Software vs Hardware

One of the things I am noticing is the trend to try to solve problems with software instead of dedicated hardware. That makes sense for a whole bunch of reasons, but the biggest ones are that the marginal cost of additional user is almost zero with software and that you can iterate your product much more quickly with software.

Smartphones make this trend possible because we have a hardware devices on us most of the time. A good example of this is the Moves app. Instead of using a device like FitBit, Up, or FuelBad, the Moves app turns your phone into an activity tracker. Right now, Moves eats your battery too much which is why it is not more popular. But fixing that is only a matter of time. Of course, iOS and Android could also make activity tracking part of their operating systems, and arguably should do that, but that's another story.

Chromecast is another good example. Why buy an AppleTV, a Roku, or some other hardware device to bring internet TV to your family room when you can buy a $35 dongle, connect it to your TV set, and your smartphone can now control your TV? It would seem that all TVs will eventually come with this feature that allows your phone to take over the screen and play whatever is on the phone. Then all internet connected TV innovation can happen in software instead of hardware.

This begs the question for me how the "Internet of things" will play out. What are the "things" that the Internet will connect to. Will they be smart cameras, thermostats, and doorbells or will all of those things run on our phones in time? And how will that be made possible?

This also makes me wonder about the health care diagnostic sector. Will I be able to take my blood pressure, blood chemistry, xray, cat scan, MRI, on my phone? Those last ones are kind of crazy, I know, but I am just aksing the question to make a point. Will healthcare diagnostics go the way of the compass, the flashlight, and the game console?

I don't know the answers to these questions I am asking. But it sure does seem that entrepreneurs are finding ways to do things with software and a smartphone that used to require dedicated hardware at a rapid pace these days. I think this is a trend to pay attention to. And it may, over time, make investing in hardare based business less necessary. Which would be a good thing from my perspective.

#mobile#VC & Technology

Comments (Archived):

  1. Sebastian Wain

    Just having the chance for an automatic basic diagnostic with basic sensors will prevent a lot of health issues without the need of a mobile MRI.BTW can you elaborate on the USV thesis for Quantified Self and the Internet of Things. Is this mainly oriented to the health sector or any other related one?

  2. arustgi

    Lets see if we can break this down a bit..A smartphone is composed of the following capabilities / components that have enabled an ecosystem of applications.1. A computing component2. A storage component3. A networking component (including all forms of communication capabilities such as Internet, Cellular, Bluetooth.)4. A series of sensors that are either built in (GPS, gyroscope, etc.) or external (like the square dongle.)Any application, current or future, can then be broken down into claims into the various components. If the crux of the value of the application is external to the phone (as we think of it.), then that application can be better done in a special purpose dedicated device. Otherwise, I would think that the an external sensor combined with a software application on the phone would be the most performant (both from a delivery and consumption perspective.)

    1. sachmo

      There are a ton of things that have specialized hardware built that do not fall into those 4 catagories. A car tire for example engineered from a new material that is impact resistant. Tell a smart phone to replace that!

  3. Yalim K. Gerger

    This post reminded me of the quote “Everything that can be digital will be.” With so many new stuff right on the horizon ( 3D printing, A.I. etc…) I think this trend will only accelerate.

  4. Avi Deitcher

    I wonder about this. It has been years since I was a deep-down hardware guy, but I have seen trends like these back and forth between software on general purpose computers (GPC) and specialized hardware for the last 20+ years.My favourite examples are networking. Originally routers were GPC with multiple drops, until specialized routers came along; same thing for firewalls before CheckPoint (we built some of the first ones on Wall Street); same thing for load balancers pre-F5; etc.But there are other areas as well. I think that things that are good enough in software, will happen in software, but over time, those that work better in specialized hardware will continue to do so.

    1. awaldstein

      The marketing part of this is–obviously–what fascinates me a lot.Things exists within common culture that need to be bridged in how the market adopts them to work. Just because they are there doesn’t mean they will get used.No one understands that more than of course Apple. Fred is right, no one should buy Apple TV cause we have Chromecast. Of course, they don’t and even I don’t.How you sell it is what people buy.We are working and living at the most perfect time imaginable.

      1. Avi Deitcher

        “obviously” – I am not surprised, but I tend to agree: the human nature part is far more fascinating than the technology part (as the engineer in me becomes more subsumed to the businessperson).Funny you bring up Apple TV. I have a v2, have been looking at Roku and Chromecast and trying to decide whether or not to buy one. I am not, and cannot figure out why I am doing what I am doing. Self-analysis is extremely difficult.”most perfect time imaginable” – how so?

        1. awaldstein

          I honestly feel like we are in a renaissance of sorts.In access to information and products. In the reality of a globally accessible local in everything from wine to art to food. In understanding nutrition, health and exercise. In a growing awareness, even at a mass market level of the importance of doing the right thing for the planet and people.And most important for myself, to continually find new applications and new ways to contribute my own expertise to a constantly changing marketplace of ideas and platforms.I’m an optimist, and feeling lucky and daily inspired.

          1. Avi Deitcher

            Serious optimist!

          2. awaldstein

            Optimism and belief drives change. Pessimism and kvetching and pointing fingers is just a bore and doesn’t get anything done.We all have bad days, shit that just doesn’t work, investments that go bad, projects that simply don’t work. Believe me I do.Why share that?

          3. Avi Deitcher

            I like it, with one (realist’s) modification: optimism and belief are required to drive change… but they are not sufficient. Some clear-headed realism about what change can be handled and their impacts, and what it takes to get there, make execution possible.

          4. awaldstein

            I’m there. I’m as practical a person i know and as focused on smart execution as a mantra as they come.But–in most everything we do, there’s this massive leap of faith, cause honestly, most everything is really hard. And no matter how many times we do it, still hard. An optimistic poise with a pragmatic approach is my answer.

          5. bernardlunn

            Somebody told me to listen to naysayers as they are the bridge from conception to execution. They maybe a drag, don’t let them stop you, but listen to them, gives you the execution to do list.

          6. Avi Deitcher

            Like. Need to find a way to incorporate into my business life.

          7. Vasudev Ram

            Read this a while ago:”An optimist is a person with no experience.A pessimist is an optimist with experience.”- Anon.Sorry, couldn’t resist … not meant at anyone in particular.

          8. LE

            In understanding nutrition, health and exerciseYou understand it but one of the problems with marketers and food is how it has become entertainment and enjoyment and not nutrition and what the original purpose of food was.Reason for obsesity and health problems.People have food addictions because it tastes so damn good and it’s all about the experience.For every Joanne Wilson that celebrates food and dining and has control (or you or me) there are more people who just want to use the food to medicate themselves. No different than alcohol abuse but way more widespread.Everything is about taste and enjoyment and experience from a food marketing point of view. (The “healthy” market is actually quite small in comparision, agree? Just walk around any supermarket. Or see how many restaurants are able to cater to healthy vs. enjoyment and entertainment.)

          9. awaldstein

            Of course you are right.But don’t start from how small the market is, start from how big it’s become. Look at the economic performance of Whole Foods, the massive explosion of green markets, the ecosystem of farm to table. At the hundreds of millions of dollars being poured into health and raw foods in the Tri-State alone.This is no blip my friend.

          10. LE

            Look I’m a big fan of whole foods. I almost located my office across from the whole foods just so I could get a buzz walking through there in the morning or picking up lunch.And while whole foods is definitely healthier they practice the same type of food marketing that I am talking about. Selling things that cause people to eat more and fixate more on food than they should.Of course with whole foods you don’t see the same “walmart” types because the up scale people either have less stress, more self control or more access to other things that they can satisfy their addictions with.By the way in my local shoprite I’ve notice the shelf expand with more offerings of the type of thing that you are selling I think with (note how disqus inserted the logo pretty cool – I didn’t disqus did.).Anyway each is heavy on marketing to get you to buy.I’ve attached a photo of one that I bought I don’t even know the cost. I was focused on the label, amount of sugar and all that. I bought two bottles. The other products had to much sugar.Marketing wise, what caught my attention was on label, the words “brazillian superfruits ; antioxidants & omegas, usda organic, original” not to mention the colors and graphics were excellent. (While uploading the pix I’m now noticing “gently flash pasteurized”. Whatever that is. Sounds good though. I love those embellishments. They go along way towards shoveling this stuff (and I say that with admiration).Funny story about zola. I was helping someone buy that domain. And I walk into the supermarket a few months ago and see the brand on the shelf. And all the sudden getting the domain took on a whole new urgency.

          11. awaldstein

            No one has been able to clearly draw the line between information and marketing.When it works for the market and the merchant, there is no reason to.You can blame your merchant for carrying improperly labeled goods, for not forcing their vendors to mark things that are important to your clientele, like non-GMO, like showing things that are raw when they are pressure pasteurized, but not that they have what you want to buy.For that, you thank them.

