ideas’ as ‘fads’ is difficult.It seems to be good news that his being so misinformed about good ideas still lets him be “the Michael Jordan of venture capital”: People who do understand good ideas should be able to do well.It sounds like Doerr has much more than money enough and has been drifting away from reality.

    1. Guillermo Ramos Venturatis.com

      The context was Web 2.0 conference. The lean start-up is Web 2.0-3.0 makes John words totally valid. In this blog, I have seen Fred´s comments that are only related to internet software companies although not being explicit about it.There are two different type of VCs: software driven and non-software ones. KPCB is playing in both at the same time. Both are important, add value and make this world a bit better, but the dynamics are totally different: the first is based on execution and the second in ideas&execution. And, yes, we need both of them.Finally, all around the world, there are huge amounts of public R&D budgets producing tons of good ideas. The big problem is that the technology transfer from labs to the benefit of mankind works very poorly. In that context, I would agree with John´s comment “Innovation without execution is hallucination, (Colin Powel)”.

      1. sigmaalgebra

        You wrote:”The context was Web 2.0 conference. The lean start-up is Web 2.0-3.0 makes John words totally valid.”I’ve got good news for you: It is widely accepted and quite possible to do much better than what John said, including in the “context” of “The lean start-up” and “Web 2.0-3.0”.By doing “better” I mean more money in the bank, and that’s a case of “better” we can all believe in, right?Your main justification appears to be: “There are two different type of VCs: software driven and non-software ones.” and “Both are important, add value and make this world a bit better, but the dynamics are totally different: the first is based on execution and the second in ideas & execution.”So you are assuming that in the “context” of “The lean start-up” and “Web 2.0-3.0”, “ideas”, in particular from “R&D”, are not important. You are assuming that “software driven” means no significant role for “R&D”; that is, if the production work of the company is from running software, in a server farm or in a client device, then necessarily the software should be just routine with no role for “R&D”. So, mostly we just read from Stroustrup, right?No. There are many topics in research where the results could be, as I wrote, “some new, powerful, valuable work, difficult to duplicate or equal, that becomes the crucial, core ‘secret sauce’ of a successful company” including in “Web 2.0-3.0”.This really isn’t the place to include a lot of explanatory detail.

        1. Guillermo Ramos Venturatis.com

          it would be great if you can tell 2-3 examples of highly succesful Web 2.0-3.0 companies that are “difficult to duplicate” because of the sofware they have built. Maybe I´m missing something big…

