The Analog To Digital Move

One of my favorite business moves over the past twenty years was the one Netflix pulled off. They started out as a subscription DVD rental business and evolved into the leader in subscription streaming video. I’ve never seen anything written about whether that was intentional all along or whether they figured it out as the business evolved. It doesn’t really matter, it was a great move in either case.

This move, the “analog to digital”, ought to be more common but I can’t think of many sectors in which it has played out that way.

But we are certainly witnessing another one in the ridesharing sector. If anyone really thinks that ridesharing is a way to get more drivers work, the news this morning that GM has invested $500mm in Lyft and will be strategic partners on “developing a network of self-driving cars that riders can call up on-demand” should put that to rest.

In this case it isn’t DVDs that will go away and replaced by bits. It is the drivers.

How this all plays out is anyone’s guess. But using an analog asset to build a large customer base that can then be leveraged in a native digital model is a great business move and I am surprised it has not been done in more markets over the past couple decades.

#VC & Technology#Web/Tech

Comments (Archived):

  1. Adrian Bye

    it may be easier to do now as tech is evolving faster so you can see the next steps more easily

  2. aminTorres

    Did Audible pull off a similar feat in your opinion?

    1. fredwilson


  3. JimHirshfield

    AMZN has done a good job going analog (books, CDs, DVDs) to digital (Kindle, Prime Music, and Prime Video). I used Prime Music more over vacation than Spotify – great experience; better on mobile than desktop, and commercial-free ’cause Prime.

    1. Tom Labus


      1. JLM

        .A unicorn in the belly of a unicorn. Hate that word yet?JLMwww.themusingsofthebigredca…

        1. Tom Labus

          a very profitable one for them

          1. JLM

            .It is also growing like a weed. Not really priced into the stock.JLMwww.themusingsofthtebigredc…

        2. Twain Twain

          Unicorns are fine if you were a little girl brought up on My Little Pony and Greek mythology about Knights in Shining Armor who save you from Medusa and the Kracken — see ‘Clash of Titans’ movie.Funnily enough, there was not a single My Little Pony in my household.I was, though, bewitched by the MECHANICS of how Ray Harryhausen (genius) did his robotics armature for ‘C of T’ and how they cut the film sequences for perspective.Nerd kid alert!

          1. JLM

            .So, grilled or smoked? I prefer smoked.JLMwww.themusingsofthebigredca…

          2. Twain Twain

            Darjeeling-smoked, haha.

          3. Joe Cardillo

            I adored Jason and the Argonauts as a kid! Harryhausen was rad

      2. Twain Twain

        Blockchain has somehow gotten positioned in the Visa / clearing & settlements space but actually it’s competing against AWS.

    2. Humberto

      I think Amazon will have an upper hand in the end. In any case, the subscription model Netflix/ Spotify isn’t sustainable in the long term.Specific brands/ movies/ studios will always leave stuff out because the platform can’t afford them. Even when they can afford premium stuff, they will have to oversell and underinvest elsewhere.Markets are value, and not price oriented. I think pay per view (simplified) will win.

      1. JimHirshfield

        That supports the model Netflix is pursuing of producing their own high quality content. Makes them the same as HBO.

        1. Humberto

          Yes, but how long can you sustain being a gatekeeper and content producer yourself? There’ll be a limit on the quantity and quality being produced

          1. JimHirshfield

            But HBO has been both for quite some time. Don’t think it’s an issue.

    3. fredwilson

      great point. i think books to hosting is even more amazing

  4. William Mougayar

    UBER/Google vs. Lyft/GM? Hmm. I would place my bets on UBER/Google, or something with Tesla in it.”GM President Dan Ammann, who is joining Lyft’s board as part of the deal, expects the automotive industry to “change more in the next five years than it has in the last 50 and we obviously want to make sure we’re at the forefront of that change.”That statement by GM has little credibility or historical proof points in my opinion. Tesla is already half-way there into self-driving cars, and Google is there. How can GM leapfrog them?Let’s revisit this one in 3 years on 12/31/2018.

    1. Martin De Saulles

      I should have thought there was room for several players – it will be a big market and competition will be essential. I cannot see Uber/Google controlling personal transport – at least I hope not.

    2. Twain Twain

      Worth reading Pando article on Uber vs Lyft:Uber received a lot of flack in May of 2014 when CEO Travis Kalanick stated that drivers would eventually be replaced with driverless cars. “‘Look, this is the way the world is going” is how he dismissed the legitimate concern of the tens of thousands of drivers that built the backbone of Uber.Lyft does not disagree with the inevitably of a driverless system. However, they implemented some tact when making the statement, and refocused the narrative how the social component will still very much exist. The dynamic will just be a bit different. Instead of drivers meeting passengers, driverless cars will allow for strangers sharing a ride to get to know one another. It still isn’t much consolation for drivers, but nonetheless, the approach was materially different than Uber’s.*…Competition between Uber and Lyft is a great thing for consumers.

      1. William Mougayar

        Competition is great, but linear predictions are not. I can bet something else that’s new will emerge that drives a wedge between these 2 apparent choices.

        1. Twain Twain

          Let’s say APPL, AMZN & Toyota / VW tag team up.

        2. Richard

          Uberfication of the lyft supply chain is one.

      2. LE

        Instead of drivers meeting passengers, driverless cars will allow for strangers sharing a ride to get to know one another.What the f is that all about? Why would I want to get to know another passenger? Are people doing that now on buses, airplanes and trains?

        1. Twain Twain

          Oh you anti-social skeptic you, :*).I’ve struck up conversations with random people on buses, planes and trains and become friends with them (one such friendship is 14 years and counting).Total strangers have seen me knitting whilst waiting for the bus and within a minute they’ve told me all about their sisters and what they’re cooking their kids for dinner.

          1. LE

            If I was sitting next to you on a plane and you were knitting I would probably ignore you. (Unless I was single and you were attractive of course). But if I was sitting next to you on a plane and you were reading something interesting (non-fiction) or doing work I might conclude you were either smart or important and I’d learn something or gain something from you so I might strike up a conversation. Not that I don’t strike up conversations with other types of people I actually do but my goal is typically to learn something about human nature or how they do their menial job. I have walked back to the stewardesses on an airplane and started to discuss their job with them “do you guys know each other” or “how are you scheduled” things like that.

          2. PhilipSugar

            As somebody that has spent 500 hrs a year on a plane for the last 20 years. I can tell you this with total assurance. The amount of time talking to your fellow passenger has decreased more than 99.99%Nobody wants to talk to you. People don’t even want to talk to you or look at you even through meals, where I like to talk. (yes I am in the front where they serve meals on plates with real silverware)

          3. LE

            That’s interesting and I love when my bias is confirmed.Why do you think it has decreased? Is it because of social media and the fact that people already have enough trivial interactions? Is it because the makeup of first class has changed? Or because you have gotten older (see below).Maybe when you were younger it was different because an older guy might get a kick out of helping someone younger out. I mean if I was on an airplane and there was a guy my age I probably wouldn’t be to interested unless my reasons stated in my above reply were met. Otoh if it was a kid in college I might want to play elder statesman and help him out. That’s just my thinking off the top.

          4. PhilipSugar

            It is purely because that device in your hand means you do not have to look at the person next to you.I actually am very strict. If I am with you or anybody else and you pull that out during a meal or a meeting, I get up and walk out. The only acceptable excuse must be pre-announced before, or an emergency which you explain.

          5. LE

            If I am with you or anybody else and you pull that out during a meal or a meeting, I get up and walk out.Well what is the exact class of people that you would do that with?Obviously you aren’t going to do it with an important customer. So in a sense the fact that you would choose to do it in some cases (but not all cases which is obvious) says to the person pulling out the phone “you aren’t important enough so I will enforce this rule on you”, right?There is a guy that I deal locally with here for some things. He is always checking his cell when I see him at, say, Starbucks. We aren’t seated but say we are standing waiting for coffee and having a conversation. In the middle of a sentence his phone rings or he gets a text and the phone has his attention. Really annoying to me. One time he invited me out to lunch and I went expecting the same thing to happen. Much to my surprise it didn’t happen. My guess was that he was called out on it and knows he has to behave. Guess what I was ready to blast him if he did. I was preplanning to walk out (take your jack and shove it thing).Here’s the thing though. Yesterday he said something like “why don’t we do this deal together”. I decided right on the spot that I didn’t want him as a partner not solely based on his attention span (appears to have ADD) but that was a large contributing factor.

          6. PhilipSugar

            I’ll expand on this. The other day I was out for lunch with my daughter. There were two families of four sitting to either side of us. It was the holiday season. There were eight sets of eyes that never came off the small screen. My daughter said: I guess that is why you don’t allow phones at the table.If you can’t talk to your family during the holiday season at the table in a festive Mexican restaurant why would you ever talk to some stranger when you are doing something boring like being in a plane or a car.I try and be polite and acknowledge you when you sit down, not strike up a conversation, just say hello, I would say I am not acknowledged 80% of the time, and my bar is not high, a simple shake of the head is fine.