          12. LE

            “but not that they have what you want to buy”You want to buy it because they have created the demand for it partly by good marketing. Actually definitely by good marketing. Which is my point. That zola bottle and wording, don’t you think that helps them sell more product?And I don’t blame them and if I earned my living that way I would do the same thing. I like to make money and make no bones about it.You may be interested in this. I only skimmed it but no question in my mind “tell a better story” is correct. That’s really marketing, right?…”For that, you thank them.”Good marketing and good product or service are not the same although they can be.They guy who shows up to my house to do a repair and has a old truck and very little social or marketing skills and is a great mechanic is the one that I tend to place a higher value on and get a better deal from. That’s the guy who is not a good enough businessman to even return my call because he is so busy because he is so cheap. [1] They guy in the new truck who has business skills – you tend to pay more. They know how to exploit opportunities. And I have to tell you that this is after dealing with this type of things for a long long time with my own money.[1] I’m chasing one of those guys now a crack mechanic, $35 per hour a total pain to deal with.Very unscientifically of course. Phil Sugar your thoughts?

          13. Richard

            Gently pasteurized is like somewhat pregnant.

        2. JamesHRH

          Established brands at a huge advantage right now. I am a tech oriented normal & i have never heard of Chromecast or Roku.Apple running downhill – the ‘i’ products all work & everything else Apple brings out has that perception preloaded (appleTV will work & it does).Why exactly would I buy some new product from someone I have never heard of, when I can purchase the perceived to be risk free Apple version?

          1. Avi Deitcher

            Established brands always have a big advantage, but I don’t know if it is any bigger. In cloud services, yes, because you worry about the security and longevity of your data. But for a device? Not too big a risk if it is self-contained.

          2. JamesHRH

            Avi – Apple’s unique situation, in marketing terms, is that their brand encompasses all new digital innovation: music, then smart phone, the web services.Its a dominant spot, weakened only by their shoddy web services track record (its not in their DNA to really get non-HW offerings).Dropbox is, by far, Apple’s greatest strategic threat.

          3. Avi Deitcher

            As a consumer, I wish Apple were more of the platform and less about the services around it, so I could choose Dropbox for file sync, Google for mail/calendar/collab, etc. I still can, but Apple is always pushing in the other direction.Would be nice if companies really could just focus on their core offering and do it extremely well….

          4. JamesHRH

            Avi- that is bang on. It fits the Apple culture to outsource non-core things. Peter Drucker would agree as well.Well played, as someone we know would say.

          5. Avi Deitcher

            I’ve been saying it about wireless carriers for years. All of their talk about content, or customer experience, or relationship… everything about, “we’re more than just fat dumb pipes.”Personally, I think there is a great business in being really reliable and fast fat dumb pipes!

          6. fredwilson

            that’s what folks who bought a blackberry in 2011 were thinking

          7. JamesHRH

            I am going to argue that point.BB & iPhone are not the same category.Had BB jumped on touch screens / iPhone as a threat & come out with a Samsung level equivalent (quickly), then I would agree.But, they were to insulated / arrogant to do MI. They never saw a scenario where CIO’s would allow consumer devices into their environments & ignored the potential competition.Apple’s unique situation, in marketing terms, is that their brand encompasses all new digital innovation: music, then smart phone, the web services.Its a dominant spot, weakened only by their shoddy web services track record (its not in their DNA to really get non-HW offerings).Dropbox is, by far, Apple’s greatest strategic threat.

      2. Anne Libby

        And privacy/security concerns we need to address, too.And I was going to link to the wikipedia entry for Gattaca, but it pulled up the DVD cover…which I thought better not to do when replying to @awaldstein:disqus !Jim and co at Disqus, is there a way to delete that attachment when we don’t want it?

        1. JimHirshfield

          Add a media URL to a comment, media will be added. Edit and delete the URL, media will be removed.

          1. Anne Libby

            I’ll try it — this is the link to the Wikipedia entry for Gattaca, not to the photo.…(Edited to add: ah, it worked. I was concerned because the DVD cover appeared in comment field when I was writing the comment. Thanks, Jim.)

          2. awaldstein

            Been a long time since I watched this.Now on my list for the next sleepless night.

          3. falicon

            Such a great movie…I often think about and paraphrase the “I didn’t save anything to swim back” story…I live that every day and as much as possible.

    2. Timothy Meade

      Which is why I can’t understand the smart (software defined) networking push, I love the protocols and am happy to see new options, but I can’t see why anyone would want their PC servers as network SPOFs.

      1. Avi Deitcher

        It is possible that the flexibility outweighs the performance. I see the value in virtualized environments, as well as if the routers/switches/firewalls are actually hardware, but software controlled.

    3. fredwilson

      yeah, but if you have to have something with you, are you going to carry two things or one?

      1. Avi Deitcher

        Depends on context. When I play hockey, or go running or swimming, I do *not* want my smartphone on me. There are plenty of times I prefer not, but still want some functionality.I think there are other uses that are too expensive to end up embedded in phones. EKG? EEG?

  5. cdelrosso

    Excellent post. The best approach is to use the smartphone as platform and enhance it when needed with external hardware accessories. I have collected several examples where applications plus smartphones and hardware accessories solve concrete problems that were traditionally approached with custom hardware solutions. Smartphones are disrupting several industries in this way, and it all makes sense. A link to the examples: http://www.christiandelross

    1. Avi Deitcher

      Cool list.

    2. awaldstein

      Just shared this list. Nice!

      1. Avi Deitcher

        Good idea. I will share it.

    3. Dan T

      great list. i love the idea of “simpler” hardware that is critical to the form factor required – chromecast falls into this category – really simple piece of hardware with clever hardware. Galaxy smart watch is better example – it may be form over substance, but it is a great example that will inspire more ideas . . similar to how google glass will / has inspired new ideas.

    4. William Mougayar

      Great list. Can you add ? It’s a wearable/waterproof heart rate monitoring for swimmers.

    5. JLM

      .Fabulous list and one that highlights a key theory of mine — there is a lot of great stuff out there but the compilation and communication of it all is still incredibly inefficient.Health care — not freakin’ health insurance — is a huge opportunity for disruption and smart application of data accumulation.Medical data accumulation is NOT medicine. Medicine starts with the interpretation of the data.If one wants to dramatically reduce the cost of the delivery of basic medicine, then digital everything with remote usage of the data is one of the big pieces of low hanging fruit.It will take a splash of tort reform and the elimination of the defensive medicine umbrella to make it work but the applications of just remote digital imaging are already here now.”If you like your remote digital medical device, you can keep it. PERIOD.” ???JLM.

      1. Timothy Meade

        Seems to be why so many of these devices are off-label “fitness” devices. Disruption will simply happen before the FDA, etc. realize what’s really happening.

        1. JLM

          .Interesting and insightful observation.Disruption is what happens when we are all sleeping.Including the regulators.JLM

          1. JLM

            .Great article.I am absolutely convinced that the war on cancer is much further along than we think if we could only get a 1 + 1 = 5 exchange going.This article nudges the discussion in that direction.YOU are always part of the conversation. Thanks.JLM.

          2. Dave W Baldwin

            Yes, we are further along due to collaboration taking place among our higher institutions. We just have to bump up ML with AI to go with the Bio.

      2. fredwilson

        check out our most recent investment Jeff…

        1. JLM

          .Looks very interesting.Good luck with this one. But then, you do not seem to have any shortage of luck, no?JLM.

    6. fredwilson

      excellent list

    7. awaldstein

      Just saw this–more on devices and software…

  6. Jeffrey Hartmann

    I really think this is an interesting trend to watch. With more capable hardware becoming available and ubiquitous we are definitely seeing software ‘eating’ hardware. Take graphics for example, you used to have to build dedicated hardware to perform some neat algorithm over your framebuffer. Now we have pixel and texture shaders that can do some very neat things that just were not possible before. This programmability leads to other uses as well, now people are using the parallel hardware in graphics cards to solve physics, engineering, financial, and machine learning problems using hardware whose original purpose was just to display CAD drawings or video games. These uses are really just an example of what is possible with graphics hardware, and it is all because we pushed to make the hardware more general, faster, and more programmable.This trend obviously is happening everywhere, and there is also a very neat trend on the hardware side that is making things very interesting. The FPGA’s that are on the market now have a large number of very capable DSP slices included, so doing complex calculations in parallel is becoming more of a software problem then a hardware one. For custom applications, like the ones I’m interesting in like being able to do the calculations for a convolutional neural networks in hardware you get things going without ever really to make a set of masks and sending your design to a foundry. There are also some neat software tools that are starting to be created that makes designing ASIC’s and FPGA much easier. Verilog, VHDL, SystemC have ways of coding that can’t be realized in hardware, but there are some people working on tools based on Scala that have the opportunity to transform hardware design from a dark black art to just another programming language you have to learn.All in all, it is a very exciting time for the software engineer. Both general hardware designs, highly available capable hardware, and evolution of hardware design tools and FPGA’s are really making us into people with real super powers. We were already immensely powerful and could leverage a lot, but software is eating the world and this includes hardware. As we absorb more and more functions, both in our industry and outside of it, the momentum of it will only make things more and more interesting.

