A Blast From The Past

I saw this in my twitter feed today.

Almost a decade ago, my friend John Heilemann interviewed John Doerr and me at Web 2.0.

It is interesting to go back a decade and see what we were talking about then.

Some of the same issues exist today. Some of the questions we debated have been answered now.

It’s about forty minutes long.

#VC & Technology

Comments (Archived):

  1. Kurt Stangl

    Great share!

  2. Tom Labus

    Is there a sound issue?

  3. Richard

    Favorite line by Fred defending his trolling comments about facebooks lack of home grown breakthroughs and and diminishing its incredible execution – “I’m just trying to be provocative”

  4. sigmaalgebra

    What I remember about Doerr at AVC is, IIRC, his:Ideas are easy and plentiful. Good execution is hard and rare.Then and now I would respond: “Bad ideas are plentiful, and then execution is hard. Good ideas are rare, and then execution should be routine.”I was astounded, took a long time, with oceans of evidence, finally to conclude that Doerr, KPCP, Sand Hill Road didn’t have even as much as a weak little hollow hint of a tiny clue about how to evaluate ideas or projects in technology They talk about new markets and products, new technology products, “disruptive”, “new, leading edge”, but all that is just window dressing. Instead, their criteria, what dives them beyond all else, is traction, significant and growing rapidly, in a large market plenty large enough to support a company worth $10+ billion. Given that situation, those criteria, they don’t care about evaluating the ideas or technology. Or if they thought about it, and I doubt they would, they might conclude that if there is the traction, then the ideas, whatever the heck they were, MUST be good.So, really Doerr, KPCB, and Sand Hill Road have essentially no insight in how to go from an “idea” through “execution” enough to get the necessary and sufficient condition of the criteria. They get in their e-mail lots of “ideas”, but, since they don’t actually evaluate them, their view is that “ideas are easy and plentiful”, and they have no ability to find, or effort on finding, good ideas. Again: They don’t consider the ideas so are free to think and say anything about them. Instead they just look at their criteria of traction, etc.So, since Doerr and Co. don’t consider ideas, bluntly don’t take seriously their remarks on ideas.”Ideas”?(1) Where do the black holes with 20 billion solar masses come from? Idea: Direct collapse in less than the first billion years.(2) At FedEx, given all the planes in Memphis, all the packages to be delivered, all the packages to be picked up, and the relevant times, which planes go to which cities in what order to get all the work done at least cost? How to find that? With what “ideas”?(3) Record some ocean wave noise. Now for some simulation experiments, want more sample paths of just such noise, that is, of a second order stationary stochastic process with the same power spectrum. How to do that? With what “ideas”?(4) Given a month of operating data of many parts of a server farm, how to monitor the farm in real time to detect “zero day” (never seen before) problems with false alarm rate known and adjustable, exactly, in advance and in some valuable sense the highest possible detection rate? How to do that? What “ideas”?Even with good ideas, the goal is still to make money so that, roughly, need to be in a large market or at least one that will become large, e.g., smartphones. Still, on the way to the market and, in particular the “traction”, ideas are crucial, we need good ideas, good ideas are rare and valuable and permit execution to be routine.It appears that Sand Hill Road is not permitted to think such thoughts: Why? It appears that the main limited partners (LPs) agree that they should think a lot like commercial bankers who write checks only with collateral, that is, make a financial investment only in an existing financial asset. For VC, the LPs are willing to bend some and regard the traction, etc. as enough of a substitute for the financial asset. So, with these attitudes of the LPs, the Sand Hill Road Gang thinks about traction, etc., and speaks about ideas with the same care that AOC speaks about her “Green New Deal” — irrational, uninformed, misinformed, wildly overly emotional, dangerously irresponsible, ditsy bimbo.Ah, the Internet is terrific! I’ve found several houses nicer than the one I’m renting that I can buy for less than I’m paying in rent!!!

    1. awaldstein

      successful execution is never routine.

      1. sigmaalgebra

        There is a LOT of really good execution in the US military, McDonald’s, the hospital that put a 1/2″ long tube of cobalt and chromium in one of my arteries and did some ultra sound showing that otherwise my arteries were all open, ….Getting another 10X unicorn is not routine and, instead needs in addition to good execution one heck of a lot of luck and/or some “good” ideas so that, then, can have the 10 X from efforts that are so common they are “routine”.