          7. LE

            My (2nd) wife and I are a bit different on this. We don’t do the forced conversation thing at all. If we go out to eat we both have no issue with kids keeping occupied and out of our hair with games or cell phones. So it’s not that we accept it and can’t prevent it it’s that we aren’t into doing the kid thing. Just the way we are (and noting that you said your wife likes doing the car pool thing my wife doesn’t). Everybody is different. When I come home (and after exercising) I like to read the paper and eat at the same time. My wife does as well. We talk at other times a great deal though. Just at the table we are both cool with reading. We are compatible like that. If we go out to brunch we might talk or we might both bring the Sunday paper to read. If we see something interesting we will discuss it.I would say I am not acknowledged 80% of the timeThat’s interesting actually. For fun you might want to do this test (on an airplane). Next time you sit down and the person doesn’t acknowledge you pull out a brochure from some expensive high end easily viewable object. Like a super expensive car brochure or a business jet and see if the person gets any more interested in you or some other obvious trapping of wealth (Manhattan apt at One 57).

          8. Lawrence Brass

            Philip, you should visit us back at Economy from time to time, we chat a lot, play games, exchange food… have a lot of fun. 😉

          9. PhilipSugar

            Send me an invite at philipsugar at the gmail service.

          10. Twain Twain

            If passenger in next seat isn’t a conversationalist, I either play games, watch the movies or knit.That friend of 14 years? He thought I was a 16 yo art student because I had a journal on my lap with sketches and I was busy taking photos out the window.So he struck up a conversation. He was pretty astounded that I knew some of the same senior bankers he did and that I was, in fact, a technology banker.

          11. Chimpwithcans

            I like your story. Fewer rules on airline interactions and engagement mean more opportunity for…..interactions and engagement 🙂

        2. sigmaalgebra

          > What the f is that all about?Hype.

      3. sigmaalgebra

        > Instead of drivers meeting passengers, driverless cars will allow for strangers sharing a ride to get to know one another.All the real driverless cars will also permit 10X speed of light trips to Mars, too. With coffee served by the Tooth Fairy.There won’t be any real driverless cars for decades, at least.

    3. JLM

      .There is actually no real American automobile industry. The car makers are parts assemblers and marketing entities.They are all using the same or similar parts suppliers. This is all just an extension of the cruise control and back up camera tech — ideas which are already in cars right now.Somebody put in a GPS and now everybody has one. The GPS brains are all the same, the monitors are pretty damn close, and the only differentiation is Volvo’s new “swipe” GPS which looks like a big smartphone. Everybody else is still strangling in their underwear using buttons.Two years from today, all GPS’s will be swipe, touch monitors.When self-driving cars come, which I predict will be a long time in the future, there will be a little first mover advantage and then there will be the same standardization which has brought Bose to my pick up as the “stock” sound system.The more things change, the more they stay the same. Still, isn’t it a great time to be alive and isn’t this a great country?JLMwww.themusingsofthebigredca…

      1. William Mougayar

        So it’s very possible that the “services” companies like UBER et al. will take a front-seat to being more visible with users, and the car manufacturers will take a back-seat. Pun.

      2. JimHirshfield

        Upvote for use of “strangling in their underwear”

      3. LE

        Noting also that a point of failure in self driving cars is GPS.As far as standards will be interesting to see if “RCA Jacks” develop or how this rolls out. Between lawyers and engineers I can see that process lasting quite a long time.

        1. JLM

          .The GPS’s in cars today are two dimensional.The auto pilots and GPS’s in airplanes are slightly three dimensional. As an example, I can dial in an altitude, program the GPS, and the auto pilot will climb to that altitude on that azimuth and level off.This is what cars are going to have to do — be able to navigate in three dimensions.This will require a lot of visual clues like flying a cruise missile through a downtown which is current technology. A cruise missile can be programmed to turn a corner and approach from a specific street direction.A car will have to know how to stop and back up out of danger by itself. This takes judgment, something that half of us don’t possess right now.Then there is the skidding on icy pavements.JLMwww.themusingsofthebigredca…

          1. ErikSchwartz

            When I drive to work I park in a lot underneath the approach to the Bay Bridge. My GPS is very confused when I ask for a route home. It tells me to drive halfway across the bridge, turn around on TI and then head south.The problem is these edge cases are all very local.

          2. JLM

            .Accidents are also very local, no?People forget that GPS signals are impacted by weather. I drove home from CO through a white out and the GPS had me in the wrong place most of the trip.JLMwww.themusingsofthebigredca…

          3. LE

            And of course the software will be rock solid and not subject to bugs and problems of other types (as well as being impervious to hacking).You know the worse part of all of this? It’s not the VC money that is being wasted it’s the brain drain. All of these young people getting sucked into careers like this that have really long time frames (and I am not saying there won’t be spinoff benefits) and not going into other fields where they can be of great benefit like there was with the space program.

          4. Richard

            bit of a overstatement? Why is transportation a brain drain? Why is time frame an issue? They are getting paid.

          5. LE

            Because it’s simply not important enough (self driving cars) and to long of a time frame. Time spent in that line of work is time not spent on a host of other things that young people could be cutting their teeth on. Ditto for things like Amazon using drones to deliver packages instead of the same brains figuring out ways to lower the cost of health care (as only one example) or get people out of poverty.

          6. Richard

            Important enough, ha! Who decides? I designed advance tactical fighters at my first job out of college. Looking back I would have rather have been prototyping the next generation of surf boards.

          7. kidmercury

            this might be the biggest landslide beef in AVC history. def siding with rich here. transportation is up there with food and money in terms of important things. solving it will be totally revolutionary.

          8. sigmaalgebra

            > solving it will be totally revolutionary.Right, but with self-driving cars, essentially totally impossible. Wanting ain’t the same as getting.

          9. Twain Twain

            THIS: This is what cars are going to have to do — be able to navigate in three dimensions.Google London used to do Tech Talks where they’d invite external developers in to tell us about some of their R+D. One of their sessions back in early 2009 was on Google Maps and their then Product Lead gave the presentation (she later went on to be Product Manager for Google Wave, the predecessor to Google Plus).Anyway, after her presentation, I wrote a slide titled, “MIA in the Valley of the Kings” which was about the obvious missing 3rd dimension.The distance may be the same but driving or cycling up and down inclines versus a flat road is a very different experience and data set.

          10. JLM

            .Even moreso when there is a bit of earth beneath the conveyance. In the air, it is no so difficult because the air is filled with ………………………………………………air.JLMwww.themusingsofthebigredca…

          11. Josh

            I (humbly) think the tech that will enable driverless cars faster than we think will be the networking, not the sensors. Sensors are important but limited; they need to be augmented. Having networked cars know the path of every other car that has travelled before it allows for knowledge-sharing. It allows for creation of best paths in general, best paths for today, and best paths for this hour. It turns mindless commutes into a consumable program. If something changes to affect the route, all cars along the network are notified and aware. Perhaps 65% of our weekday routes are well-known and repeated, and should be on autopilot.In addition, oncoming cars broadcasting their position is just as revolutionary. Auto fatalities should vanish.The best marketing platform to me is the elimination of young driver fatalities and then DUI fatalities. Ending driver fatalities in general is the ideal, but the marketing pitch that strikes a chord is: “Worried parent hands keys over to child, child never comes home.” If I as a parent can eliminate that harrowing proposition, that’s a winning platform. That’s something that can emotially charge the issue and perhaps drown out the insurance lobby.It’s also got environmental legs. Eliminating traffic jams, providing greener alternatives, etc. If I can keep my F-150 parked 5 days a week and get shuttled for my 25-minute commute in a driverless Prius, that’s a huge win for me (hello productivity) and for emissions.I agee the insurance lobby is the biggest impediment. It will be interesting to see if it can be overcome.

          12. sigmaalgebra

            > A car will have to know how to stop and back up out of danger by itself.There is a great variety of such situations. Generally the result will be two biggie problems”(1) Unsure of what to do that would be safe, the self-driving car will go too slowly or just stop. People will get pissed. Laws will be passed to forbid self-driving cars.(2) There will be accidents, and the insurance companies will get pissed, the insurance rates for self-driving cars will go way up, and people won’t want to pay and will give up.Long term, self-driving cars could work under ideal conditions for 90% of car usage. The ideal conditions would require mostly all new roads.Biggie issue: Some of car driving takes actual intelligence that now is easy for humans and absolutely unavailable otherwise. Reality check here, guys.The whole pattern of computer science using words that sound like the work is on the way to matching what humans can do with human intelligence — intelligent, learning, thinking, neural — is just deliberate, deceptive hype and a misuse of words.Sure, there is software that plays one heck of a good game of chess, but that software and the computers it runs on, unless programmed carefully in advance, couldn’t find a sugar packet, tear it open and not spill the sugar, pour the sugar into some coffee, find the spoon, and stir the sugar until it is dissolved in the coffee, without spilling any of the coffee and without letting the wet spoon stain the table cloth. Bluntly, computers just are NOT anything at all like intelligent, and so far no one has even as much as a weak little hollow hint of a tiny clue how to do anything about that. Instead, computers are still carefully programmed for very specific, well defined, limited tasks well understood in advance. Occasionally computers can look a little more capable than that, but that is really just an illusion. Sorry ’bout that.So, we will not have genuinely intelligent cars for a very long time.Self-driving cars are a silly publicity stunt. For GM to spend $500 million in that direction is brain-dead. At most, self-driving will be like tail fins — a fad that maybe has to be responded to in the short term but otherwise will go the way of all fads. Besides, the tail fins were just 10-20 pounds of wasted steel — no biggie, no $500 million.IIRC, the electronic management of ignition, fuel, and transmission shifting, etc., is now an industry standard. Okay. Then let any self-driving car computing be standard. So, in that case, just wait for the standard. Offer a self-driving option for a few years, wait for the fad to die, and then f’get about it.With real experience, soon enough people won’t want to pay for the option, the extra insurance, or the maintenance and won’t want to accept the safety risks.For GM, save the $500 million.