    1. sachmo

      It’s funny you mention FPGAs… They are often used in embedded systems for *very* particular / dedicated applications. Though of course, they are essentially ‘programmable’ hardware.But for example, say I need a real time video capture system that can display an image after some type of analysis has been performed.The system w/ the best performance specs may well be an FPGA. But once you release the board, due to highly specialized pin-outs, your system is way less configurable than say a computer with data acquisition card and GPU that could be built at similar price point.Why? B/c a full fledged computer will have expansions slots a la PCIe.So If I’d like to do a field upgrade later, the FPGA doesn’t buy me much, whereas the computer does. Now if all I’m doing is changing my image processing algorithm, maybe you don’t need to add in new hardware.But If I’m adding in motion control capabilities to my system, that same FPGA may require a board redesign, whereas a computer would not.So at surface level the FPGA seems like programmable hardware, but in reality often times modular and more traditional hardware components are more flexible / supportable / configurable after product release.

      1. Timothy Meade

        Of course, your smartphone could let you run DSP modules on it’s own DSP. In the Qualcomm devices the modem DSP (mDSP) and the application DSP (aDSP) are separate, so there’s no real risk from letting you run your own modules.Why would adding motion control require a hardware change, doesn’t your DSP/FPGA allow communicating over an internal bus with some kind of framing?

        1. sachmo

          If I’m doing real time image analysis, I’m not using a smartphone, I’m using a camera that’s capturing images at 30 to 60 frames per second let’s say, performing some type of operation, and then outputting within say 16 ms (so I have a refresh rate of 60 Hz, and barely perceptible latency).So I’m envisioning a couple RT cameras communicating with an embedded board with an FPGA. Of course this is an imaginary application, but one that I think is realistic.In a later release of my product, I want to mount this inspection system on a robot and reduce need to manually correct any defects noticed by my system.If my FPGA is on an embedded board – which most are – I’d have to anticipate what type of hardware I’d be attaching to it in Release 2 at the time of Release 1. This is so that I can design this type of connection to the pin-outs of the FPGA. These two releases may be a couple years apart.With a modular computer system, I don’t know and don’t care. The technology may change, but there’ll likely be something compatible with my empty PCIe slot.Now there are cases, where I’d rather have the FPGA anyway. But if I want to be flexible, I’d probably stick with a mini computer system.

          1. Timothy Meade

            Right I can see that for industrial applications you would want hardware, though I can see the value in modules that can be added to a machine and give it machine vision capabilities without each manufacture having to design and build an addon board for their own machine. These could be further customized from a software/firmware perspective for more specific applications.I was looking at your comments more in line with the focus of this article, personal technology, and commenting that the limitations on real-time processing in mobile applications are somewhat artificial, as the hardware can easily support these things.

  7. Tom Labus

    Those fitness devices tied to heath reading seems an interesting route.

  8. jason wright

    i read this and i feel i was born too soon. so much is to come, but how much of it will i get to see?

      1. Avi Deitcher

        Cool, but disturbing. I don’t want to grow old, sick or die, but the philosopher in me is disturbed.Ever read the Silmarillion by JRR Tolkien?

        1. Bala

          Nope, have not read that one… adding to the To Read list. Well, I did not say I approve of the finding but for those who “don’t want to grow old, sick or die” there is possibly a solution. I personally would like to be worn and torn and totally spent when my time comes, what a glorious way to go

          1. Avi Deitcher

            It is Tolkien’s legends, the First and Second Age prior to the Hobbit and Lord of the Rings. There is a whole tension in there between the firstborn (Elves) who do not die, and can go to the undying lands of the West, and Men, who grow old and die, and are barred from going to the West.The “blessing” of death is one of his themes. Makes you think…

          2. Bala

            “The fear of death follows from the fear of life. A man who lives fully is prepared to die at any time.”Mark Twain

          3. Avi Deitcher

            Dunno. I feel I have so much still to accomplish, that I fear the loss that comes with death. Ask me again in 40+ years…

          4. Bala

            you know where to find Lin28a 🙂

    1. Kasi Viswanathan Agilandam

      There is nothing better than what you have right now ….’now’.

  9. ShanaC

    Smartphones are both over and under poweredIt can’t lap count for me (water would break my phone)…Is better.Same with why a lot of peoplelike fitbit (and I wish fitbit had something for swimmers)

    1. JimHirshfield

      Cool device. Do you have one?

    2. Anne Libby

      There are waterproof phone cases, though. Downvoting the idea of swimming with a phone…

  10. obarthelemy

    Hint: look at your examples:- the first doesn’t work because it kills the battery and prevents my phone from doing it’s main jobs (actually, not phone, pocket computer)- the second is actually a hardware dongle, not softwareSo.. .maybe hardware is not that passé ?

    1. fredwilson

      well they will fix the first issue. apparently they have on iOS already.and the second issue will be a non issue when chromecast is built into TVs

      1. obarthelemy

        1- The Moto Droid X has some equivalent setup too; smartphone battery life is still too short, on both iP5S and Droid X.2- so hardware stops to be hardware when it’s integrated ? A SoC inside a TV isn’t hardware, but a SoC hanging off a TV’s USB is ?You’re not making any sense.

        1. fredwilson

          my point is more innovation is possible in software when the hardware disappears into devices like cars, phones, and TVs. i am not saying cars, phones, and TVs will go away. i am just saying i don’t have to invest in them.

      2. ci5er

        Mr. Wilson,I find your iOS example fascinating, but more directly I am finding that the “Internet of Things” that ITTT and RaspberryPi enable to be also relevant.One of my projects has me using a SDR (Software Defined Radio) stack to embed a cell-phone base-station to do environmental traffic scanning. With an incremental hardware cost increase of maybe US$35 on a US$10,000 digital signage unit.If you had told me 10 (or even 5!) years ago that I would be able to embed a cell-phone base-station to do foot-traffic analysis for US$35 … I would have called you crazy.As Yakov Smirnoff didn’t say: “In the future, sign looks at you!”

  11. Carl Rahn Griffith

    Software is the new soldering-iron.

    1. JimHirshfield

      Minecraft is the new Legos.

      1. Timothy Meade

        Ahhhh… my one hope for the future is kids playing Minecraft.

        1. JimHirshfield

          It’s a marvel to see my 8 year old at it.

        2. bernardlunn

          minecraft is amazing, seeing 12 year old kids programming without realizing they are programming ie it is just fun

  12. Jan Schultink

    I think people who design medical diagnostics equipment do not yet think of what would is possible if they can sell millions and millions of sensors rather than a few thousand to hospitals.Not all medical scans are needed real time (MRI for example). What could change though is if very basic vital signs monitoring was widely available on smartphones, researchers might put in more effort to find simple tell tale signs of upcoming doom. Maybe a heart attack starts with a certain beat distortion that can easily be picked up and trigger a warning signal. Currently, nobody can check this, but with a large of pool data you can.

    1. awaldstein

      The big leap for medical/healthcare has been getting to digital. It’s there now the rest will follow.It’s still a damn revelation that doctors offices have Digital Xray machines that are useable unscheduled, and provide the results digitally to the doctor in real time and shareable back to the patient.

  13. Matt A. Myers

    “Will I be able to take my blood pressure, blood chemistry, xray, cat scan, MRI, on my phone?”Star Trek FTW!

    1. JimHirshfield

      “Damnit Jim, I’m a blogger, not a doctor”

    2. Timothy Meade

      It’s just waiting for somebody like Elon to figure out how to build a coil-based MRI that operates on household current.

      1. sigmaalgebra

        Ah, just need to miniaturize the liquidHelium supply for the superconductingcoils and keep any iron or steel kitchentools well away!”Dear! Be careful with that cast ironfrying pan!” CLANG.

  14. panterosa,

    I agree with the phone as a can do many things given it’s current abilities and current and future software upgrades.I’m not sure I agree with the part about the TV connections. A lot of people are moving from TV to tablet for watching, or in our case, we’ll get an LED mini projector for movies for audience of 1+.

    1. awaldstein

      Yes, of course that is a trend but for me I want screens bigger, deeper in texture, cheaper, stupider and open to any input as well.

  15. Shaun Dakin

    The I had a nest and almost got divorced. Here is the thing. When Facebook doesn’t work it is nothing. When your. Ac doesn’t work and it is 100 degrees in July in dc you are screwed. I’ll try the Internet of things as long as the thing is not critical to quality of life. The Nest is long gone and my wife is happy.

    1. Dan T

      i’m counting on pioneers like you to forge the trail on these devices so they can become reliable and I can use them when all the kinks are ironed out. thanks for giving it a shot. . . .one more reason to wait for me.

    2. LE

      Exactly the point I was making in reply above.That said if the nest doesn’t work you can always yank it off the wall and simply cross the wires to get some cold air. All a thermostat does is either run or not run the system. When the system achieves a certain temperature it stops running. When the temperature rises it runs again. And so on. [1][1] Source: Observation of mercury thermostats with those springy things and how the circuit in one is closed. I’ve attached a picture. The mercury is where the arrow is. It simply closes a circuit.

    3. fredwilson

      your wife sounds like my wife. we don’t use new fangled stuff when it comes to heating and cooling either.

  16. Richard

    Technology Forecasting can be grouped into four buckets:Judgement or Opinion (the genius forecast)Extrapolation or trend analysis (Moores Law)Modeling Behavior (network effects)Scenarios (global warming happens)Each of these buckets has useful tools for forecasting technology. Forecasting the commercial success of a technology adds another layer of complexity.