          1. sigmaalgebra

            You wrote:”it would be great if you can tell 2-3 examples of highly successful Web 2.0-3.0 companies that are ‘difficult to duplicate’ because of the software they have built.”I don’t think that there are “2-3 examples”. Right: Now we can all LOL. Now that we’re done LOL, we can continue.If there were 2-3 examples, then that would not be so “great” because the next example would have a little less of an advantage!We need to continue further: In his post “Storm Clouds” athttp://www.avc.com/a_vc/201…and there in his reply athttp://www.avc.com/a_vc/201…Fred wrote:”web apps have little defensibility”large networks of engaged users have enormous defensibility”the move is to get from web/mobile app to large network”i’ve just given you our entire investment strategy”So you have also noticed Fred’s”web apps have little defensibility”.That has been the situation.Continuing, you wrote:”Maybe I´m missing something big…”Actually, apparently, yes:So how are we to evaluate my claim that an idea can be “some new, powerful, valuable work, difficult to duplicate or equal, that becomes the crucial, core ‘secret sauce’ of a successful company” including in Web 2.0?One way is what you suggested: Take the history of Web 2.0 and see what others did. Or, take empirical data and an empirical average and/or an empirical distribution.Well, actually such ‘data analysis’ has some powerful ‘best possible’ properties at separating the wheat of the good from the chaff of the bad if GIVEN two biggies: (1) We do need the empirical data and (2) that empirical data is all we have. Maybe here we have (1) but we fail miserably on (2). We have MUCH more data than just this empirical data. With the extra data, we can do MUCH better than just from the simple, old empirical data.We need an example: How to evaluate what the Wright brothers were attempting, before they had yet tried to takeoff?If we used just the empirical data as you are suggesting, and noticed that well funded, well respected Langley had just gotten all wet falling into a river, we would LOL at the Wright brothers.But if we got more data, we might not LOL: If we had visited the back of their bicycle shop in Dayton, we would have found a little wooden box that had the world’s first, good wind tunnel. They were able to calculate wing lift and drag and propeller horsepower and thrust. Yes, they did miss Reynolds number, but still they were close enough. And they understood the crucial importance of three axis control for stability and had a good enough solution there, also.So, given their wind tunnel data and their corresponding calculations and given their three axis control, their chances were quite good. I believe that they knew that!They weren’t just guessing. That Langley and many others had failed was essentially irrelevant. And, uh, the key was not “execution” but the “ideas” of the wind tunnel, the calculations, and the three axis control. Sewing the linen or whatever onto the spruce or whatever was “execution” and just routine.All that additional data, or, if you will, information, was visible right there in their bicycle shop. So, with that additional data, the old empirical averages were irrelevant.Then we can see, for a large fraction of firsts, the simplistic empirical data from ‘the rear view mirror’ is irrelevant or nearly so.Uh, just looking in the rear view mirror this way and following a herd of sheep can lead to remarks like “sheep get sheered” if not fleeced.More generally, we’re trying to do something new. So, we’re not trying to plow the lower forty acres just like they’ve been plowed productively for the past 100 years.Still, we want to know, in advance, long before ‘traction’, that our “ideas” and plans are solid. How to do that? The Wright brothers provide one example. I insert here: Such “ideas” are rarely “easy”.For more the US DoD has provided many examples since about 1941. Commonly their advanced projects got funded just from “ideas” on paper and long before any “execution”.E.g., for GPS, it would be good to have accurate data on the gravitational field of the Earth. For that it would be good to have a satellite in orbit and that had no drag from the gasses up there. So, how to build a ‘drag-free’ satellite? For that we need an idea, one we can have confidence in, before, during, and after launch. So, we need an idea for a drag free satellite, and a way to evaluate that idea, early on, just on paper. The DoD did that. How’d they do that? Uh, as we’ve been told, “Ideas are easy”, so I won’t bother you with something so easy!So, how to get powerful ‘secret sauce’ for a Web 2.0 project and, just from a