          13. JLM

            .It will come into focus when the first self-driving car v self-driving car accident happens and the lawyers, the judge, the court, the jury is asked to apportion blame.”My car is smarter than your car.”It will happen.JLMwww.themusingsofthebigredca…

          14. sigmaalgebra


      4. SubstrateUndertow

        “When self-driving cars come, which I predict will be a long time in the future”How long in your guesstimation ?

        1. sigmaalgebra

          Until the insurance companies can be happy — a long time.

    4. Anne Libby

      Yes on revisiting. I think that the AVC crew has a very optimistic timeline on adoption for self-driving cars.

      1. pointsnfigures

        I was at an event where Mary Barra, CEO of GM, spoke last year. She said there already is driverless tech in a lot of their cars. However, she said the same thing Anne said. In order to get it to work, roads have to be redesigned, sensors and standards developed. She thought a true driverless car economy was realistically 20 years away, not 5. I know of some cool tech going into cars, so they will certainly change.Of course, great leaps in the betterment of human life happen when innovation creates new ways to communicate, and new faster ways to transport humans and goods. We are seeing that happen in real time today.

        1. LE

          Of course, great leaps in the betterment of human lifeIt’s unclear to me how this is a great leap in the betterment of human life. A great leap would be self driving helicopters maybe. A good leap would be allowing cars to communicate with each other so in bumper to bumper stop and go traffic flowed better and was less stressful. [1]Also the problem with true self driving cars is the edge cases. Simply having a solution that gets people around Manhattan (agree that’s possible) is not that big of a leap.[1] That is actually possible and in fact I took my wife’s car through the Holland Tunnel on auto cruise control (only doing steering) and it worked pretty well. Still took a great deal of attention to make sure nothing went wrong but it’s easy to see how that could be fixed and improved.

          1. Lawrence Brass

            The final solution for self driving cars has to be based on them working as a network.The demos of single self driving cars are too synthetic and isolated from reality problems such as emitter/sensor interference. Cars moving in coordinated swarms make more sense than isolated self driving cars.

        2. awaldstein

          Nicely said.

        3. Anne Libby

          And beyond the engineering problems to be solved, there’s an enormous, klugey, hyperlocal financial structure that’s dependent on (error-prone) human drivers: traffic ticket revenue, people employed in enforcement/courts, and so forth…

          1. pointsnfigures

            I am not worried about that fallout. There will be jobs created that we can’t even dream of. There will be productivity gains that we can’t measure. Most of that stuff is leaching money and productivity out of the system!

          2. Anne Libby

            That’s not fallout. It’s friction, and the reason it matters is that it’s so local.(Think about title insurance, for example — also hyperlocal, klugey and pretty resistant to “efficiencies” and productivity gain.)

          3. sigmaalgebra

            Sorry. I don’t believe that cars will change more in the next 50 years than in the last 50, which means very little, for the past 30 years, nothing significant.The main changes I’d like to see: Easier maintenance with good documentation — e.g., how to replace a burned out bulb in the instrument panel. Solve the many corrosion problems. Strong bumpers. That’s about it.The current interest in changes is to keep up with the hype from Silicon Valley, Google, Apple, the Internet, and smartphones. 100% silly talk.

          4. pointsnfigures

            I think they have changed in some pretty massive ways since 1970. My first job was selling adhesives, coatings and chemicals for 3M in 1984. Computers were first coming into cars then. The kind of paint, the kinds of gaskets used to put things together are very different. My high school had one of the best auto shop programs in the US. We used to be able to take cars in and work on them. You could change the timing, carb, all kinds of things to try and get it to go faster. Now, that sort of thing is impossible. It’s all computerized and it’s even a challenge to change your oil at home.

          5. sigmaalgebra

            I agree. I explained more in my post in this thread at…There I listed as a significant change in the last 50 years electronic controls for ignition and fuel. It’s a biggie, huge, to save spark plugs, keep from having gasoline dilute the oil and wear out the engine, e.g., from the old chokes that were a disaster, no longer need to use exhaust to heat the intake manifold and, with a still cold manifold, have a lot of raw gas dilute the oil.Then there was the big carburetor boil off: Turn the engine off, and the heat from the engine boils away the gas in the float bowl, likely leaving deposits. Then start the car again and flood the carburetor float bowl with cold gas and warp the cast aluminum.So, don’t repair a carburetor — just replace it because the old one is warped anyway.Carbs, points, centrifugal advance, etc. were grand disasters for performance, engine wear, maintenance, fuel economy, exhaust system corrosion, the air, etc.Some of the adhesives are really nice — the weather stripping on my driver’s side door came loose, and my local auto parts guys knew just the right stuff right away. It was a tube of some sticky yellow stuff, and it worked great — designed just for that purpose.In the rest of my post I mentioned hydraulic valve lifters, automatic transmissions, sealed cooling systems, much better rubber for tires, and a few more.I should have included the now quite good ball joints — my guess is that some lawyers applied pressure there.Significant changes in the past 30 years? Nothing.I used to be a car nut, was once made a Full Member of the SAE.

          6. Chimpwithcans

            +1 on making maintenance easier – My car feels like it is carefully designed to corrode, erode and implode at different, specific milestones, and i feel a bit helpless.

          7. sigmaalgebra

            My solution: Keep it all quite simple and standard, e.g., Ford or Chevy. Then have plenty of parts and mechanics. And they make them by the hundreds of thousands with relatively few changes from year to year so do relatively well getting the bugs out. Also get to use English tool dimensions and not metric.So, no Lexus, Mercedes, BMW, Volvo, Toyota, Kia, etc. Again, for a simple Ford or Chevy, my local parts store has a lot of stuff, and my local mechanic can do really well with it. Often so can I.For simple, any fancy stuff don’t get, don’t have to pay for or maintain.Also, want rugged — e.g., based on a light truck.If my startup works, then I’ll get a high end Corvette, but that would be just a toy. For actual car transportation, a Ford or Chevy SUV.

          8. Chimpwithcans

            Sounds like a good strategy. I live in South Africa where we manufacture only a few brands of cars, mostly for export. I am not even sure there are many parts available here for these exported cars – I will research which car is most straightforward and has cheapest parts in South Africa, and go from there. Thanks for the tip.

    5. awaldstein

      I’m with you on this.

    6. LE

      “change more in the next five years than it has in the last 50Dan is totally drinking the coolaid. No comparison of cars today than 50 (or even 30 years ago). And nobody in their right mind thinks that 5 years is even close to a realistic time frame. This is totally “concept car” type thinking.

      1. William Mougayar

        yup, seems like a hyperbolic statement to me. show me, don’t tell. i’m from Missouri when it comes to GM/Ford/Chrysler.

        1. JLM

          .Hyperbolic? Like hyperbole? Or is there no such word in Canadian?JLMwww.themusingsofthebigredca…

          1. JLM

            .It’s not as much fun when I have to ‘splain the jokes.JLMwww.themusingsofthebigredca…

          2. William Mougayar

            I was pulling your chain too 🙂

        2. Donna Brewington White

          “I’m from Missouri”Impressive for a Canadian. Wonder how many people got that?–Native Midwesterner

          1. William Mougayar

            haha…i have been working with Americans for a long time 🙂

      2. sigmaalgebra

        30 years? Cars have made no significant changes in the past 30 years. The only significant change in the last 50 years has been electronic controlled ignition and fuel. Before that, the significant changes were just hydraulic valve lifters, automatic transmissions, sealed cooling systems, disk brakes, better headlights, and much better rubber for the tires.

    7. ErikSchwartz

      Electric cars (with current battery tech) and ridesharing don’t go well together.It takes me 3 minutes to “load” another 300 miles of range into a car with an internal combustion engine.It takes hours to load 170 miles of range into an electric.Optimizing asset utility.

      1. William Mougayar

        but won’t that improve over time?

        1. ErikSchwartz

          Somewhat, but chemistry is not greatly impacted by Moore’s law.

        2. LE

          You would have thought that the spam problem would improve over time and quite honestly they expected cancer to be cured by now as well. Conversely when AIDS came out they didn’t even predict a cure timeline it totally confused everyone that wasn’t a rhesus monkey.