  17. pointsnfigures

    On the medtech side, I have already seen a phone that can do an EKG. It sends the data to a cloudbased data base and the docs can analyze it. I know a paper algo that is a breath test to predict a heart attack. It’s waiting for someone to turn it into a 24/7 digital scraping EMR data constantly to be used to predict heart attack before it happens and alert medical professionals.We aren’t that far away.Eventually, there will be thousands of B2C apps for medicine to allow individuals to take more control over their care. Eventually, there will be machines that turn commoditized medical procedures, like reading x-rays and anesthesia into robot/machine functions that turn data around faster and cut costs.

    1. William Mougayar

      Here’s another one tackling ear infections in kids. It’s a very common thing, and you typically rush to the doctor…millions of visits. CellScope has a smartphone-enabled otoscope for remote diagnosis of pediatric ear infections.

      1. pointsnfigures

        The interesting thing to me is that instead of opening up and embracing it, the legacy hospital software systems are fighting it. They are extremely hard to crack vertical silos with some tough barriers to entry.

        1. LE

          legacy hospital software systems are fighting itLegacy people fight things for many reasons. One is they run systems that need to operate 24/7 that need to always work. They can’t introduce things into that system (no matter how trivial) that might jeopardize the functioning of those systems.It’s like my dad comparing when his macintosh doesn’t work and he has to call apple to me who has to have systems operate 24 x 365 with no excuses. Or in one of my past businesses where the machinery had to operate or you didn’t get the job done and you lost the big customer that took you a year to land. So if someone comes in and tries to sell you on something you are focusing on the downside of the decision not the upside. These institutions are slow moving because they need to be slow moving.The culture is much different on the web. Many people giving things away for free or charging very little and people have come to expect that things don’t work all the time or aren’t perfect.

          1. pointsnfigures

            they also have big fans in the finance dept-but not such big fans in the medical practice dept. they intentionally make it tough to innovate within the silo-like Microsoft.

      2. LE

        For home usage this will totally depend on the price point. Personally I don’t think the amount of times kids have some ear issue will drive a large adoption by parents. Not to mention reimbursements if the pediatrician can’t get reimbursed for remote diagnoses or viewing an image it’s not going to go over big.

        1. William Mougayar

          right. the reimbursement rules will need to change for all this remote thing to work. that’s why doctors resist email communications with patients, because they don’t get paid for email advice.

          1. Mordy Kaplinsky

            I read an interview with a doctor out of Scripps (forgot his name but he’s a leading pioneer in E/M Health) about how doctors will be prescribing apps and peripherals as the future of medicine, and this dovetails with the point you’re making.

          2. William Mougayar

            Good point. So, just like they prescribe drugs that they are familiar with or that are approved, they’ll recommend Apps and devices that are “sanctioned”.Maybe there’s a business to be made for approving/certifying these new toys.

          3. Mordy Kaplinsky

            that’s primarily the FDA/FCC/UL’s job but a specialist house that adds on certification services would be interesting.There are many other opportunities here, including helping make these things, productions, combined platform for data as a single one is useless, conferences, combined modular devices (Phonebloks) for the likes of aircraft, buildings, stadiums, first responders and ambulances, HIPAA (international versions) compliance. lobbying, etc.These just name a few opportunities, so lemmeknow if you wanna get into this 🙂

          4. LaMarEstaba

            Half right. http://www.healthleadersmed… | Yeah, patients don’t want to pay doctors for email advice, but healthcare orgs have been able to bundle it with the existing insurance premium and other things. Whether or not doctors can get paid for it is actually up in the air right now – in the US, at least, that varies by state. The CMS, which runs Medicare and Medicaid, is thinking about authorizing reimbursement for telemedicine for chronic conditions, if the patient agrees to stay with one doctor for a year. There are a handful of other conditions, but it’s something that’s getting a lot of interest right now.

          5. LE

            Also another reason is that email communications is not finite. If someone is in your office you can control the interaction and end the visit. With email when you write back there is no “end” you would get follow up questions possibly and also it takes much more time to thoroughly answer a question by email (in many cases) than by in person interaction. And assumes physician can type at a good clip. Also creates a paper trail which would work against the physician almost certainly in a lawsuit.

  18. Robert Heiblim

    Thanks Fred, an excellent post and it does seem quite true that software is overtaking hardware. While some of the effects are great, the one that is taking profits quickly out of hardware (and why you probably don’t prefer that) is also concerning. Software needs a platform, but if there is no profit in hardware what will happen to innovation and progress there? One can expect innovation at some levels such as sensors to capture the data you note on health and fitness, but what company can afford to keep upping the hardware content when margins are so depressed. As another post notes, perhaps this is cyclical in nature, but one of the issues for software then is fragmentation, uneven deployment and often uneven experience. It is certainly an interesting time.

    1. SubstrateUndertow

      That evoluting relationship between hardware/software seems very analogous to the evolutionary interplay between our hardwired biological-wetware and our programable cultural-software.At some point human biological-wetware evolution reached a saturation tipping point where generic abstract processing power, plus input sensing, plus output actuators were more than adequate to meet all our low level survival functions. This evolutionary stall-point become the new base-camp platform beachhead on which to construct the next layer of the human evolutionary stack, namely, cultural-software extensions(ideas-attatudes-language-tools-techneques etc).Sure we will still build out some physical extension to our low level biological sensing and physical actuators but the bulk of the human evolutionary story form here on out will almost certainly be about reprogramming our cultural-software for orchestrating recombinant human/machine behavioural synchronicities.(short of strange-loop interference by genetic engineering)I do sense a disturbance in the force though ?So many cultural-software-extention Apps are pouring out of the gate and yet little to no collective formal conscious effort is being expended on the development of any cultural-OS-Kernel software.First things first !Emergent synchronicity is the pixie dust, the mojo, the engine that drive all evolutionary complexity trajectories. Without any credible cultural-OS-Kernel software, namely some new form of effective digital-social-contract, around which social-software synchronicity can coalesce, our new cultural-software based evolutionary platform simply flies in the face of historical evolutionary reality.Sure atoms and cells can coalesce the necessary, evolutionary synchronous trajectory by simply riding in on synchronicity’s self-selecting and self-reenforcing, statistically-emergent, strange-attractor, self-organizing-dynamic, all driven by the statistically consistent causal necessities exhibited by their underlying substrate components.The underlying substrate components, the pennies in the currency of our new cultual-software based social-evolution, namely humans, exhibit no such consistent underlying causally necessitated behaviours.Coalescing that necessary evolutionary synchronization-trajectory by simply riding in on synchronicity’s self-selecting and self-reenforcing, statistically emergent, self-organizing-dynamic atop the statistically consistent causal necessities exhibited by the underlying substrate components is simply not on for culutual-software evolution given the immense volitionally self-serving behavioural volatilities inherent in humans as substrate.This new evolutionary breakpoint existentially mandates some sort of collective, conscious, cooperative, strange-loop-driven Social-Operating-System design effort on our part.(network-effect synchronizing political-framework)Final evolutionary stall point or just a really hard nut to crack ?<hr/>Coalescing synchronicity is theSTRANGE-ATTACTOR-MOJOdriving all evolutionary complexity.

  19. John Frankel

    Software is eating the world, but it has to run on hardware. The Phone is getting more capable, and more multifunctional all the time. The world, however, is physical and thus there will be hardware components in the physical world that can interact with smart phones and other devices.One example of something that will not be replaced by a smart phone but augmented by one is you home deadbolt. You still need a lock; but it can be smart and UniKey’s Kevo lock provides security intelligence that is a combination of hardware and software. See… for details.The same goes for InteraXon’s Muse. It is a smart phone and tablet accessory that reads your brainwaves. for details. Note that it replaces large dedicated EEG equipment which something that is a portable as Google Glass.[Note that both of these examples are’s portfolio companies.]These examples show that either we are going to embed chips in places we have not before (e.g. locks) or we are going to slim down the devices that have had chips before (e.g. EEG measurement devices) to surface information that did not exist before or add functionally that we will not be able to live without.

    1. Dan T

      home automation – like the locks that you mentioned and the thermostat (nest) that was also mentioned is an interesting category. Alarm systems with newer technology have also been around for a long-time. Sonos – – now that is MY personal fav. Again, another example of clever software with simple hardware – leveraging the cell phone. The monitoring / security cameras that use simple camera systems with cellphones and recording in the cloud are cool too. . . this stuff is passe from a technical standpoint but at an infancy stage in adoption. the simplicty, reliability and marketing all needs to improve.

      1. John Frankel

        Totally agree that we are in the first innings. You nailed it by referencing simplicity, reliability and marketing as key. These are the key criteria that lie behind the iPhone and its dominance. Platforms will be built and consumers will get a layer of information they never had before, and pay for it.