description on paper, be quite sure the idea is good, good enough to build a major company, say, worth over $50 billion, and where execution is routine? Yes, the evidence from some of the best sources in our civilization is that we can do this.Then it is fair to say that, yes,”Maybe I´m missing something big…”For what you are missing, below is a little more, that I drafted but omitted to reduce length:For a hypothetical project, consider exploiting the ‘social graph’. Okay, that is data. In the exploitation, we will have software manipulate that data, and the manipulations will necessarily be mathematically something, understood or not, powerful or not. For more powerful manipulations, some mathematics can be crucial. What mathematics? High school? College? Graduate school? Original? Emphasize the last two and not the first two. I’m not talking high school second year algebra here, guys.For a little more, suppose there is an answer from exploitation of the social graph and that we want to report to the users. Likely we can only estimate or approximate that answer. So, we should consider some techniques of estimation and approximation. Some of this material, even if not original, is deep enough to be known by only a small fraction of the mathematics professors.To be a little more clear, some of the mathematics is astoundingly general and also applies well to using available data about the social graph. Or we can take available data and find ‘new connections’ in the social graph.For a little more, in Web 2.0, beyond furthering gossip, we very much have another ‘data exploitation’ in mind: We want to get ad revenue. Here a crucial point is ‘ad targeting’. So, the question is, given the data, what ads should we display? We can believe that in principle there is an ad that will make us the most money. So we want to ‘estimate and/or approximate’ the selection of that best ad, and here we are essentially forced into some mathematics. If we are to do this work well, that is, that looks good on the way to the bank, then we better not rule out graduate school and original mathematics.For a project not hypothetical, could I say more, with theorems and proofs, with some original work based on some advanced prerequisites, written out with D. Knuth’s TeX? Modesty, shyness, and confidentiality constrain me. Might my work be only for ad targeting instead of more in Web 2.0? No.My conclusion is, in the future of computing, starting now, as we seek to ‘automate’ work, including using the Internet to help young women gossip, the present ubiquity of intuitive and heuristic techniques for manipulating the data will much less powerful than we want, and original mathematics based on advanced prerequisites will become more important than Moore’s law. Likely you read this first here.The crucial issue isn’t “Web 2.0” or gossip: The crucial issue is manipulating the available data to yield money in the bank, and for such manipulations appropriate mathematics, commonly quite advanced, makes a 900 pound gorilla look like a flea that can’t bite. Again, intuitive and heuristic techniques can’t hope to compete, eventually maybe not even in helping young women gossip.Look, appropriate parts of the US DoD have for decades known very clearly, and have strongly acted on, the power of mathematics; the mathematics of estimation, approximation, optimization, stochastic optimization, etc. have been warmly greeted and well exploited. I started my career in such work; like essentially everyone with a US technical Ph.D., the US DoD paid for nearly all of it; when my wife encountered delays in her Ph.D., we ran out of our budget and I took a part-time job to see us through, again paid for by the US DoD. Thank you!The Wright brothers, Fat Man, Little Boy, the SR-71, GPS, the F-22, and much more provide examples that we really can have good “ideas”, that we can evaluate well just on paper, ideas good enough to make execution routine. In particular, our key can be some appropriate, possibly in part original, likely advanced, mathematics.So, borrowing from the techniques the DoD long used with high success, we can be one of the first with such techniques and ideas for more powerful, valuable ‘secret sauce’ ways to manipulate data in Web 2.0, be quite confident quite early on just from work on paper, have ideas good enough to make execution routine, and go ahead and get a major company.Much of the challenge of this whole ‘context’ is being successful doing things that are new, and we need to take this challenge seriously.Gee, I thought that we all understood these things long ago.I do suspect that the important audience is missing — the LPs.