      2. LE

        Did you also notice how it’s great that electric is cheaper than gas but the people who would care about that would be those that drive a great deal and therefore have to worry the most about where they will charge up? (Even at our formerly high gas prices but even more so now..)I really question how much Fred would be in love with his Tesla if he didn’t live in the city and take cabs and the subway most of the time. It is close to a non-starter with most people who use a car everyday at least currently. You know when I come home I am not even interested in having to plug the car in every day actually.Also you left out the most important part (other than speed of refueling which is 100% valid). You know that famous web usability book “Don’t Make me Think? [1] Well one of the problems with electrics is that you have to think and plan where you are getting gas. Both because of the range and also because filling stations are not ubiquitous (or even close to it). Quite honestly that’s an extra load of thought that just makes it a non-starter for most people. And even if you find a filling station it won’t have 20 pumps it will probably have just 1 or 2. So very possible that someone else is using that. Even if they aren’t it’s like going swimming in a pool and worrying if you will have a free lane to do laps.[1]

        1. ErikSchwartz

          A friend with a Nissan Leaf offered to give me a ride home.We spent 20 minutes trying to find the recharge station near SFO and another 30 minutes waiting for enough of a charge to get us down to Los Altos.

          1. LE

            I’d be curious what your friends rationalizations are for owning a Leaf. Is it some “I’m green” party in their brain and it just makes them feel good about themselves? [1]Actually I really would like to know.[1] If so seems more practical to just use a vibrator.

          2. ErikSchwartz

            His usual commute is 5 miles. He was only in the city for some meetings.

          3. LE

            Right and to my other point. If he is only typically driving <5 miles to work then his carbon impact (or whatever) isn’t that great and his fuel costs and repair costs are trivial.

          4. ErikSchwartz

            He bought it because it’s cool.

          5. LE

            His brain thought it was cool. And sure there are those that agree with him that think the same way.Classic party in your brain.The definition of my saying (this is 100% my creation) [1]:”It’s not what people think it’s what you think they think”.[1] Me saying that is the party in my brain wanting credit for making up the saying.

          6. sigmaalgebra

            No, you left out the real problem: Your friend with the Nissan Leaf didn’t get the optional pedals.I heard that next year they will have an optional connection for a donkey.You could also get the option of the rooftop solar panels that permit speed of 0.0000001 millimeters per hour, on very smooth, perfectly level ground, with very high pressure tires, on a clear day, in the midsummer, in the tropics, without running the air conditioning, or the radio, or the car computer.Bigger problem: Didn’t realize that being Green would be so challenging.Still bigger problem: Being a sucker for Saint Laureate Al Guru’s flim-flam, fraud, quasi-religious earth-worship, guilt ridden scam. Al Guru wants sacrifices, now our gasoline powered cars, if his primitive religion goes as standard, later our virgin daughters.

        2. JLM

          .The source of the energy used to generate the electricity in the first place will be investigated and it will turn out we are using coal, in part, to drive electric cars.There is still a lot of Disney in electric cars.I still like them but we need to stop bullshitting ourselves. We may be increasing coal usage while getting away from fossil fuels? Huh?JLMwww.themusingsofthebigredca…

          1. LE

            How about if we just roll back on soccer moms schelpping their kids to afterschool activities in big ass SUV and minivans and save the environment that way?Or how about if we give cars the right of way instead of pedestrians. In my state people walk without regard to cars which have to stop instead of the people looking out for the cars and proceeding when they can. (In parking lots I am talking about of course..)

          2. JLM

            .Whole new world when gas hits $1/gallon. Will be like the 1960s again except that weed will be legal.Do not try to take an SUV from a soccer mom or a former soccer mom. Not going to happen.JLMwww.thenmusingsofthebigredc…

          3. LE

            Best part is Ivy League people (of which I am one sniff) underestimating what could be done by street smart people running oil rich countries. You know the people that you can just sit down and reason with.

          4. sigmaalgebra

            We’re not “getting away from fossil fuels”. People wised up to the totally wacko, quasi-religious, earth worship, flim-flam, fraud scam of Saint Laureate Al Guru.What sank Guru:(1) In his big movie, his big graph of temperature and CO2 concentrations of the last few hundred thousand years from the Vostok ice core data showed that the CO2 levels rose about 800 years after the temperature rose. So, the CO2 increases did not cause the temperature increases. Instead, the temperature increases caused more biological activity and, thus, the CO2 increases. What caused the temperature increases? Guru didn’t say. I doubt that he knew. Also Guru didn’t have a cause for the CO2 increases.(2) The 2006 report by the US National Academy of Sciences, that is,Committee on Surface Temperature Reconstructions for the Last 2,000 Years, National Research Council, Surface Temperature Reconstructions for the Last 2,000 Years, ISBN 0-309-66264-8, 196 pages, National Academies Press, 2006.available as PDF at…reported that the temperature in year 2006 was essentially exactly the same as in year 1000 before the earth went into The Little Ice Age. So, roughly we’ve been pulling out of The Little Ice Age.(3) Starting roughly 20 years ago, the climate science community made some predictions of the temperature of the earth between then and now. Their predictions were nearly all wildly too high. There is a graph of the many such predictions athttp://www.energyadvocate.c…In science, when make a prediction that turns out to be wildly wrong, we junk the science. Period.(4) Even over just the past 100 years, CO2 and methane changes didn’t fit with temperature changes, e.g., IIRC the global cooling from about 1940 to 1970 — below I give a reference with careful data.(5) Sure, can light a match and, thus, warm the planet, but the warming is not significant. Well, now we know that CO2 and methane at anything like reasonable levels, say, up to 10 times current levels, have no significant effect on the temperature. In simple terms, CO2 and methane just have nothing to do with the climate. Moreover, CO2 and methane from human sources are just trivial compared with natural sources, e.g., volcanoes.CO2 and methane — these are not the causes you are looking for. Nothing to see here. Move along.(6) Here is the cause: As is totally clear from observations, but for reasons we don’t understand, the activity level of the sun varies, significantly, commonly over intervals as short as just a few years. More solar activity causes more sun spots which cause more solar wind, which at the earth block more cosmic rays from hitting the atmosphere. The cosmic rays are from essentially the universe at large, e.g., supernova explosions 5 billion light years away, and likely and apparently arrive at rates that don’t change at all over millions of years. But the cosmic rays in the atmosphere cause water droplet formation which causes clouds which have a cooling effect. So, a more active sun results in fewer clouds and a warmer earth. A less active sun, e.g., in The Little Ice age, results in more clouds and a cooler earth.Details are in, right, a movie, yet another movie, this oneThe Great Global Warming Swindle,…Look, guys, I hate to pull out my dusty old prof’s hat, but we’ve just gotta learn how to evaluate evidence carefully and to do so.For Al Guru, I don’t know what’s been going on in that little space between his ears, but it looks a lot like what drove the ancient Mayan charlatans to kill people and pour their blood on a rock to keep the sun moving across the sky.Or, pick something people need, say, sex or fossil fuels, claim that at least in part it’s evil, create a pseudo-religion, ask for sacrifices, and get money and power.Guys, understatement of the millennium — we can’t let ourselves fall for a flim-flam, fraud, scam.An electric car has an advantage: As I remember from my physics course on E&M, a series wound electric motor has, from the first-cut physics, infinite torque at 0 RPM. So, an electric car can have just fantastic acceleration starting from 0 MPH. Okay. Good for golf carts, too. And some fork lift trucks?How about 18 wheel trucks? Super tough to compete with those two, big stainless steel cylinders as saddle tanks with ballpark 200 gallons of Diesel each. Some arithmetic I did, check me on this, starting with the data on the Tesla battery, says that on an 18 wheel truck, for the range of that Diesel, the battery pack would have to weigh about half the maximum weight of a fully loaded truck. And a truck stop to refuel such trucks quickly would need nearly its own electric generating plant.Net, for an energy source for cars, super tough to beat a 20 gallon tank of gasoline. Similarly for Diesel for trucks. If someday we can’t get that gasoline from fossil fuels, then we should get it elsewhere, e.g., from electric power, coal, and water, and get the electric power from nukes. Or, maybe get the electric power from solar cells in the desert, ship the water and coal to there, and put the gasoline into a pipeline back to real people with real cars. I’d bet on the nukes.The brain-dead train wreck of the claims of Human caused global warming. What a flim-flam, fraud, scam. What a waste — billions of dollars just wasted on total nonsense for no good reason.

        3. Susan Rubinsky

          There weren’t many gas stations around either, until Ford started mass-producing.

      3. Susan Rubinsky

        Tesla has new technology that does it in less than an hour. I suspect that will improve over time.

    8. Richard

      This feels like a act of desperation or tax strategy by GM.

      1. William Mougayar

        that’s what i thought. if it was accompanied by something to show, it would be better.

        1. Richard

          GMs only good decision over the last 50 years has been its investment in china. I haven’t looked at their financial statement in a while. Do they have a half billion on their balance sheet?