        1. LE

          Part of the reason more products aren’t like the iphone is the people who develop those products aren’t able to think down to the level of normals. So what is simple to them is complex to regular folks. And a non starter.In order to be successful with something on the consumer level my feeling is you need a number of people on staff at higher levels that are actually clueless and that things don’t come easy to.You see this in everyday things and even here at AVC when using disqus..Guess what? It’s not obvious to everyone how to post a picture, flag a comment or collapse a thread. Guess what? Some people don’t even know what a thread is. I’m using small examples to illustrate my point. And of course beating on disqus like a parent who loves a child but wants them to try harder. (Note how I softened what I said by doing that.)Anyone who says “simple all you have to do is…” is missing that regular folks don’t have the attention span for anything that isn’t intuitive unless there is some big reward for figuring it out. (Like sex or a remote control for kids). They will just bail and move on to the next thing.I call this my “puny brain” theory.For years of course computer people have taken advantage of the fact that consumers don’t really know jack squat to sell them things that they don’t need.Apple on the other hand shows them a screen (on the new ipad mini) that they can clearly see is sharper and then make a buying decision. Auto manufacturers are pretty good at this as well. Same with restaurants. Sell the experience (and food taste) not the food quality.

          1. Mordy Kaplinsky

            This is a genius bit of product advice hidden in the comment.”In order to be successful with something on the consumer level my feeling is you need a number of people on staff at higher levels that are actually clueless and that things don’t come easy to.”

          2. Timothy Meade

            But doesn’t that work against you just as much? If the people can’t understand that something is inherently complex they can’t understand how important it is to simplify it, and why shipping the complex and unusable version won’t work.Having someone that understands all the pieces making the decision is what makes it not design by committee or feature list.An HP exec looks at the lineup and says we don’t have a product for X market, go make a new one. The mid-level HP person comes back with the U15X4M and the exec is happy. Then HP’s sales/marketing calls Best Buy who says they need it in three versions for their back to school push in different parts of the country.Apple looks at their line-up and says Free iPod touch with any computer for back-to-school.

          3. Mordy Kaplinsky

            Obviously that person wouldn’t be working in a tech company, so the point is that while one can “get it” on an intellectual level the implementation of an inherently complex problem may be beyond them.What this approach gives you is in effect a focus group / usability testing at the product design phase. The more diverse the group is and the earlier the questions arise, the better you can design the product to address the broader market otherwise known as “normal’s” needs.

          4. Timothy Meade

            By “higher levels” did you mean in charge or thinking at a higher level?My concern is with the former.

          5. Mordy Kaplinsky

            Its a quote I was highlighting from the post I responded to by @domainregistry:disqusI understood it to mean people involved at various levels but removed from the direct product. This can be anything from financing to strategy or marketing, etc.If I misunderstood @domainregistry:disqus will hopefully correct me.

          6. Timothy Meade

            Yeah, I mean I was wondering why you pulled the particular bit of advice out and labeled it genius.I’d be curious about @LE’s reasoning as well.

          7. LE

            It means both removed from the product generally but more importantly not to swift on the uptake in the particular area in question. Does not mean stupid. Could mean stupid. I used to show marketing pieces to a stupid person to see what they looked at and if they got the drift. It worked pretty good.For example if you are designing signage for an airport you want someone both not familiar with the airport but also slightly older and not able to see that well. Then you will be able to design signage that anyone can read.I regularly go in restaurants with dim lighting. And I have a really hard time reading the menus without my iphone “flashlight” and/or glasses.It’s quite obvious to me that the type is to small on the menu.But the person designing the menu a) knows the dishes very well (so they can auto complete) and b) is probably viewing it in normal lighting and c) might also be younger and have better eyes than I do and d) knows all the words, sauces etc. so they can bridge more gaps in knowledge e) is not around to view what is going on and f) the waiters and waitresses don’t feed back what they see. So nothing changes.

          8. Timothy Meade

            Understood, and I agree 100%.The last case is often the restaurant showing how sophisticated it is by using words the customer doesn’t understand, but that help sell the ambiance of the place. Even with bigger type the menus are unreadable (and certainly unscannable) to most normal people.

          9. Mordy Kaplinsky

            You highlight why I believe diversity is a plus which makes a company better and is not just something that one should do because its the right thing. Historically countries which were open to all for commerce did well while those that closed themselves based on religions or nationality suffered financially.

          10. LE

            “U15X4M”One of my beefs is the fact that people who are familiar with their product naming simply don’t understand that normals are confused by their product naming. And it’s so obvious that is the case. [1]Same as we know that a phone number, no matter how it’s formatted is a phone number.2129947880212.994.7880212-994-7880(212)-994-7880212 994 7880etc.But not with an overseas phone number, right? Your brain can’t immediately figure out the pattern. It has to pause and think. And a pause is a lost sale.[1] Funny I mentioned “sku” to my wife the other day and she doesn’t know what a “sku” was. So I told her. My guess is if I ask her next week she will have already forgotten as well.

          11. Timothy Meade

            Let me preface this by saying I typed random letters and numbers starting with U and containing an X.”U15X4M”Obviously, that means an Ultrabook (what’s that?) with a 15-inch diagonal screen and 4 Gigabytes of RAM.Best Buy guy: It’s like a 15-inch MacBook Pro that can run lot’s of apps at the same time.Customer: Oh, I want that. Show me that.I used to use this joke about Sony who’s naming convention is just as insane, but it applies here as well.

          12. LE

            “I used to use this joke about Sony”Right but this shows a bit about legacy thinking in terms of how products used to be sold.Used to be you walked into the store and knew you wanted a trinitron (you may be to young to even remember that). So you saw a bunch of TV’s and there was actually a salesman who had a clue that would answer questions. And you would see the product and make a decision.But with things selling online it’s much different, right?Apple of course being simple handles both issues.At least with their new inventory once you move off new and go to used or older it becomes confusing.

          13. Timothy Meade

            Yeah, Apple is getting a bit confused here. Moving the old models back was working for a while but probably hurting the sales of new models (or had the potential to).I think 5C is a bad decision, it’s hard for somebody to choose that over the 5S based on price alone, and they aren’t really going to grab the lower end which Android has essentially consumed. (Look at the prices on some of the Motorola devices, without a contract.)It also serves to fragment their push for fingerprint recognition.I would probably push the devices back for 6 months, then firesale and prepare for the next new version. (Really iPhone is upgraded every two years with an interim version with a minor feature but same case.)The iPad 2 could live a long time if Apple doesn’t outright kill it, it’s good enough and has a huge ecosystem of devices and cases, still more for it than iPad 3 from what I can see.

          14. Timothy Meade

            Had a TV and two monitors with a Trinitron tube, or at least a flat CRT if it wasn’t by Sony.I may be young but I’ve been exposed to a lot of hardware in my life, from old cabinet style TVs with sonic remotes, turntables, and my C64, as well as oscilloscopes, cassette players/recorders, VCRs, reel-to-reel tape. I may not remember the day when you walked into the neighborhood Color TV dealer personally, but I’ve had enough of those experiences related to me I feel a bit like I do.I appreciate the concept of a knowledgeable dealer, who’s also the repair person, parts supplier, etc. and would prefer that to somebody that intentionally mislead customers or attempted to cover up for their lack of product knowledge by recommending the overpriced or high-margin product that doesn’t solve the customer’s problem.

          15. LE

            One of the reasons that I think google could improve as well as Amazon (AWS). They think the idea is to simply hire (their idea of) the best, brightest and the most well educated people. But those people’s brains work differently than the people who are using the products and services that the company sells.An example of this is in a comment I made here today. When someone mentioned NEST I said:you can always yank it off the wall and simply cross the wires to get some cold air….knowing full well that most people probably can’t and aren’t going to do that and don’t even know what I am talking about. But the point is I know that people don’t know that. The best and the brightest by my theory think they are super smart but can’t put that intelligence into the right perspective with regards to end users. They are to far from normal.

          16. Mordy Kaplinsky

            I couldn’t agree more which is why I pulled the quote 🙂

      2. sachmo

        Sonos speakers are still custom HARDWARE though. They have supporting apps for sure, but what they’ve built is intelligent hardware.

    2. William Mougayar

      +1 for InteraXon, a Toronto company.

    3. Timothy Meade

      Right, we should find the sweet spot of devices that already feature chips (and 3.3v DC) for other purposes and augment them with radio chips and security.

    4. sachmo

      Couldn’t agree more, you’re going to need a lock of some kind. That’s not going away. DUMB hardware will be replaced by SMART hardware.

    5. Timothy Meade

      Interesting, hadn’t heard of Omek before, or that it had been acquired by Intel.I wonder if they are going to get serious about licensing chips for TVs and other devices again where this stuff could be really useful, if PrimeSense is off the market.

  20. John Wilbanks

    Health related response only: Blood pressure, heart rate, breath tests (asthma etc), and some “dry” diagnostics via accelerometer and altimeter are already there for mobile. And you can make the hardware of the phone incremental to add things like electrocardiogram etc – anything with galvanic skin response is on the way already. For things in your blood chemistry the movement is towards connected patches with microneedles that can do almost-noninvasive sampling and communicate back.The hard part of the move is that the “hardware” of health normalizes the data stream very effectively. We know where, when, and how the samples were collected and converted into the digital realm – whereas a patch on my back communicating to my phone has no idea if I’ve been feasting or fasting or running or anything else. The level of monitoring required to automatically create that kind of normalization makes existing network surveillance look skimpy, and the software required to do that normalization post hoc isn’t out there yet. The dirty data cleanup software and the data stratification tools that work on the web have error margins that aren’t acceptable in health care. And the sample sizes aren’t growing fast at the moment. It’s going to happen but it might not happen as fast in health as it seems it rationally should…

    1. Timothy Meade

      I’m confused, why would it lose samples? The existing protocols are broken, and don’t incorporate PKI at the source (figured that out watching the Frontline episode on SCADA a few years back), they also assume a connected orientation and seems to be derived from Bluetooth, which was derived from IrDA.I think raw ethernet would actually be superior, monitor the packets and reassemble into a stream.