        2. paramendra

          You should start a blog. Blogs are good places to rant and rave.

          1. sigmaalgebra

            So, you have no substantive content and just want to attack with three ‘ad hominem’ remarks.

    2. paramendra

      Oh, my. Easy now.

  14. vruz

    Hallucinations are what other people think you’re seeing.Until they see your hallucinations realised.You know, like Twitter.

  15. awaldstein

    Fred…well done. Congrats.I just watched this with a glass (well two glasses) of Clos de la Roilette ’09 Fleurie Beaujolais. A great panel. A terrific glass of wine…organic and only $20 a bottle.

    1. fredwilson

      we need a buy button in disqus for comments like this!

      1. Jonathan Cardella

        Funny you should mention detecting products in blog comments because I’m working on a system and for a Valley stealth-mode startup that does just that… It has a lot of practical applications beyond just what you’ve contemplated here with the Beaujolais and I’m thinking more along the lines of underlining the product name and an overlay on mouse-over, as opposed to a “buy button”, but I like the way you’re thinking. Betcha I can beat Disqus to the punch. Moreover, how about a Disqus killer? I’ve seen some of the comments that get removed by moderation around here – “Canadian Drug Pharmacy” is your username, really? Should we add a buy button to that one to? Disqus has done a helluva job with the $500K check you wrote them, don’t get me wrong, but the opportunity is immense and someone is going to come along and take it. Mark my words.

        1. perpetuity

          Can’t ship alcohol across all state lines, alas! Sounds like a good idea but annoying to me. 🙂

        2. fredwilson

          why not work with them to do this?