          1. JLM

            .It was GM’s management that created the auto crisis of the first Obama administration. They pandered to the unions, collected their options, and wrecked the company and the industry.They are smarter now, why?Go see The Big Short and know that the subprime auto loan market is huge and getting ready to go the same path. Here is a review on the movie.http://themusingsofthebigre…JLMwww.themusingsofthebigredca…

          2. Richard

            I was ankle deep (not knee) in the inner workings of pricing algos of credit default swaps of the subprime housing dabacle. Suprime auto loans are maybe 100billion. Not sure how many CDS have been sold against this, but I doubt it’s anywhere near subprime housing.

          3. JLM

            .Housing — 100 kilotonsCars — 4 kilotonsStill, an explosion.Keep one’s eyes on student loans. This is a credit failure. Nobody wants to pay them off though everyone wanted them.President Bernie Sanders will sort this out retroactively.JLMwww.themusingsofthebigredca…

          4. Richard

            Auto loans are small potatoes. Student loans, that’s another issue. But most are federally insured and most importantly not eligible for bankruptcy which just effective means that the IRS will be the final debt collector for perpetuity for a lot of people

          5. JLM

            .The IRS will be disbanded and replaced by Wm M and the blockchain in a Trump administration, no?There are a huge number of auto loans out there. Dollar volume lower but the number is huge.JLMwww.themusingsofthebigredca…

          6. William Mougayar

            haha…that’s too funny.

          7. Lawrence Brass

            My AI based site crawler has just issued a warning:webcraw v 0.3: Warning: JLM forecast for winner candidate switch detected.The ironic preprocessor module logic is deeply flawed, so I will tag this as a false positive.

      2. Salt Shaker

        Not at all. The auto lease angle alone could be a decent rev driver for GM. The driverless angle, at least at this point in time, is just PR spin IMO. For Lyft, a large investment by GM pre-IPO helps legitimize the company’s valuation. Gives them more credibility. Win-win.

    9. SubstrateUndertow

      US electrical production 2014Coal = 39%Natural gas = 27%

  5. LIAD

    Analogy would work for Amazon too.Perhaps also: – Telcos leveraging phone customers to become leading wireless broadband providers – Banks leveraging physical branches then switching down the road to online only – Nintendo building huge live multiplayer game subscription business off analog disk sales

    1. fredwilson


    2. Rob Larson

      so far, zero banks have made that switch, amazingly. The few banks that are online-only started that way, and never had branches. Traditional banks are at least building fewer branches now, which is a step in the right direction, but none have abandoned their branches. I agree with you that eventually it will happen, “down the road,” but that is a verrrrrrrry slow industry when it comes to embracing change.

      1. LIAD

        i’ve read reports banks are closing branches at a accelerating pace.

        1. Rob Larson

          In aggregate, that’s true. They’re also building new branches. But currently they are closing more than they’re opening. Generally the closures consist of 1- thinning the least profitable locations among their branch networks (a continual exercise) and 2-some small banks going out of business.

  6. Ben Kinnard

    I’m not sure how the business is split % wise nowadays, but the Financial Times has done a good job of going digital first (e.g. there is no print deadline now, they just put in the paper the latest stuff from the website) and is one of the few papers succeeding with its paywall – although it obviously helps if you have corporate/wealthy clients and a 125 year history of excellence

  7. Kate Huyett

    Fantastic podcast episode on Netflix’s strategy here, including how they thought about this transition:

    1. Anne Libby

      Also, Reed Hastings addressed the shift from disc to streaming at Stanford:…(h/t Gary Chou.)

      1. Kirsten Lambertsen

        Wow, this is a great share, Anne.

      2. fredwilson


  8. Twain Twain

    Go LYFT.

  9. Eric VanderSchaaf

    Hey Fred, check out the book Netflixed – I read it and it was really good. Some friends of mine alerted me to the book (friends that used to work at Blockbuster during the whole war!). Enjoy and happy 2016!Link:

  10. Eric Friedman

    I am re-reading the Innovators Dilemma right now, and this seems to use the books framework to invest in the disruption happening to their business. Even if its the wrong horse in the race, their move to do this seems a positive for the Company.

  11. TeddyBeingTeddy

    Great hedge by GM, and nice move by Lyft. if you were Uber, how would you respond? If I were Travis, i’d buy a ton of Google stock under a cover name, buy some way out of the money GM call options as far out as possible, hire the former heads of every big city’s transit authority, buy a mapping biz (already done?), partner with Dominos to start by delivering hyper local pizzas with automated cars, and give employees huge bounties/bonuses for cracking various milestones needed to win the race. Google will still win, and Uber’s only hope is to buy Elon Musk, which won’t happen.Google Maps + search + Boston Dynamics + Google Glass + VR = unfair competitive advantage.

    1. Richard

      Buy GM calls?

      1. TeddyBeingTeddy

        I wouldn’t want to own GM right now, as this might be setting money on fire for a big slow company temporarily doing well. But I’d want to own the option to own them way down the line in the off chance they are one of the victors. Hence buying call options. Likely not a popular pick, but probably merits consideration.

  12. btrautsc

    Great comparison.Will be fascinating to see the businesses that carve major niches. Who would have though 7 years ago that we would get 8-12 episode series (really “movies”) with major budgets that launch “fully available”, directly to consumers’ living rooms.

  13. Stephen Bradley

    At NPD (entertainment market research) we worked directly with Reid from the early days of Netflix and I can say with reasonable clarity (short of climbing into his head) that modern-day Netflix was never part of the original plan, at least beyond “someday maybe” status. The original plan was fully focused on disrupting Blockbuster et al, which was progressing swimmingly until the disruptor ran into a disruptor of it’s own called Red Box. Red Box came out of the gate fast and started capturing market share at a phenomenal pace, well outpacing Netflix, and Netflix was in real trouble. This is where I think Reid was most fearless and brilliant. He changed the model and the playing field completely, and at a blinding pace, in my view. Reid was anything but Wall St’s darling at the time, and it was a hell of a big bet. He bet the business and his own career on the mega-move, and it has clearly paid off.

    1. fredwilson

      thanks. super helpful.

  14. Tom Labus

    Sme analysts are saying that this is the best shape the auto industry has EVER been in. I’m surprised it wasn’t Ford.

    1. pointsnfigures

      Ford is in pretty good shape I think and also working feverishly on innovation.

  15. awaldstein

    I participated personally in the move from analog to digital in the movie and cinema worlds.Certainly I bolted on top of the roll in of digital projector to launch RLD but the larger movement that enabled 3d was analog to digital in every aspect of the Hollywood ecosytem.3d was a subtext to a huge industry shift that of course led the way to streaming as well.

  16. Stephen Bradley

    Intrigued but I could use more clarity around your definition of “analog” and “digital.” What qualifies as an “analog” asset that can be digitized (or digitalized)… beyond media, of course? In your Lyft example, are humans (drivers) the analog asset? Apologize if I’m being dense…

  17. Dana Hoffer

    The GRANDADDY of them all, in my view, would be Computing Tabulating Recording, a maker of typewriters and calculators evolving to the International Business Machines Corporation and proceeding to set the standard by which computing evolved from the 60’s for nearly four decades, in the process becoming one of the best examples of how to run a corporation and lead a market. Alas, even the best lose their place at the top of the leader board (Fred – per your comment about what is ahead in 2016), Apple has also done this analog to digital transition MP4 player –> iTunes electronic music/Podcast/video/TV delivery,Others qualify, also in my view, GOOGLE, a search engine, now an digital ad leader, mobile OS, and with GOOGLE INDEX, a player for the coordination of Android apps in your life.One other may be AmEx which began as a paper “on account” tracking system (memory aid help here please) to be the first standalone electronic payment network.This is an excellent way to teach business students about real value enhancement and Netflix is a solid example of switching before the option to switch has past. Can’t wait to see how Twitter evolves.

    1. LE

      Except that Google didn’t make money from that search engine until they started selling ads. Netflix (in theory) made money from DVD’s prior to digital delivery.This is an excellent way to teach business students about real value enhancementBusiness school isn’t or should be like art history or philosophy or american history. It should be about learning skills that can help you make money. Not concepts that quite honestly don’t really matter to 99.9% of people.Entrepreneurially better to think in the present as far as a niche that you can fill. The discussion today is purely academic and bears little relevance to any particular individual trying to actually earn a living. Would also think that the majority of people that are working for someone else (in a corporation or otherwise) don’t have to give any thought or effort to these type of things as it’s not part of their job and isn’t going to land them a promotion (unless it’s part of their job – nice work if you can get it I guess). Good academic stuff though or for bloggers or a cheezy and poorly done report (they typically are) on CNBC hosted by Carl Quintinella.

      1. Dana Hoffer

        Hi LE,Regarding Google… while I am not entirely sure about the timing, Google did tie up with AOL for awhile around search technology. It may have been a straight rev share from advertising and you’d be right but I recall it was about their tech.Regarding B-school… On your art history point… we have a honest difference of views on this. While I could teach (baseball) batting skills without pointing to great examples like Ted Williams and let them understand what has worked in the past so they need not “discover” it potentially discovering way too late.On the point “value for individuals actually earning a living” … consider this… everyday tens of millions go to work and their managers would thoroughly appreciate well informed, insightful and helpful recommendations about how to make the business more viable and more profitable. It happens too infrequently. The best way to earn a living is to know how to spot and how to recommend improvements like Fred is pointing to here. And if the person is truly right and the organization is incalcitrant, then that person has found the opportunity to go do it as an entrepreneur.