        1. awaldstein

          i honestly have no idea what you are saying.

          1. sigmaalgebra

            I’m pretty good at understanding and explaining things, so I’ll try again.PreviewFor just a fast preview, clearly enough there are several important largely separate parts of our civilization, and here we are concerned with about three of them(i) research from mathematics and physical science,(ii) the economy and business, and(iii) the US military.For (i), for what I am saying here, I focus just on math and mostly set aside physical science. For (ii), the economy, I’m focusing on information technology startups essentially as of interest to venture capital on Sand Hill Road. For (iii), the military, I use that for some examples important for (ii) from using (i) with both the math and physical science.John Doerr’s RemarkIn particular I am responding to Fred’s post with an old remark from John Doerr.So, the context of Doerr’s remark was how to do a successful Sand Hill Road style information technology startup.For such a startup, and of course for business more generally, commonly early on the startup entrepreneur has an idea. E.g., Bill Gates saw a big future in microprocessors and wanted to “put a PC on every desk” for “document preparation” — i.e., get rid of the typewriters.Of course, to be successful, the entrepreneur has to execute the business based on, at least initially, the idea.So, sure, some entrepreneurs are proud of their idea, believe that it is good and the key to a successful business, and contact Sand Hill Road for funding.In this way, KPCB and Doerr get lots of contacts with ideas. Then Doerr formulates, IIRC,Ideas are easy and plentiful. Execution is everything and rare. My response isBad ideas are easy and plentiful, and then execution is everything and difficult. Good ideas are difficult and rare, and then good execution should be routine. By “routine” I do not mean easy or universal. Instead I mean that in the US economy many tens of thousands of people are good at execution, and their companies execute well.Or there is a point: We have a short list, that even Tennessee hill boys can count without taking their shoes off — Microsoft, Intel, Cisco, Google, Facebook, Amazon. That’s six. They all execute well. Still there are tens of thousands of people, and companies, in the US that execute well.Point: We have only those six companies that thrill Sand Hill Road but tens of thousands of cases of good execution.Sooooo, John, good execution is not rare but common and, say, “routine”. What is rare are the business ideas and/or luck good enough to be one of those six companies.So, good ideas are rare, and then good execution should be routine.To draw from the US military, as JLM has outlined in fantastic detail for both the military and business, the military is just terrific at training people to execute very well and has on any day in uniform tens of thousands of officers really good at execution of at least just what they have been trained for but no doubt lots more needing innovative exploitation of situational awareness and special opportunities. So, again, good execution is common, common enough to be routine. Again, I didn’t say easy or universal.Good IdeasAs is easy to see from examples of especially the US military, it is possible, sometimes even crucial, to do much better than can be done with just good execution.E.g., in WWII in the Pacific, the US started to execute well in the Coral Sea and then brilliantly, with some luck, and a lot of willingness to take big risks, charge into the unknown, and lose US lives, at Midway. Then slowly the US built one heck of a navy, went from one, two, maybe 2 1/2 carriers to at least 24, say, at Okinawa. We put radar in the planes. We took the Marianas with Tinian and built the B-29 which could fly nonstop, no refueling, from Tinian to Japan and back. And we built Fat Man and Little Boy.Then, presto, bingo, we got unconditional surrender in less than a week and saved maybe 1 million US casualties from a direct invasion of Japan.Uh, ROI? From some R. Rhodes writings, the Manhattan Project cost about $3 billion. So, that would be $3 K per US casualty saved, a grand bargain, great ROI.So, with Fat Man and Little Boy we used (i) above, math and physical science. Some of the math is now in very clean form in an E. Cinlar book with Markov “branching processes”. In grad school, I used that result to solve an exercise better than the prof did and got an acknowledgment! Some of the S. Ulam work led to, say, random number generation viaX(n+1) = 5^15 + X(n) + 1 mod 2^47which passes the severe Fourier test and I once used to estimate the survivability of the US SSBN fleet under a special scenario of global nuclear war limited to sea.Ah, I can like some of what Ulam did: I used his classic result the French probabilist Le Cam called “tightness” in my paper on anomaly detection.And, bluntly, one of the keys to both the Coral Sea and Midway was a guy Commander Rochefort who broke one of the Japanese codes. The real hero of Midway was not Spruance or Nimitz but Rochefort. Sorry ’bout that, in a way, especially because Rochefort, back at Pearl, was at no risk of shedding even a drop of blood.I know; I know; it was Monty and Patton who drove Rommel out of North Africa. Uh, …, it also helped that, maybe due heavily to A. Turing, that a German code was broken and a huge fraction of Rommel’s supplies ended up on the bottom of the Mediterranean. At one point, Rommel’s son claimed that his father thought that there was a leak in the Italian military. Nope: Turing and Co. broke the code and, thus, knew when the shipments were on the way so that the British forces in the Med could sink the shipments (there’s a little more to the story).I know; I know; “… so much owed by so many to so few”, but the real key to winning the Battle of Britain was British progress in radar, now commonly in microwave ovens in kitchens.In Gulf War I, there may have been more allied casualties from accidents in R&R ball games than from enemy action. Astounding. The keys here were beyond good execution and were due to the (i) above, math and physical science, e.g., the F-117 essentially invisible to radar and the plane of choice for the first attacks into Saddam’s military infrastructure. The design of being invisible to radar uses a lot of math and physical science in (i).So, as promised, we have some astounding examples of “good ideas” based on (i) and used with astounding, world-changing, effects by the US military.Good Ideas in BusinessSo, now, on to “good ideas” in information technology startups in business:From the examples above, we want some ideas good enough to be as powerful and valuable as Fat Man, Little Boy, the F-117, the code breaking, the branching process calculations for critical mass determination, the infrared to “plink” Saddam’s tanks, etc. Then, as we showed in Gulf War I, we can win astounding victories from, then, merely “routine” execution by the soldiers in uniform, where one of the astounding results is that not only do they devastate the enemy but also have not much risk to their own blood.Yes, of course, for our success, we need a suitably large market where we can apply our powerful technology with valuable results.ParadigmBack to (i) with math and physical science:In an information technology startup, we take in data, manipulate it, and report results. E.g., Google.Well, for more valuable results we can hope to use more powerful manipulations, and as the military successes in information processing showed so clearly for the more powerful manipulations we can proceed mathematically.Mostly Sand Hill Road doesn’t understand this paradigm. Sure, then, the flip side is an opportunity.While Sand Hill Road is not willing or able to evaluate such powerful new technology, thankfully the US NSF and NIH, along with some other parts of our civilization, are able.So, that’s what we want in “good ideas”, ones as powerful as some of the military examples and that permit using just routine execution. say, as in the crew of the Enola Gay or the pilots of the F-117s, to get the desired successes.SummaryMath, physical science, the STEM fields, and the astounding results are now all over our civilization. The US military and some parts of our economy are good at exploiting the STEM work. What is astounding is that Sand Hill Road won’t evaluate such work. In effect they pretend that the STEM field work is irrelevant, e.g., nowhere nearly as important as “execution”.