  16. Michael Weiksner

    This is one of the most important videos I have seen in a long time. I loved your argument for the rise of entrepreneurship outside of the valley in general and in NYC in particular. I also loved that you were able to stake out controversial but profound positions (e.g., Android >> apple). Still, I think John had some persuasive points to me that differed from your view. Was there anything that he said that has caused you to reconsider your views on the arguments that you made? Put otherwise, what did you learn from this great interview?

    1. fredwilson

      that reasonable people can disagree

  17. Mark Essel

    I’d love to watch but with 60kbps @ $14/night in the Fairmont at San Jose, streaming video isn’t going to happen. If only I could funnel the free wifi downtown to my hotel, it’s 100 times faster but my Twitter ID would be up for grabs to any Firesheep users in the city :(.Weak net connectivity blows.Update, in the lobby I get 7mbps down and 800kbps up, woohoo camping out here all afternoon after a good hike.Update2: whoa. on my laptop I’m getting 20mbps up and down. This is what I need in my room!

  18. Cristian Joe

    lol Worshiping at the Temple of Ideas love it

  19. Jeremy Campbell

    Loved that debate, very enlightening and intelligent. If John is the Michael Jordan of VC, then what are you Fred? I saw you speak in Toronto a few weeks ago, and truly respect your place in the VC world. What you have to say is empowering to entrepreneurs like myself, so I thank you for sharing your knowledge and perspectives.I really hope we aren’t in another web/tech bubble, I hope you are wrong about that Fred and I’m sure you hope so too!I have ADD but this was time well spent watching this video!

  20. sigmaalgebra

    You wrote:”The context was Web 2.0 conference. The lean start-up is Web 2.0-3.0 makes John words totally valid.”I’ve got good news for you: It is widely accepted and quite possible to do much better than what John said, including in the “context” of “The lean start-up” and “Web 2.0-3.0”.By doing “better” I mean more money in the bank, and that’s a case of “better” we can all believe in, right?Your main justification appears to be: “There are two different type of VCs: software driven and non-software ones.” and “Both are important, add value and make this world a bit better, but the dynamics are totally different: the first is based on execution and the second in ideas & execution.”So you are assuming that in the “context” of “The lean start-up” and “Web 2.0-3.0”, “ideas”, in particular from “R&D”, are not important. You are assuming that “software driven” means no significant role for “R&D”; that is, if the production work of the company is from running software, in a server farm or in a client device, then necessarily the software should be just routine with no role for “R&D”. So, mostly we just read from Stroustrup, right?No. There are many topics in research where the results could be, as I wrote, “some new, powerful, valuable work, difficult to duplicate or equal, that becomes the crucial, core ‘secret sauce’ of a successful company” including in “Web 2.0-3.0”.This really isn’t the place to include a lot of explanatory detail.

  21. sigmaalgebra

    Needs to be deleted.