  18. Christopher Fong

    Hi Everyone,I’ve been a member of this community for almost a decade. It’s my first post but I’ve been to different community events including the fundraiser at Houston Hall. It’s this very concern regarding disrupting many individual’s livelihood and incomes that motivated me to leave Google after 8.5 years.I looked at optimizing drivers’ pay across on demand apps e.g. Driving for Uber and Lyft during their peak hours then Instacart and Postmates during theirs and after working with a few hundred drivers, I realized the only way to help maximize the income of these folks is by providing skills and jobs in growing sectors of the economy. The first field I’ve chosen is in Digital Marketing where I’m providing training and job placements for those that have have absolutely no background in this area but has the motivation and ambition to learn.If there are other concerned individuals that would like to work with us to help create a job training and placement platform for growing industries to get ahead of the disruption Fred mentions I’d love to discuss offline and see if we can come up with other solutions. Here’s my post on the launch and the reasons behind it:….Fred – I know you don’t want your blog to be a promotion platform but thought this was a relevant mention and I’m looking to partner with other individuals that are a part of your community.Cheers

    1. Donna Brewington White

      Training and job placement is a powerful combination. I work with more senior and executive level hires but a lot of my work involves filling digital marketing openings and I see this increasing significantly in 2016. Let me know if I can be of help with this worthwhile initiative. Much success to you!

      1. Christopher Fong

        Hi @donnawhite:disqus, thanks for confirming our research and offering your partnership and help! I love the AVC community. There were over 70,000 digital marketing job openings on, these and other jobs can only be filled via training additional professionals in this field. I’d love to chat and see if you might be interested in placing the candidates after they go through our job training platform.

        1. Donna Brewington White

          Hi Christopher — My work is with more senior level marketers but I will often be aware of openings at my client companies even though not directly working to fill them — and when I place a senior level marketer that person is often charged with building a team. So, yes, let’s chat. Plus I believe in the kind of thing you are doing. Can you reach out on LinkedIn and we’ll go from there:

      2. Richard

        Do you candidly feel a digital marketing job placement program can make somebody job ready?

        1. Donna Brewington White

          In general, I think that training people to provide new skills and then helping them to find jobs is a great thing. Christopher’s post on LinkedIn indicates that they already have interested employers.My perspective is biased because I work with more senior level hires and the people coming out of Christopher’s program will probably not be candidates I can place. I am also working, primarily, with funded startups that can compete for top talent and even these companies are having a difficult time.So my best answer to your question given the limitations of my perspective is that what Christopher is doing will be beneficial given that it is focused on entry level hires.1. As the market becomes even tighter, companies will become more flexible in their requirements and the entry points will increase, possibly more ramp up time allowed.2. It also depends on what is being referenced as digital marketing. If it is specific aspects of execution under the guidance of a more senior marketer who is then freed up to focus more on strategy then yes.Also, there is a lot of opportunity for freelance work for small businesses.Your thoughts?

          1. awaldstein

            DunnoI do see a change happening in the recruiting space that is making peer networks in some ways where recruiting is evolving to at the sub senior level.People at mid to senior levels move often in the Google/Facebook ecosystem and carry with them massive networks and a significant amount of hiring experience.There are a number of people who are going in house as long term contractors not just to hire but to actually design teams, recruit and then focus on making the hires successful.Kinda opened my eyes as I’ve run into this.Seems like this will continue.

    2. JoeK

      Shameless plug Christopher.You clearly don’t know this community well enough.

    3. fredwilson

      promote away! i appreciate and applaud what you are working on and why

      1. Christopher Fong

        Thank you @fredwilson:disqus, I appreciate your continued encouragement as well as Nick and Albert’s! Kirsten and Kevin, thanks for your support. Are you ever in SF? If not, it would be great to meet and get your thoughts in person or perhaps during my next visit to NY.

  19. kirklove

    Is this really an analog to digital move? Still need lots and lots of physical (analog) cars.

    1. Anne Libby

      And real life people to say “yes” to the shift.

    2. fredwilson

      well netflix still has servers. but i do get your point.

  20. LE

    But using an analog asset to build a large customer base that can then be leveraged in a native digital model is a great business move and I am surprised it has not been done in more markets over the past couple decades.Why? Because in order to do that you have to see the writing on the wall in your business. Otherwise you are to comfortable and secure to make a move like that. For Netflix it was both easy and necessary since they saw how things would play out. With self driving cars it’s nowhere near as clear and actually not even close.

  21. mikenolan99

    Nice insight – I’m prepping to teach a round of Entrepreneurship class, and found this figure in Steve Blank’s The Startup Owners Manual:It shows four quadrants of possibility – Physical and Virtual Products vs Virtual and Physical Channels. Cool to think about the pivots between these quadrants – and how many of the big players are innovating in all areas.

  22. pointsnfigures

    Have to think about how this happens in health care.

    1. fredwilson


  23. ErikSchwartz

    Self driving features on cars are coming soon.Autonomous cars on public roads are not coming soon.

  24. Salt Shaker

    Too much of this alliance is focusing on driverless cars when the very real and immediate benefit for GM is by creating a lease agreement w/ Lyft and its drivers. That’s money in the bank, while driverless is still prospecting for gold, and antithetical to Lyft’s current biz model.”America’s largest automaker will offer Lyft drivers vehicles for short-term rent through various hubs in U.S. cities, the companies said in separate statements on Monday.”

    1. LE

      Exactly. Similar to fleet sales to Hertz Avis et al.

      1. Salt Shaker

        Better lease terms w/ GM can give Lyft drivers an edge against Uber, although this agreement can presumably be easily emulated w/ another auto man. Lyft/Uber are aggressively competing for drivers as much as they are riders. It’s a good move on GM and Uber’s part, but the notion this is about driverless, which is nothing more than a long term hedge bet, seems vastly premature from a practical standpoint. That said, it does give Lyft enormous credibility for an IPO move. It helps legitimize its valuation.

        1. LE

          GM just wants to be able to send a message to investors that they are not asleep at the switch and that they are moving and grooving with the times ie “I talk jive”.Not to mention that GM is the type of car builders that saves 5 cents on a part and then cleans up the mess afterwords when people die or have problems.

  25. Kirsten Lambertsen

    We all knew as early as 1996/97 that some day streaming video would be the default. My first job in tech was breaking “Metropolis” into 2 minute chunks for people to stream for free online. If you bought the DVD from us, you’d get access to watch it (in a truly crappy experience) online for free.When Netflix launched, as a former home video executive, to me it was an obvious move. The amazing part was that someone finally got enough studios/labels on board to cooperate. Hastings was the guy who finally worked *that* miracle. But we all always felt that for the *consumer* this was how it *should* work. We also knew that it was an interim step to streaming until the bandwidth could catch up to the vision.The motto of the home video industry (when I was in it) was “watch what you want, when you want” (don’t know if it still is the motto). With that as a compass, it was always clear that streaming and subscription would be the way it should go.

  26. Arthur Schnitzler

    From what I have heard, downloads were the original Strategy for the shift to digital until streaming was possible and they switched. This was in the late 90s mind you.

  27. Brandon Burns

    California also just passed a law a couple weeks ago stating that driverless cars actually have to have a human driver in the car ready to take over the wheel when necessary.The tech world loves the “software is eating the world” mantra, and the assumption that technology will replace humans. News flash: no it won’t.Humans are humans, and the world is the world. People will never allow a world to develop where people themselves are rendered obsolete. And if you were the legislator writing the laws, or the average voter voting to keep your relevance as an individual, neither would you.Analog may evolve, but the car is the asset being changed, to get more out of each car’s idle time — we’re not seeing a decrease in drivers, but rather an increase since Uber and Lyft have provided more jobs for them.

    1. LE

      California also just passed a law a couple weeks ago stating that driverless cars actually have to have a human driver in the car ready to take over the wheel when necessary.Well in theory if there is enough money involved they could help get legislators elected that pushed the agenda in the direction of no human driver. I don’t see a revolt by the people either doesn’t fit the pattern for that type of thing.For the record I think the idea of having (except in super limited situations) totally self driving cars is 100% absurd. The world is not Disneyland or WDW where things can be controlled this is all a pipe dream try driving around a mall parking lot during Christmas shopping (I don’t).

      1. Richard

        I totally can see highways lanes designated as being selfdiving It would make highways safer and more efficient.

        1. LE

          That’s limited use and I agree with that. However that is an edge case that could actually be pulled off.

        2. Salt Shaker

          But one needs to get to the highway first. Perhaps park ‘n ride is a good use case (e.g., driverless buses).