      1. Timothy Meade

        Actually, similar to the M8 chip in the new iPhone, a hardware chip could gather those samples remotely when the AP is asleep, busy, or just doesn’t want to devote cycles to it.

  21. Doug Gibbs

    First, there will always be hardware. The software has to run on something, or it is just people sitting around thinking. That works, but is a difficult investment.Qualcomm has their tricorder X Prize. http://www.qualcommtricorde…The phone solves a major problem for sensors. You have a portable display to configure and monitor devices. No need to have separate display devices for every item in the house. For the price of a low power bluetooth module (about $15) any device has a smartphone connection.

  22. Matt Zagaja

    I was thinking of buying a scanner until I discovered I could use the TurboScan app on my iPhone ($2!) to make better quality scans with my phone camera than most scanners I’ve seen. The only reason I use my point and shoot camera is because it has a 12x optical zoom my iPhone does not have.

  23. Yukon Palmer

    Great post!The key issues with the IoT are:1. Is a hardware sensor required for the application? Some IoT applications will require a sensor. Examples include monitoring and remote control of home appliances, irrigation sprinklers, etc. Eventually, the OEMs will incorporate them into their products, but it come at a cost.2. Who pays for the IoT solution?In my above examples this would be an enterprise play. A majority of IoT deployments (at least initially) will be within the enterprise. Depending on the pricing model, it will likely require a sales cycle and thus a high CPUC.This article is timely for me because we have a “work around” in process to address these issues. We are using a freemium model to get our IoT application within the enterprise. Driving down CPUC is the primary driver for this.

  24. iggyfanlo

    Yeah… they will do many things on smartphones and may very well be the “remote” of life, but you will still need hardware devices at the edge

  25. William Mougayar

    I’ve had a long experience at the HP Medical Products Group selling diagnostic and monitoring equipment for 10 years. In 1987, we introduced one of the first solid state Holter monitoring device (previous ones were like a tape recorder), and the system was sold with a standard PC attached to the units. Both the PC software and the device’s firmware would update via software, and this was 1987. I was the specialist for that product and top sales guy worldwide, so I got really close to it.I’m surprised that we haven’t made big progress in this area of smart devices since then, but it seems that this is about to change.What we will have is not just an Internet of Things, but rather an Internet of Everything. The smartphone accessories market will explode, as you accessorize your smartphone with all kinds of sensors, electrodes, wired or wireless, etc.I’m also seeing the trend in decoupling and unbundling of the sensing vs. computing. But your smartphone is one device, so it may not be as good for the dedicated tasks, but better for the on-demand ones.The Edge of the Network is getting smarter. To a computer, your body is just another Thing. For healthcare, this will also change how we interact with our doctors.

    1. LE

      one of the first solid state Holter monitoring deviceA throwback to the days when you used to see “in living color” and “stereo” and “via satellite”. I’m sure many AVC readers weren’t around when most things used tubes.And as it happens I have a tube tester [1] that I used to use as a kid to test tubes for my dad. He’s been hocking me to put it on ebay and I’ve been putting if off.(I also have a 40’s or 50′ blaupunkt radio.)[1] So any musicians out there looking for a Hickok 533a please contact me.

      1. pointsnfigures

        We unbundle Facebook and Craigslist, why not unbundle everything else?

    2. Timothy Meade

      By the way, I had an idea to make IoT/IoE chips truely economies of sale and give-them-away cheap.Only a few companies in the world could pull this off though.

    3. fredwilson

      “What we will have is not just an Internet of Things, but rather an Internet of Everything”word

  26. William Mougayar

    And to prove this is a hot market, Apple just bought PrimeSense for $345M, gestural technology that was at the heart of Kinect.

    1. Timothy Meade

      So Apple just confirmed the Apple TV?Maybe PrimeSense should have gotten into the personal photography business, I hear that’s easily worth 4Bn USD.

  27. JLM

    .It is not the smartphone of today — which is already pretty damn impressive — it is what the smartphone is going to become as it converges with tablets and laptops.The tablet and the smartphone are already on a collision course. I routinely use my Nexus 7 to make Bluetooth enabled phone calls and to Skype and Hangout. It is almost my phone already.It is already better than just my smartphone and it fits in my back pocket. I just have to remember to take along a BT device.It is only going to “converge” more violently in the future.THAT smartphone is going to be THE smartphone on which all this stuff gets done.And, it is happening right now. Convergence.JLM.

  28. george

    There’s been a major shift in focus to software, that’s where all the creative and solution based ideas have accelerated to enhance our lives. I still give tremendous credit to Apple, they really set the blueprint for others – that initial vision of symbiotic software and hardware, created so many possibilities.

  29. george

    Why buy AppleTV or Roku? Familiarity, preference, quality, integration, etc…it’s not always about cost.

  30. Emil Sotirov

    The “dedicated” portion of the hardware may end up being what “peripherals” used to be in the PC world – input, output, specialized functions (printing, measuring, etc.).The question then would be – which form factor would become the dominant generic computing device. The phablet seems like a good candidate.

  31. James Ferguson @kWIQly

    It’s been happening since the…”The Jacquard loom was the first machine to use punched cards to control a sequence of operations. Although it did no computation based on them, it is considered an important step in the history of computing hardware.”

  32. Kasi Viswanathan Agilandam

    ” Will I be able to take my blood pressure, blood chemistry, xray, cat scan, MRI, on my phone?”You can get all those displayed on your smartphone…ALREADY …happening across the world … but can’t ‘take’ as in ‘detect’ ….each needs specialized source/detectors-sensors other than computer …computer comes in only as a data analysis and display unit….not everything…If i can attach a xxx-kg magnet to my smartphone and carry along…or minimize it to xxx-mg … then yes….i can take MRI scans on my smartphone….If I can attach a xxx-kg x-ray generator to my smartphone and carry….or minimize it to xxx-mg …then yes…i can take my ct-scan on my smartphone…There are some small piece of world left out there beyond puters :-).

  33. JamesHRH

    The natural limit on multipurpose usage of a single device is 7. Lots of consumer studies show people max out on remembering choices & that the max is ~7.So:- talk to someone- message someone- email some thing to someone- music player- video player- print ‘player- health monitorI think your TV will become one of many screens for your ‘players’, so that Apple will comeout with one w AirPlay integrated ( or they should, people would buy it ).I am already looking at printers that AirPrint to skip the middle man steps.

  34. jason wright

    if novelists and science fiction writers were software programmers…

  35. Dave Goldberg

    Chromecast seems like the ideal solution, and fits into the category of ‘dump pipe’ devices or ‘slim clients’. But, I think this idea has limits, for some particularly mundane reasons.I own an AppleTV, Chromecast and Roku3 device for the two TVs I have in my house. Given the price point, I bought the Chromecast without hesitation, and hooked it up in my bedroom. But – it turns out the Chromecast is a pain to use. Every time I want to use it, I have to get my phone or laptop and launch some app (whether it’s the browser or an app on my smartphone). This is not the best experience for bedroom viewing.My sense is, having a remote control is a better experience in this case. Yes, it can get lost etc… but, it’s single purpose, is fast, and there is no software based confusion about it at all.Compare this with my ideal in-car experience, and the one I’d like to see. Right now when I get in my car, my phone connects to the bluetooth automatically. This is super convenient. I don’t even need to take it out of my pocket. The ‘hardware’ controls are built into the car itself, which differs from the Chromecast experience, where there are no hardware controls at all. What I’d like to see in my car is the replacement of the proprietary Honda GPS system, with my Google Maps account – via bluetooth. This is the best example of thin client I can think of. The controls are built into the device (the car’s dashboard) and are all generic enough and readily accessible, but the software comes from my phone. It can work like magic, unlike the TV experience, which is a bit cludgy.People laugh at the idea of a screen on the refrigerator, but I think I would prefer having a dedicated kitchen screen vs. dragging a tablet into the kitchen only to get food and crap all over it. And I definitely WANT a screen in the kitchen and I want it to have all my recipes from Epicurious or whatever.So I think some nuance is required when thinking about what the best experiences in our connected world. I think application specific hardware controls are a plus in most cases, and this is where Chromecast fails.

    1. Timothy Meade

      Maybe, but the screen on the fridge idea assumes you want to stand there and look at the recipe while you’re cooking, maybe it should display only the ingredients while the iPad 2 in your stand displays the actual instructions. Maybe it should allow for gesture control with the FFC?It’s possible this would require too much buy in from clueless industries to be successful though.

  36. William Mougayar

    btw- even the chromecast has been softwarized. you can use your android as a chromecast with Cheapcast

    1. Elia Freedman

      Someday all the hardware will be superfluous.