  22. sigmaalgebra

    You wrote:”it would be great if you can tell 2-3 examples of highly successful Web 2.0-3.0 companies that are ‘difficult to duplicate’ because of the software they have built.”I don’t think that there are “2-3 examples”. Right: Now we can all LOL. Now that we’re done LOL, we can continue.If there were 2-3 examples, then that would not be so “great” because the next example would have a little less of an advantage!We need to continue further: In his post “Storm Clouds” athttp://www.avc.com/a_vc/201…and there in his reply athttp://www.avc.com/a_vc/201…Fred wrote:”web apps have little defensibility”large networks of engaged users have enormous defensibility”the move is to get from web/mobile app to large network”i’ve just given you our entire investment strategy”So you have also noticed Fred’s”web apps have little defensibility”.That has been the situation.Continuing, you wrote:”Maybe I´m missing something big…”Actually, apparently, yes:So how are we to evaluate my claim that an idea can be “some new, powerful, valuable work, difficult to duplicate or equal, that becomes the crucial, core ‘secret sauce’ of a successful company” including in Web 2.0?One way is what you suggested: Take the history of Web 2.0 and see what others did. Or, take empirical data and an empirical average and/or an empirical distribution.Well, actually such ‘data analysis’ has some powerful ‘best possible’ properties at separating the wheat of the good from the chaff of the bad if GIVEN two biggies: (1) We do need the empirical data and (2) that empirical data is all we have. Maybe here we have (1) but we fail miserably on (2). We have MUCH more data than just this empirical data. With the extra data, we can do MUCH better than just from the simple, old empirical data.We need an example: How to evaluate what the Wright brothers were attempting, before they had yet tried to takeoff?If we used just the empirical data as you are suggesting, and noticed that well funded, well respected Langley had just gotten all wet falling into a river, we would LOL at the Wright brothers.But if we got more data, we might not LOL: If we had visited the back of their bicycle shop in Dayton, we would have found a little wooden box that had the world’s first, good wind tunnel. They were able to calculate wing lift and drag and propeller horsepower and thrust. Yes, they did miss Reynolds number, but still they were close enough. And they understood the crucial importance of three axis control for stability and had a good enough solution there, also.So, given their wind tunnel data and their corresponding calculations and given their three axis control, their chances were quite good. I believe that they knew that!They weren’t just guessing. That Langley and many others had failed was essentially irrelevant. And, uh, the key was not “execution” but the “ideas” of the wind tunnel, the calculations, and the three axis control. Sewing the linen or whatever onto the spruce or whatever was “execution” and just routine.All that additional data, or, if you will, information, was visible right there in their bicycle shop. So, with that additional data, the old empirical averages were irrelevant.Then we can see, for a large fraction of firsts, the simplistic empirical data from ‘the rear view mirror’ is irrelevant or nearly so.Uh, just looking in the rear view mirror this way and following a herd of sheep can lead to remarks like “sheep get sheered” if not fleeced.More generally, we’re trying to do something new. So, we’re not trying to plow the lower forty acres just like they’ve been plowed productively for the past 100 years.Still, we want to know, in advance, long before ‘traction’, that our “ideas” and plans are solid. How to do that? The Wright brothers provide one example. Such “ideas” are rarely “easy”.For more the US DoD has provided many examples since about 1941. Commonly their advanced projects got funded just from “ideas” in paper and long before any “execution”.E.g., for GPS, it would be good to have accurate data on the gravitational field of the Earth. For that it would be good to have a satellite in orbit and that had no drag from the gasses up there. So, how to build a ‘drag-free’ satellite? For that we need an idea, one we can have confidence in, before, during, and after launch. So, we need an idea for a drag free satellite, and a way to evaluate that idea, early on, just on paper. The DoD did that. How’d they do that? Uh, as we’ve been told, “Ideas are easy”, so I won’t bother you with something so easy!So, how to get powerful ‘secret sauce’ for a Web 2.0 project and, just from a description on paper, be quite sure the idea is good, good enough to build a major company, say, worth over $50 billion, and where execution is routine? Yes, the evidence from some of the best sources in our civilization is that we can do this.Then it is fair to say that, yes,”Maybe I´m missing something big…”For what you are missing, below is a little more, that I drafted but omitted to reduce length:For a hypothetical project, consider exploiting the ‘social graph’. Okay, that is data. In the exploitation, we will have software manipulate that data, and the manipulations will necessarily be mathematically something, understood or not, powerful or not. For more powerful manipulations, some mathematics can be crucial. What mathematics? High school? College? Graduate school? Original? Emphasize the last two and not the first two. I’m not talking high school second year algebra here, guys.For a little more, suppose there is an answer from exploitation of the social graph and that we want to report to the users. Likely we can only estimate or approximate that answer. So, we should consider some techniques of estimation and approximation. Some of this material, even if not original, is deep enough to be known by only a small fraction of the mathematics professors.To be a little more clear, some of the mathematics is astoundingly general and also applies well to using available data about the social graph. Or we can take available data and find ‘new connections’ in the social graph.For a little more, in Web 2.0, beyond furthering gossip, we very much have another ‘data exploitation’ in mind: We want to get ad revenue. Here a crucial point is ‘ad targeting’. So, the question is, given the data, what ads should we display? We can believe that in principle there is an ad that will make us the most money. So we want to ‘estimate and/or approximate’ the selection of that best ad, and here we are essentially forced into some mathematics. If we are to do this work well, that is, that looks good on the way to the bank, then we better not rule out graduate school and original mathematics.For a project not hypothetical, could I say more, with theorems and proofs, with some original work based on some advanced prerequisites, written out with D. Knuth’s TeX? Modesty, shyness, and confidentiality constrain me. Might my work be only for ad targeting instead of more in Web 2.0? No.My conclusion is, in the future of computing, starting now, as we seek to ‘automate’ work, including using the Internet to help young women gossip, the present ubiquity of intuitive and heuristic techniques for manipulating the data will much less powerful than we want, and original mathematics based on advanced prerequisites will become more important than Moore’s law. Likely you read this first here.The crucial issue isn’t “Web 2.0” or gossip: The crucial issue is manipulating the available data to yield money in the bank, and for such manipulations appropriate mathematics, commonly quite advanced, makes a 900 pound gorilla look like a flea that can’t bite. Again, intuitive and heuristic techniques can’t hope to compete, eventually maybe not even in helping young women gossip.Look, appropriate parts of the US DoD have for decades known very clearly, and have strongly acted on, the power of mathematics; the mathematics of estimation, approximation, optimization, stochastic optimization, etc. have been warmly greeted and well exploited. I started my career in such work; like essentially everyone with a US technical Ph.D., the US DoD paid for nearly all of it; when my wife encountered delays in her Ph.D., we ran out of our budget and I took a part-time job to see us through, again paid for by the US DoD. Thank you!The Wright brothers, Fat Man, Little Boy, the SR-71, GPS, the F-22, and much more provide examples that we really can have good “ideas”, that we can evaluate well just on paper, ideas good enough to make execution routine. In particular, our key can be some appropriate, possibly in part original, likely advanced, mathematics.So, borrowing from the techniques the DoD long used with high success, we can be one of the first with such techniques and ideas for more powerful, valuable ‘secret sauce’ ways to manipulate data in Web 2.0, be quite confident quite early on just from work on paper, have ideas good enough to make execution routine, and go ahead and get a major company.Much of the challenge of this whole ‘context’ is being successful doing things that are new, and we need to take this challenge seriously.Gee, I thought that we all understood these things long ago.I do suspect that the important audience is missing — the LPs.