    2. JLM

      .Execution is the real advantage in business. The Special in Special Forces is execution which in that line of work is called “battle drill.”http://themusingsofthebigre…Execution advantage is not a single pill a company takes; it is the product of excellence at every level and at every action point in the process.It is hard to achieve, but when achieved is the critical driver of culture.Elite companies, elite military units know they are elite. They have a certain swagger that comes from winning.JLMwww.themusingsofthebigredca…

      1. sigmaalgebra

        “Execution is the real advantage in business.”Yes, for the URL, Amazon may have had another leadership or execution advantage: When they were designing all the parts, pieces, data gathering, record keeping, data base queries, and notifications sent to customers, some high pressure middle manager might have burst into the meeting, seen the resources in time, effort, head count, meeting hours, design documents, calendar months, etc. thought “Does it have to take THAT long, take THAT much effort, be THAT complicated, and cost THAT much?” And the head of the meeting might reply that the plans were just first cut and minimal and that much more was needed, in hardware performance, overall system performance, data backup and recovery, system monitoring, and, then, correction when disaster struck: E.g., someone hacked my e-mail address, put a bunch of products in my Amazon shopping cart, and changed the e-mail address of my account. I called, and Amazon cleaned that up, in every detail, any rebates, with roll backs to what was correct, etc. So, that required major surgery on my data in their database. That TOO. took a lot of planning, designing, coding, testing, documentation, training, etc.Point: Bezos and Co. were ready to take on all that, sweat the details, take the time, head count, meeting hours, planning document pages, code, testing, support staff training, etc., all so that I could get an order for some SATA cables or some such!Smart cookies, and Bezos had the money to fund it.”Execution is the real advantage in business.”In practice, nearly always, yes. But the VC game is 10X unicorn hunting, and for that need luck and/ or something in addition to good execution.I’m quite prepared to believe that the US military is really good at training thousands of people a year to be really good at execution. Their training is solid, practiced, proven, effective and, thus, for them routine. Dad was long the head education expert for one such training program, 40,000 students at a time. When I visited his office, the streets were spotless, the concrete curbs were all freshly painted white, the floors were spotless and freshly waxed and buffed, etc. And the soldiers the US military trains are nearly all really good at execution, especially the execution they were trained for, And I’m quite prepared greatly to favor a former US military officer for a COO slot if I do get that far.But in military terms, it also greatly helps “execution” if have some B-29s and one Fat Man and one Little Boy. And it helps if have some F-117s as the first airplanes into enemy air space. And some infrared on the F-111s were terrific for Gulf War “tank plinking”. Our air to air missiles are nearly beyond belief. Phased array radar, adaptive, phased array sonar, GPS, laser guided munitions, …, all great helps for making execution routine!Praise from the enemy can be welcome: In one of the Gulf Wars, an enemy tank was parked between two buildings. We fired a missile, knocked out the tank, and left the buildings. “American technology is beyond belief.”There are not many 10X unicorns. IMHO, to create one more, need one heck of a lot of luck and/or some big advantages, some “good ideas”, good enough to make execution, say, by a really good US military staff officer as COO, “routine”.

      2. Richard

        I’m not a user, but one could argue that FB execution since 2008 has been brilliant – particularly for what it has done behind the scenes, and for what it has not done to its product.

  5. jason wright

    I remember this one. I think i’ve been here for too long.

  6. William Mougayar

    Do you still think that FB is just a photo sharing app 😉

    1. sigmaalgebra

      By all means, say what FB really is! Contribute to the debate! Inquiring minds want to know! E.g., why is it so much more valuable than anything on a BBS, Compuserve, AOL, Prodigy, Google Groups, MySpace?I have a FB account, but I rarely use it. I find their Web pages badly designed and ugly. Their Web pages jump around while I’m still looking at them, all for no apparent reason. The pages are awash in undocumented triple tricky functionality — e.g., in their multi-line text box, Enter sends the message. Yet, if format a message with blank line paragraph separators and paste it into their box from the system clipboard, then that does not send the message and the blank lines are retained. To edit a message sent, have to find some obscure three dots (to quote a famous movie, “I HATE that hacker crap.”) and hope. And their other stuff, especially privacy, is a black hole of undocumented obscurity.Just what the heck do people like about FB?Uh, such a ‘social thing’ is about particular to a single natural language. So, a US company should have no advantage in India, Romania, Kenya, Egypt, Latvia, Finland, Japan, etc., yet somehow FB does.My guess is that it’s important to understand these things, and in advance!

  7. Mike

    Facebook now has a market cap of > $500B. Easy to predict looking back of course. A lot of the familiar names – Google, FB, Amazon, Apple. Microsoft?Related to the earlier post this week. VC is not the only business where hit rates are low. It is very difficult as well for large, well resourced tech companies to deliver new innovations, or any company or organization deliverng a new product. Poker players have a good model for looking at this called pot odds that I am sure many here are familiar with. As long as the pot payout exceeds the risk of the bet you win in the long run. Need to get the math right and you can’t go bankrupt in the interim of course.

    1. jason wright

      Large tech companies generally seem to acquire rather than produce innovation.

  8. gmalov

    Web 2.0 was a great conference, miss that! Massive amount of wealth created around web over this period. It would be fun to see You and JD reconnect at the 10-YR mark and redo the discussion, at present.