      1. William Mougayar

        I’d like to see a Lego approach to smart hardware or accessories that you can plug and play or move around, for the house, your body, car, appliances, etc..

        1. Elia Freedman

          Or no “hardware” as we know it at all.

    2. fredwilson


  37. William Mougayar

    That Moves App is very interesting…I’m trying it now. So, these sensors are going inside.

    1. fredwilson

      i think they fixed the “battery issue” on iOS with some new features in iOS7. i don’t think they have fixed it on Android yet

      1. Paul Smalera

        I have been running Moves on iOS for many months now without battery complaints, FWIW. Even on iOS 6 it was fine, but then again I am an ABC’er (Always Be Charging) usually chained to a desk, and don’t expect very much from my battery at all.

  38. LaMarEstaba

    I agree that it makes more sense to focus on the software than the hardware. I have my Pebble and they keep making it better with what feels like constant software upgrades. They said that they didn’t want to keep upgrading the hardware – at least one of the founders is ex-Apple and understood how people felt when each new generation had such better hardware than the previous.I actually work in healthcare IT on a mobile app for hospital patients; we’re a relatively small team in one of the biggest healthcare IT companies. Withings and Fitbit already integrate with our patient portal so that you can send stuff like blood pressure data back to your provider and get it filed into your chart. We just opened our stuff up in an effort to increase interoperability and to play nicer with the other kids in the sandbox.I was particularly interested in Mark Cuban’s recent investment in Validic, since that’s essentially what part of our end game was when we opened up.There are a lot of people trying to disrupt healthcare IT right now, but it’s incredibly complex and highly regulated. Right now, we’re open to hospital patients 18 and up; the legislation surrounding adolescents’ medical information is a nightmare and varies from state to state.

    1. Timothy Meade

      Does sending data to their chart require patients to actually do anything? Does it work in a medical facility?

      1. LaMarEstaba

        Yes and yes. You have to connect your accounts to the patient portal – both sides want the health information to be safe and secure. Yes, it works in a medical facility, although normally Withings products are used for home monitoring. There are other interfaces if you do things in a hospital, but there’s no reason why a Withings scale for an admitted hospital patient wouldn’t work.

        1. Timothy Meade

          Right, but I mean does the patient have to send every single individual sample to the chart, or does opting into it for a period of time suffice.I’m interested in the real-time monitoring aspect of it, if it could be used in care facilities that won’t invest in fully networked monitoring, individual patients and their families could provide it.

          1. LaMarEstaba

            Opt in should suffice. There’s no need to send every single sample to the chart. We’ve had automated patient entered flowsheet data for a bit now; the use case that I’ve heard for that was glucometers.Real time monitoring is awesome; right now, it’s really for home monitoring for outpatient stuff. However, my team, which works on the inpatient side of patient-facing software, is probably going to integrate those devices into our side.There’s a lot of inspiring work being done with this.…Hospitals face reduction of reimbursement if their congestive heart failure patients are readmitted within 30 days, so it’s in their best interest to really closely monitor patients who are at high risk for readmission.

  39. Semil Shah

    A few notes. (1) The M7 processor inside iOS may be a threat to the FitBits of the world, we will have to wait and see. (2) Remaining on iOS, interoperability across iOS devices + iBeacon (or BTLE) will enable many of these things to happen automatically if integrated properly.

  40. daryn

    I feel like I’ve been seeing a lot of the opposite: people solving simple problems that can easily be handled with software and existing devices using some newfangled custom hardware solution. It very much seems like choosing the shiny new tool over what’s appropriate for the job.

  41. sigmaalgebra

    Gee, I missed out on ‘Fun Feature Friday: EmbeddableMedia In The Comments’ so here ’embed’ or attach apicture!Apparently so far no one’s mentioned ‘softwaredefined networking’ that has been keeping Cisco CxOsup at night?And what about ASICs and FPGAs — right, ASICabbreviates ‘application specific integrate circuit’which is mostly from some software in the designprocess and FPGA abbreviates ‘field programmablegate arrays’ where essentially just from softwarecan change the thing to look like, say, an Intelprocessor or a RISC (reduced instruction setcomputing) processor.I have a friend who keeps telling me that little IP(internet protocol) routers are really just a littlecomputer running a version of ’embedded’ Linux andsome software.Gee, it has long been the case that a printer hadone or more microprocessors and was awash insoftware. I still have a Diablo/Xerox daisy wheelprinter (great for addressing letters) that, as Irecall, has an early Intel 8080 chip in it and a lotof software.Heck, in a sense, Intel was started by the idea ofreplacing hardware with software — the deskcalculator Intel designed using, as I recall, anearly 4 bit processor and some software.And the early HP calculators that did so much forHP? Yup, a little hardware and a lot of software.That is, the thing could calculate all the morepopular mathematical functions, and that’d be a lotto do in hardware!So, make the peripherals and their hardwareinterfaces and software device drivers standard(some years Intel had a project for the softwarepart) so that all the hardware is COTS (common offthe shelf, and typically cheap) and the rest issoftware.Of course for the standard interfaces, etc. there isMicrosoft’s Plug ‘n Play — a nice step forward ifrecall the earlier hardware configuration issues –so use software both to identify the device andconfigure its interface.There is a rumor that actually IBM did well withRISC computing quite early on — used it, with somesoftware, as the core of an I/O ‘channel’ on their’mainframe’ computers, that is, a little RISCprocessor instead of a lot of hardware.Then there’s the old ‘wake on LAN’ (local areanetwork) where can have 1000 computers in abuilding, with all the computers powered off, theNIC (network interface card) for the LAN still ableto receive a command over the LAN to power up thecomputer and receive software from the LAN andexecute it and, thus, let a central systemsmanagement computer, overnight, install new versionsof the operating system and applications. So, saveon the ‘hardware’ of some interns with removablemedia going around the building using the media toupdate the computers.And think of the Raspberry Pi, really cheap hardwarethat can run Linux, etc. So, the thing is cheapenough to use for, say, a remote camera that couldupload its data to, say, a drone overhead, with, ofcourse, the drone some COTS hardware and the restsoftware.With some irony, some years ago much of academicresearch electronic engineering changed over fromhow to build electronic hardware to what can be donewith data, especially signals, e.g., with stochasticintegrals, which are a long way from traditionalhardware.Of course, the unique, world-class grand championhardware interface is the Internet so that for atleast the first few layers of the ‘communicationsstack’ any two devices with Internet access cancommunicate, right, with the rest software.Apparently currently Seagate is building hard diskdrives with an Ethernet interface. So, bye bye tothe hardware of SCSI (small computer systeminterface), EIDE (extended integrated deviceelectronics), SATA (serial advanced technologyattachment) and hello to some standard Ethernetsoftware with some more software on top of that.Gee, think of the analogy of humans or even just akitty cat: The ‘I/O or ‘transducers’ are reallypretty simple, and what’s crucial is the ‘software’.So, right, have the peripherals COTS and usingstandard interfaces, and then plug together thehardware for a new system like building with Legoblocks and have the rest just software, also quitemodular.

    1. Timothy Meade

      Printers now are just servos and sensors, with software either in the device (needed for ‘cloud printing’) or in the host computer (the old Windows GDI printers) driving them.Seagate has the right idea, but the wrong execution. They should look at the ATA-over-Ethernet (AoE) protocol and adapt that to their usage. They seem to be caught up in the cloud and some theoretical Storage Area Network technologies like OSD. They have created the next Itanium where they could have easily proposed the next AMD64.

  42. Mordy Kaplinsky

    My take on the topic is:1. If a component has meaningful cross-over value for other apps, the sensor/capability will be embedded in a smartphone or other general purpose hardware and developers will build applications on top of them. If however it’s a unique capability for a single use case it will require a dedicated peripheral.A case in point is the M8 motion detector which has uses is multiple app use cases, as seen with Google doing other work to essentially solve the same issue, but as Apple proves that doing it at the chip level is optimal.Medical devices fit into the above category. While some capabilities have cross over value with fitness, many sensors are dedicated for a relatively niche market and there’s no justification for smartphone makers to embed them into devices.2. Application that require dedicated remote sensing or an actual presence will either be packaged into a dingle device with multiple sensors, or if it’s a single data point that’s required and there’s no justification for the combo it will simply have sensor+power+connectivity.The medical device and the connected home are both examples where both of these scenario’s can play out.3. If a particular sensor has an intrinsic thirst for power which for whatever reason cannot be quenched with small sips, it will be relegated to a dedicated device unless its a use case where it can be plugged in.4. Requiring dedicated hardware (that will not become obsolete per the gist of this blog) is a double edged sword, for on the one hand its more difficult and costly to gain market share, but on the other hand once you’ve gained that market share its an even more difficult barrier for competitors to cross. Also the loyalty of users that “invested” real money in purchasing an accessory is hard to understate.An interesting note which can change the dynamic and needs to be taken into account is the PhoneBloks/MoTo ARA project. If one can effectively buy modules to suite their specific needs, one can get a suite of app+module combos and embed them into their phones. This project can also go beyond phones and apply to tablets or to various IoT implementations akin to the way hobbyist/hackers use Arduino.