    1. paramendra

      This guy is no JLM.

      1. sigmaalgebra

        You are angry. Why? There are some standard answers, but none of them justifies your anger.I’ll boil it down to a ‘demi-glace’ (as in Escoffier or maybe in Myhrvold’s recent books; combine with shallots sauteed in sour cream butter, a 1959 ‘La Tâche’ reduction, and add truffles and heavy cream) of just three easy tablespoons for you:First, Doerr said:”Ideas are easy. Execution is everything.”andThere are “plenty” of ideas.These statements are nonsense, uninformed, misinformed, just plain wrong, arrogant, ignorant, destructive, dangerous, etc. nonsense.Second, here’s how to patch up his statement to make one that is true:Bad ideas are plentiful and easy to find. Given a bad idea, all that is left with any chance of value is execution so that the execution is forced to have everything of value but, starting with just a bad idea, tends to be from challenging down to foolish and hopeless.Third, what is also true and much more productive is my:Good ideas are rare and difficult to find. Given a good idea, execution tends to be routine, and the US is awash in ability to do routine execution well.Uh, as we know very well, all across the US are Main Street entrepreneurs who serve food, repair machines, pave roads, supply products, etc. and execute well, quite routinely. Bluntly, the main difference between them and a valuable Internet business is just the ‘space’, that is, the Internet versus Main Street. For these millions of entrepreneurs, ‘execution’ is not a problem and should not be for Internet entrepreneurs.Further, for ‘good ideas’, we know a lot about how to find them and to be quite sure they are good quite early on, just on paper, and the US DoD and US Congress have ‘got it’ on this point quite well at least since 1941. In about 1945, V. Bush, J. Conant, and D. Eisenhower ‘got it’. As an important special case, I went on to outline a ‘paradigm’ from new to nearly so for ‘social Internet’ entrepreneurship:The software takes in data, manipulates it, and reports results. For manipulations that are more powerful and yield more valuable results, exploit mathematics, possibly new with advanced prerequisites.With this paradigm, you can become angry, fight with ad hominem attacks, etc. but have less hope of being correct than getting silly putty to cut granite, and here is why: The paradigm stands on some of the most solid foundations in all of civilization and has been proven powerful over and over by the US DoD since 1941 or so. It takes a determined degeneracy to ignore this paradigm.But a question is, does the paradigm apply to getting valuable results in the ‘social Internet’? Actually, from looking at the internals of the paradigm, the answer is clearly “yes”, and here is much of why:Again, the software takes in data, manipulates it, and reports results. Commonly we cannot get (and for part of the opportunity, are not getting) the desired results with just the manipulations from direct, simple, intuitive, or heuristic techniques. Thus powerful mathematical techniques for estimating, approximating, optimizing, etc. can become valuable and possibly crucial. Looking at these techniques closely it is easy to see that many of them are much more general than needed for the data in ‘social Internet’ and do apply there also. Here’s how to look closely: Start with a good undergraduate major in pure mathematics and then have 2-3 years of well focused graduate study in pure and applied mathematics, do some original research that passes peer-review and solves some practical problems, and then spend some years making applications, possibly to US DoD problems. For a faster path, look for names Lebesgue, Hilbert, Birkhoff, Riesz, Kolmogorov, von Neumann, Doob, Dynkin, Rockafellar, etc. Look for more recent authors Halmos, Royden, Rudin, Loeve, Neveu, Cinlar, Luenberger, Breiman, etc. So, emphasize Stanford, Berkeley, U. Washington, Princeton, Cornell, etc. For a more intuitive explanation, I’m talking an F-22 against a WWI biplane or maybe just a bicycle or a horse. On this point, I’m standing on the shoulders of giants standing on the most powerful constructions of civilization; you cannot beat me; it’s hopeless to try.Someone angry might be eager to jump to misinterpret my claim and paradigm, erect a straw man, and knock it down. So, I should add: Intuitive insight is difficult to check or evaluate but remains important. So does luck. Clearly it is possible to make a lot of money in ‘social Internet’ following Doerr’s claim instead of mine and without my ‘paradigm’. Yes, in ‘social Internet’, so far my claim and paradigm have had little to no role.Still, I’ve provided strong support for my claim and paradigm. Moreover, if we are the first to bring powerful material to a new field, then necessarily there will be no history of success of that material in that field. While this situation is inevitable, first, it is not very important IF the material has strong support anyway and, second, is much more of an opportunity than a problem.There is a similar situation, and more of an opportunity than a problem: So far in the economy, high returns and major successes are rare. So, clearly at