          1. Richard

            For we only get to driverless highways it will be a huge win

      2. Brandon Burns

        I think the sentiment that the biggest innovators were at some point called crazy gives folks the balls to say absurd shit like “soon there will be no drivers!” and then use the advent of the computer or the car replacing the horse and buggy or something to “prove” that their craziness is right.But these people are pundits, not true inventors. A true inventor understands the one constant that will never change is human behavior, and they understand how their products fit into it. And they know that humans in government will only make Uber lead to more jobs, not fewer, make AirBnB eventually lead to a better financial situation for home and lease owners, not a worse one, and so on.Sometimes we get so excited with where the future is going that we forget about common sense.

        1. LE

          Yeah people (simpletons) just love talking about how Woz tried to sell HP on his stupid little wooden box computer but they were to stupid to see the potential. Problem is he was guy number 10,000 that they rejected the rest of the ideas didn’t go anywhere.

        2. Twain Twain

          COMMON SENSE is an intelligence we have that the machines don’t:* http://www.technologyreview…*…So even before we get to @ErikSchwartz:disqus & @bfromkansas:disqus debate on whether a morality algorithm is needed to decide which life is more valuable (nun vs child) and @domainregistry:disqus ‘s insurance problem…The machines have to be able to deal in 3D as @JLM:disqus points out and be able to detect and calculate the likelihood of some random item is falling on the car from above and sudden potholes in the road it needs to swerve.All COMMON SENSE which you, JLM and others here have aplenty.

    2. ErikSchwartz

      Who writes the morality algorithm?Is it better to run over 2 adult males than a single child?Is it better to sacrifice the driver of the car than to hit a nun? What if the driver is pregnant?All of that would need to get codified into a very public (and probably written into the vehicle code) algorithm.

      1. LE

        Exactly. Another component that people don’t talk about is how auto insurers will deal with this type of dilemma.

        1. ErikSchwartz

          The insurance rates that will go way up are the manufacturers of the cars. Right now operating liability is on the driver. When the car drives itself operating liability is on the manufacturer.

          1. LE

            Essentially any PI or class action case is limited by the greed of the attorneys who handle the case.Here is an example. On most policies you have what 1 or 2 million dollar limits on liability? That is the low hanging fruit of opportunity for most attorneys in most situations. They can make a quick buck by maxing out on that without a great deal of work. So it often doesn’t pay to go for personal assets which could take years to get. If the policies had 5 million limits then they would go for that (and settle for less).Anyway my point is that in the case of a class action you have to have a class (for big dollars) but in the case of a person here or there getting killed that’s a million here or there because no attorney is going to want to wait for 50 million (statistically) they will go for the lower amount that is the sure bet (and move onto the next case). As such I don’t see the risk on auto companies as great as you do.Likewise auto companies are good at dragging out these cases for a long time and surviving. It’s not like people haven’t been killed because of faults in non self driving cars.I was looking at the cost of airplanes in the olden days vs. today. I would suspect (and JLM can confirm) that the main driver of cost increases has been legal and insurance related to manufacturing. So you could be right but I think cars are different simply because there are so many of them vs. airplanes which don’t crash as frequently (as odd as that sounds).

          2. JoeK

            They will drop significantly. Look at the causes of road accidents – speeding, failure to follow traffic rules, drunk driving etc. Automotive or software defects are a very minor fraction of all this. And will be even smaller once manufacturers can code rules in their cars to the effect that improperly maintained vehicles are disabled automatically.

          3. ErikSchwartz

            The insurance rates for owners will go way down.The insurance rates for manufacturers will go way up.

          4. JoeK

            I disagree with you on ‘way up’. Google’s prototype cars have had no at fault accidents in more than 2 million miles of driving. When in doubt, bet on technology.

      2. Brandon Burns

        “But we’re all gonna be the Jetsons, man! That who dies first stuff will all just get figured out and we’ll be driven through space and time at the click of a button!”Lol.

      3. JoeK

        You don’t need a morality algorithm. Why should the car have to run over anyone? Have you recently had to choose between a nun and a child?

        1. ErikSchwartz

          Unless it’s driving under 10 MPH all the time then it’s going to have to make those decisions at some point.Kid steps out on a scooter from between two parked cars. There’s a car coming in the other lane. What does the software tell the machine to do?

          1. JoeK


          2. ErikSchwartz

            So you’re in the stay under 10 mph camp?A car (self driving or not) can only stop so fast. F=MA is also not really greatly impacted by Moore’s law.

          3. JoeK

            No, I mean that the algorithm is simple. Brake.What would you do?

          4. ErikSchwartz

            But it may be more moral to swerve and kill the occupant of the car.Will people be comfortable riding in a car that under certain conditions is programmed to intentionally kill them?

          5. JoeK

            That’s a moral decision, which you do not want anyone coding for. What’s next, kill the Republican over the Democrat?

          6. ErikSchwartz

            So your algorithm is protect the life of the passenger of the car at all costs?

          7. JoeK

            That’s not what I am saying.The decision process does not have to involve ‘moral’ decisions. We have guidelines in place for choosing recipients of heart transplants as an example, which is a much more tricky decision to make.With cars, minimize the risk of collision, without doing anything that would lose you control of the car – the sole exception being when the car is empty, in which case it is acceptable to destroy it.

          8. ErikSchwartz

            See child to close too brake for.optionsbrake but still hit childswerve into clear laneswerve into parked carswerve into oncoming car (occupied)swerve into oncoming car (unoccupied)Further complicated by the number of people in the computer controlled car. It should make a different decision if the car is empty than with a single passenger, and probably other decisions if there are multiple passengers.Those options need to get codified into an algorithm. That algo needs to be public, the same for all automakers and probably included in the vehicle code. Call it a morality algorithm or don’t. But that is what it is.

          9. JoeK

            Little children do not spontaneously appear. If they did, your solution list would need to include the case of see another child too close to brake for in the middle of swerving to avoid the first child who was too close to brake for.Automated cars will not be required (in the first instance) to do anything a human driver would not be expected to do. Once they are prevalent, these niche scenarios you describe will be non-existent, every car on that road will coordinate to minimize the risk of a collision, for example by the parked cars verbally telling the child to stop as another car is coming.But as far as algorithms go, the reason you don’t blindly swerve is that you don’t know what chain reaction you will trigger. Look at current driver instruction manuals and guides – you won’t see collision target selection rules that reference anything other than minimizing risks to yourself.

          10. ErikSchwartz

            Pedestrians step out from in between parked cars all the time. As far as the sensors on the car are concerned, that is spontaneously appearing.But this is all 20 years off so while it’s an interesting discussion, it is largely a moot one.

        1. Stephen Voris

          Well, there’s certainly a decent chunk of reading material out there on the human-morality end of things – The Righteous Mind (book) and Slate Star Codex (blog) come to mind for quality contemporary thinking on the matter.N-thousand-foot view, if it were up to me I’d probably break it down into something like a “moral action dictionary” paired with detection algorithms for morally-important situations. Bit of fractal detail there – “murder is bad” in the aforementioned dictionary also has to be backed up by “killing a person counts as murder unless it was a mistake” and the various definitions for ‘killing’, ‘person’, and ‘mistake’, the sensors to actually detect people/morally-relevant-entities… and likely would have to have different values for different cultures.

          1. Twain Twain

            Google, FB, IBM Watson, Baidu et al have accumulated the largest data corpuses possible (as Fred pointed out recently) and applied every algorithm known to try to understand that data. That includes every book published since Gutenberg because Google made a special effort of digitizing that analogue treasure.Yet … for all that data and all that Maths thrown at it … not a single one of those techco’s can solve the persistent problem in AI of language understanding (pls see slides).How ethical+moral and capable of understanding us are the machines? Well … how ethical+moral and capable of understanding us was the AI that ran the trading and risk management systems in the banks that led to the global financial crisis of 2008/9?Because those are the same algorithmic approaches and flaws that are now emerging in consumer AI. The network multiplier effect means that any algorithm flaws will get bigger, faster.The BIGGEST OPPORTUNITY for Analogue => Digital is if we were somehow able to take our analogue human intelligence and make the digital machines more intelligent like us (moral+ethical, language coherent, considerate etc).Now, I’m part pragmatist-part philosopher, mostly problem-solver.It’s not going to be solved today, tomorrow or the next day …But maybe when Fred looks back in 20 years’ time to this ‘Analog to Digital Move’ post, we’ll have progressed some way closer to solving the problems in these slides.@JLM:disqus @domainregistry:disqus @samedaydr:disqus @pointsnfigures:disqus @donnawhite:disqus @MsPseudolus:disqus @lawrencebrass:disqus @philipsugar:disqus @wmoug:disqus @disqus_Awy3Cl8ObF:disqus @JimHirshfield:disqus @kwiqly:disqus @TomLabus:disqus @ErikSchwartz:disqus @annelibby:disqus @aexm:disqus @sigmaalgebra:disqus — Keep AVC buzzing with brilliance and LOLs whilst I’m away …There are some upcoming AI / Deep Learning events where I’ll need to show the great+good giants of the sector (Stanford Professors, Chief Scientists of all the major techcos, AI investors) that there IS a way forward for the problems that have befuddled us for a few millennia related to data and intelligence, and how we measure, model and gain meaning from it.So I need to work towards that and not play on AVC.Yes and if LE sat next to me on a plane whilst I was knitting and ignored me, he’d never know the “mad scientist” things being knitted in my mind that then get knitted in code by my hands.LOL.