most only a tiny fraction of the population knows how to achieve such success. So, even given a good approach to success, not many people will understand it. This situation is also inevitable.But the battle between (A) emotion, passion, and intuition and (B) rationality, logic, mathematics, and science has been going on at least since the ancient Greeks, and so far (B) has made steady and major progress against (A). Now as we program computers, we give up too much if we ignore (B) and make the software just routine and otherwise plan the business based just on (A). Going forward, for more value from computing, we stand to have more exploitation of (B), in my opinion over the next few decades enough more to make (B) more important than Moore’s law.While over the past 10 years some venture firms have done well, it appears that a large fraction of venture fund limited partners (LPs) are disappointed in their returns and, thus, will be looking for the goal in that famous one word description “more”.Eventually the LPs will notice: We can start 256 = 2^8 stock picking newsletters and in 128 of them say “Buy” and in the other 128 say “Sell”. After three months, keep the 128 winning newsletters and have 64 of them recommend “Buy” and the other 64, “Sell”. After three months keep the 64 winning newsletters and have 32 say “Buy” and the other “32” sell. Continue until have just one newsletter that has been correct eight times in a row and then tell the people who received that newsletter, “Clearly we are the Michael Jordan of stock picking news letters. For more of this fantastic, sure-fire advice, send $20,000 for your subscription to our newsletter for the next quarter.” Now I’m not saying that this is exactly what any 256 venture firms do; I’m just saying that in principle this is what someone could do!When the LPs notice this situation, they may start to look for better means of making money. They can begin to set aside plentiful, easy, bad ideas and weak techniques and move to rare, difficult, good ideas and powerful techniques; for this change they can just borrow from the track record and work of the US DoD since 1941 and the main reason the US Congress has generously funded research at Stanford, Berkeley, U. Washington, Princeton, Cornell, etc. The DoD and Congress both ‘got it’ decades ago. Maybe someday the LPs will also ‘get it’. In the meanwhile, there’s a LOT of opportunity for good ideas and powerful techniques.But I’ve got to get back to work. I’ve got the crucial, core, challenging work done and now am working with routine nonsense, today trying to be quite clear (rebuilding my software installations after an absurd virus attack from just one use of the Akamai Download Manager rot — all their programmers were on two quarts of 100 proof vodka a day? — just to get a PDF file on a motherboard) about the T-SQL statements to create a SQL Server ‘user’ with appropriate ‘permissions’ (should be called ‘capabilities’) ‘schemata’ (a really inappropriate use of ‘schema’, especially in the context of relational data base) to ‘map’ (a really bad word choice) to the SQL Server ‘login’ for Microsoft’s IIS. Some of the ‘best’ documentation has the semi-, pseudo-, quasi-great “‘class’ specifies the class of the securable” where apparently nowhere else in the 1186 pages of the book is there any other discussion of ‘class’. We’re talking mud wrestling here, and throwing the mud against the wall overhand, with back spin, etc. until it appears to stick. I may have it worked out by lunch and then get back to writing the rest of my few, simple Web pages.For your anger, you have nothing to be angry about. For what I’ve written here, you likely have some good things to learn. Yup, learning can seem painful, but it’s really not. To avoid such anger, just accept working with ideas ‘objectively’.Buthttp://www.youtube.com/watc…yes, likely the Charles Laughton character (or me, for that matter!) thinking about the Maureen O’Hara character, should make you feel better and is available at Amazon as an MP3 download, with somewhat better sound quality, for 99 cents! Spring for the 99 cents and feel better! Of course Alfred Newman knew about this music!The composer of this 99 cent music very much knew just what he was doing: He was from Austria about 100 years ago and musically close to Richard Strauss (the ‘Also Sprach Zarathustra’ Strauss, not the Johann Strauss, of ‘2001’), Anton Bruckner, and Gustav Mahler.But, gee, maybe this music will also just make you angry!

        1. paramendra

          Oh my. What did I do?

  23. paramendra

    Watched last night and promptly blogged it. Too long? I say not long enough.

  24. Mordy Kaplinsky

    Fred,I finally got the chance to watch this video, and I wanted to say thanks.Thanks for pointing out the lack of new product innovation at Google.Thanks for pointing out the lack of home brewed innovation at FacebookThanks for your comments on Facebook being a glorified photo and chat platform.When I say this to people they look at me like I blasphemed against the gods of tech. It’s time to create new innovative products out of existing tech companies and new startups.Thanks for your lack of fear to state the uncomfortable truth.