      4. kidmercury

        gotta factor in occupation into the morality algorithm. when unable to determine via facial recognition, should use other proxies.time to start on the morality as a service API!

    3. fredwilson

      i hope you are right

  28. David LoVerme

    John Gallaugher (Boston College IS Professor) talks a lot about Netflix and this shift in his textbook I don’t recall if he touches on intentionality but I know lots of professors are teaching his book so I’d expect more young entrepreneurs to use try this approach in the near future.

  29. Salt Shaker

    Meanwhile, in the world of reality, the markets today are tanking in response to China instability. Now that’s something in need of a real Lyft. Wtf, I may actually buy a chunk of Alibaba 🙂

  30. jason wright

    on the theme of prediction.i predict that car manufacturers will stop selling their cars.

  31. Collin

    The analog to digital transition is the most exciting thing to me. Every facet of human life will, in varying degrees of completeness, shift from analog to digital. We will interface with physical reality through a digital layer, but every business, government, and non-profit will be digital.Inefficiencies are a result of the analog world and more of life as we know it will move digital thanks to greater computing power and as a result of the search for efficiency. It’s going to be awesome.

    1. SubstrateUndertow

      Isomorphicly minimalist digital-abstaction is the new visceral efficiency glue !

  32. Josh Davis

    The home improvement industry will fully adapt to digital in the next ten years. Here is a quote from May 2015. “This transition from the analog process to that of the mobile world, no doubt, changes project management as we know it. The ability for instant communication and convenient mobile tools is beginning to revolutionize how we define a “contractor,” today. Recreating the role and placing the menial jobs in the hands of technology, it is the crews and skilled laborers now driving the man-power, focusing on what is truly important: best practices and quality work.”

    1. JLM

      .This is not even close to being new.I built high rise office buildings in the 1980s where all the scheduling — PERT, CPM charts — were completely digitized and the entire financial model was computerized.The productivity of iron workers tying rebar and concrete finishers finishing concrete was crunched every night to two decimal places and compared to the original estimate which was also digitized.The design of the structure was completely computer based using Strudl. Now you can get such software for free. It has been tied for years into reinforcing design, cut sheets for fabrication, and pick sheets for delivery. It even produces tags for the ironworkers to find on the job site.A thousand identical reinforcing bars each have a single, unique location to be installed.You know who really knows the stuff? Donald Trump.JLMwww.themusingsofthebigredca…

      1. Josh Davis

        The quote from above is in regards to the Home Improvement Industry. Most contractors in the Home Improvement Industy have very poor systems. Companies like my company Viirt, are building vertical platforms for a specific trade, we then operate a step by step process that manages between installers and homeowners, automating much of the communication and promting the rest. All while measuring the homeowners experience and educating the installers on what they can do better. – What you are saying is like saying; there has been transportation software for bus fleets for decades, lyft and GM are dumb.

        1. JLM

          .Not at all. If you can figure out how to build 50 story buildings, then you can marshal those systems to build homes, as an example.You go to the moon first, you invent GPS, you develop the FAA and then every car in America has a GPS.Most home contractors are craftsmen first and businessmen second.This is why if you understand construction you are not afraid to deal with pick up truck contractors.I had my floors refinished. High bid $16K. Low bid $3K. Not sure they were all on the same page. Big contractor has a website, a quality control staff, a writen warranty. Low bidder does better work and operates out of a van and a cell phone.I used the little guy but, of course, I had built high rises and renovated office buildings. Favorite rehabs attached.Nobody these days is dumb. The right amount of tech can triple the profits for a small contractor who is just trying to make a living.JLMwww.themusingsofthebigredca…

          1. Chimpwithcans

            Hi JLM, I got a query for you – how would you get a contractor to finish the work first, then get paid? We were recently ripped off by a fence/balustrade man who ran off after a sizable down payment was made – I now don’t want to make down payments. Any recommendations? Any useful guidance for how much of a down payment to make which suits both parties? Happy New Years.

          2. Josh Davis

            You and I are in two different worlds. I know what you are saying though. Beautiful building.

  33. kevando

    Great song, too :)…

  34. Ana Milicevic

    Netflix’s move was brilliant and it created the whole OTT premium video category. Not only that, but their content strategy in the early days was double-brilliant: they focused on niche content (e.g. European art house flicks, anime) that was comparably easy and cheap to source and had avid followings (coming just short of mainstream).As a digital native I readily see opportunities for traditional businesses to improve by adopting some aspects of the digital realm. The reason this is neither easy nor obvious to non-digital folks is that in most cases the current way of doing things meets the good enough muster. I was surprised at the extent this type of (to me) legacy thinking was present and prevalent in some of the world’s largest software companies — those who have in the bast been considered innovators and trailblazers. Other industries have an even bigger hurdle.

  35. Kim

    Great post! Here’s an interesting read on Netflix by the way, from a former employee:

  36. chadbyers

    I was lucky enough to see Reed Hastings speak a coupe months back. He said streaming was always the goal from the start, but the tech wasn’t ready yet. In my opinion, having that foresight and vision is more impressive than if they figured it out as they went along (which no doubt would have been amazing as well). He was a wonderful speaker and an even better person from the sounds of his charitable work.

  37. Phil Hayes-St Clair

    AirShr is starting to bring about this move in terrestrial radio. Early days with positive signs.

  38. Simone

    Reading today this blog post and this news http://uk.businessinsider.c…, I was thinking that sooner or later technological unemployment will become the important topic that it should be. Nice PR move for both GM and Lyft, but is it going to deliver medium/long term and for whom.Love Netflix and Amazon and looking fwd for more and more content to be available for instant streaming/subscription models etc.

  39. James Ferguson @kWIQly

    As soon as we consider irrational numbers eg PI 3.142… or the ratio of a circle to its diameter we see digital solutions have an inherent problem. They are at best an approximate. And these problems are common (toss in e and root 2 ).Perfect is often the enemy of the good. So digital can beat analog (in domains where a poor substitute is adequate)Consider the earliest known digital networked telegraph of a facsimile image circa 1846…The images though of poor resolution (not a technical limit – but a time limit) , but made interesting precursors to the .png (or indeed the internet)This illustrates that analog/digital change is (always I think) limited by technological change influencing quality vs speed trade offsKodak Film -> digital photography is another case in point (and is a great counter case to @fredwilson point ) where kodak dropped the ball big time.>> “using an analog asset to build a large customer base that can then be leveraged in a native digital model is a great business move”So what comes next – comfort – or is this a trick question ?Comfort (in the sense of temperature and humidity) are analog inputs to analog systems (humans) which we digitise with thermostats to set thresholds or promote action (digital decision – putting a wooly jumper on).HOWEVER – We merely use an anolog of a more precise digital model. Temperature depends on counts (integers) that transfer energy in quanta (or dont) so temperature IS inherently digital at small scale – it only first becomes analog as we step away from it using probabilistic physical models .Stepping away turns our pixelated circle into an analogue (in our brains) represented by digital firing of retina neuron paths and modeled.So (hesitantly) I argue digital and analog are the same except for scale.Start with something that doesnt scale (analog) understand it (digital automation perhaps a state engine or a RESTful service) , justify it (analogue risk assessment) then scale it digitally until so analogue metrics work again CAC/LTV and cohort analysis models

  40. Johan Brenner

    Square and credit cards will undergo same shift. Building merchant user base and trying to create consumer network as well. Not alone though.

  41. Humberto

    Personally, this GM move seems to me like an awful bad idea. If they’re investing for the technology, they might have done better acquiring a whole university team doing driverless car research; Team Caltech comes to mind for unnamed cars; for pure lyft “sharing” algorithm there’s a million options. if they’re investing for the business, they are overpaying for something which is still a big bet. Bmw/ sixt did a much smarter move by launching DriveNow. Can be extended for AI quite well; it’s an actual, real business (already making a profit I imagine) and quite scalable geographically, temporally and technological.

  42. JLM

    .Very interesting comment.Car design is the “best thinking” that is available to any car company. No design can touch the Ford Mustang for its business significance. The people involved were the best.I will be checking those wheel wells.I never parked cars but I washed plenty. When I had big parking garages I offered car washes and oil changes and tire rotations.One of the guys, Jimmy, cleaned golf clubs and shoes. He made more than $100K one year. I always remembered that.I thought I was the entrepreneur. No, it was Jimmy.JLMwww.themusingsofthebigredca…

  43. SubstrateUndertow

    Bingo !Everything that exist in abstract-space (digital-space) is meaningless unless it can be effectively/consistently mapped back onto the its tangible corresponding physical/visceral equivalencies.(including abstract assets)That is the core existential reality of Mind Over Mater be it human mind or machine mind.That abstract/digital to physical/viscera isomorphic transposition is the core generic concept animating stability at all strata of “living adaptive systems” all the way up the reality stack.