    1. Timothy Meade

      You should talk to Jeff Hawkins about PhoneBloks.I’m with Steve Jobs on this one, a fit for purpose dedicated hardware and software combination is preferred over a modular approach the consumers don’t understand.

      1. Mordy Kaplinsky

        I agree when it comes to phones but the connected home will require that you can customize your home, as with any decor. Homes have very different use cases than mobile devices.The complexity of training customers is why I think that whoever does power the connected home/lifestyle will require Apple store style galleries where customers can be educated and assisted in this new lifestyle.

        1. Timothy Meade

          I think the reason there’s no smart home is there’s no IKEA of the smart home.This is a role that Best Buy or HH Greg should take before they disappear. Why have LCDs in one part of the store all piled up and refrigerators in the back getting ignored?I want to know that all brands work together, that adding a certain sensor won’t confuse my watering controller or whatever. I want a basic control panel that glues on the wall and introspects every device on the network, then decides what form it’s interface should take. (Lighting in the AV room, security and garden in the foyer, etc.) But always the same device, a basic LCD with Bluetooth LE and ZigBee with a low amp AC or DC feed that can be run out of the wall without the need for conduits or new work boxes. I grew up with X10 and it’s licensees’ products, nothing really evolved to replace the basic transmitters and bring them into the modern age.

          1. Mordy Kaplinsky

            What u’r saying fits with the gallery concept I described. I dont want a stack of devices but devices in context and ways to use them, along with any help getting started or setting up new devices akin to an Apple Genius Bar or their training classes.Hence Gallery which is more like an interactive museum or Ikea style immersive experience.

          2. Timothy Meade

            Best Buy should have applied their Canadian acquisitions better to their merchandising. The Best Buy Magnolia stores I’ve been in have been the most internally consistent part of the stores, though understaffed.Their only take away seems to have been to shrink the SF, but they’ve left so many dead zones within the stores that it has been a failure.

          3. Mordy Kaplinsky

            Can’t fix Best Buy’s management problems,i just hate that they blame the customers and the etailers but refuse to take any blame.

  43. Timothy Meade

    We first, it’s app store distribution vs. the CueCat, even free hardware has to be manufactured and distributed to the masses at a logistical cost.Smart TVs aren’t designed to adapt very much, limited app APIs make it hard to build anything more complicated that another wrapper of a video podcast, whether for Roku, Samsung. Google, and probably Apple will at least deliver a full application environment, giving access to input devices, protocols, background processing capability and notifications.The questions about medical devices are interesting, it probably takes the form of dedicated sensors for people with specific conditions apart from the more general market of health awareness devices. I think pulse rate/oxygenation could be the next big one, along with other minimal invasive technologies. There’s always this far-in-the-future viewpoint of remote surgery and being always connected to your own personal physician, but this isn’t borne out in the real world and economy. The ratio of patients to doctors is too great to expect real-time monitoring for most patients. Something that could page a nurse could be more interesting, if you are in a retirement home or have a home health provider.The last thing I want to discuss is what I think the smartphone will ultimately turn into: a single solid piece of hardened glass/plastic hybrid where every square millimeter can emit signals including colors of light, infrared, sound, tactile regions and respond to heat/infrared, touch, proximity, and pressure. Something like this could be easily adapted to gather information from a body it was near, identify the user, communicate on multiple different frequencies and standards simultaneously including near-field, low power RF, wireless broadband, Wifi, Super WIfi etc., using a versatile SDR and MEMS RF path. Software updates could actually transform the hardware, what networks it works with, and allow introducing new lower power protocols even after the device has left the factory.It would be the pocket hub of your smart watch. (If we all go back to wearing watches ones that also notify the user of events on their phones will be popular, but I don’t think that dedicated smart watches will necessarily be the focus. The AVC party was the first time I saw a modern smartwatch, and two other people mentioned having one.)

  44. sachmo

    As a hardware engineer I think this posts sets up a false dichotomy between hardware and software. It’s really not one or the other, because in most cases, hardware is built for things that simply cannot be done by your smartphone.I’ll give you an example…Check out –…These are both new types of bike lights that aim to light up the cyclist peripherally. They are ‘intelligent’ and use sensors to determine phase of the wheel in real time. It’s even conceivable that a smart phone may connect to them someday and control how the LEDs light up.But the lights themselves will never be replaced by software.I do think that the hardware of the future will be more intelligent and by nature have software components in it. I think they may be required to have accompanying apps to support the hardware, but replaced by software? Hell no.I’m not really sure what the Internet of Things is. I’ve heard about it a lot, i think the idea is sensors continuously monitoring everything.I think we’re still a few decades away from really seeing this come to fruition. The #1 application that comes to mind for me, is general stuff around the house.Turn on the sprinklers? Wait, the home computer control unit checked the moisture contents of the yard, and it seems fine, no need.There’s a leak inside the wall behind my shower head. How do I know? Sensors embedded in the piping are alerting me.It’ll be sometime before this happens, b/c building codes need to change, and standards still need to be written about how very simple sensors can be integrated into basic materials.I think what is going obsolete is ‘dumb’ hardware.

  45. Scott Jenson

    Fred, you make a classic point: adding software to ‘a proper phone’ clearly can substitute for hardware. But just because an app *can* replicate hardware doesn’t mean it’s equivalent. At some point the ‘swiss army knife’ effect kicks in and the hassle of doing things through a crowded app launcher starts to count against it.It all depends on how much interaction is involved. With the new Lockitron, there is zero interaction, you just walk up the the door and it unlocks automatically so win all around. However, anything that requires me to pull the phone out, unlock it, find/open the app, etc needs to be pretty important to worth my trouble. Sometimes it’s just better to push a button.The valley likes to take bets on whether Google is getting better at design faster than Apple is getting better at the cloud. The same might be said for your hw/sw dichotomy: Is smart ‘iot hardware’ getting interactive faster than phone apps are good at mimicking hardware? We’ll see. My point in the end isn’t that earth shattering: design friction plays a role and needs to be included in this conversation.

  46. Pete Griffiths

    It’s all about the sensors.You can only process the data you capture so the question is how do you capture it?You can process on a general purpose device e.g. a phone, but you can’t necessarily capture with such a device.Hence the issue is all about sensors?What data can we capture we currently don’t? And why don’t we? Do we not have the sensors at all? Are they too big? Too heavy? To expensive?…

  47. Sebastien Latapie

    I personally am looking forward to getting these things happening on my device. I think its a great way to be more in touch with your body and how it’s reacting. The constant monitoring would allow for a host of health benifits including early diagnosis of underlying conditions. The earlier you know the sooner you can take steps to improve your health.Eric Topol strongly champions these ideas and even though this Ted talk was given in 2009 I still find it captures many of the thoughts above:

  48. ryan kulp

    Wrote my own post on Svbtle, “Hardware vs. Software”…

  49. Todd Breeden

    To your question on health care diagnostics on mobile, interesting company does this with thermometers through your smartphone to produce geotagged map of fever onsets in communities. Still a bit of hardware involved, but IA is currently involved as Series A, and looking to see what sales/adoption rate will be v. traditional options.

  50. Michael Rattner

    My company works on an emerging hardware based medical technology. It seems unlikely that your cell phone will be doing MRIs anytime soon. Processor power is fungible, you don’t really care where it comes from and why you need it. We happily switch between CPUs, GPUs, and ASICs because fundamentally all these things are optimizing for the same thing: maximum performance at the lowest possible power.There are some generalized sensors that are pretty easy to make and that have generalized uses. A pressure sensor and a microphone are the same thing. A camera and a blood oximeter are pretty similar. And a GPS is basically some antennas and a clock. But even if it were physically possible, there just aren’t that many generalized applications for a low power 3 Tesla magnet. It can’t be used for many activities other than MRIs.Likewise, we work with some pretty cool specialized lasers. Right now the module we use is the size of a couple of soft cover books. There’s no reason it can’t be made a bit smaller and lower powered. But even if it were one cubic centimeter, I can’t imagine that it would be useful for many applications outside of where it’s used right now. It still wouldn’t be worth the space it would take up inside of a cell phone.Think about a modern car. Very often they have somewhere in the neighborhood of 30 processing units – or more. I want the phone to replace the brains of the AV unit and the GPS. But the computers that control fuel injection or battery management (in an EV)? Even if my cell were completely reliable and fast enough – there’s just no point. The car is a specialized piece of equipment that is separate from my phone. It does things my phone will never replicate. And there’s no reason to bring my engine or my engine management computer with me when I’m away from my car.So for a lot of machines that interact with the physical world, I think they will get smarter and have smarter phone interfaces, but it’s unlikely that our phone will replace most of what we use to interact with the world. And physical constraints are ultimately very limiting in this regard.

  51. Michael Aronson

    Clearly, as others have said, Iphone 5s is “tipping” point. Our portfolio company Neat’s scanning app on my Iphone 5s is almost equivalent in quality of end results to dedicated scanner hardware of which they have sold millions. Speed issue has pretty much gone away as well with new processor and wifi pretty much everywhere. Although as I like to remind people you can now run the app on a $700 device and gets results “almost as good” as special purpose devices that cost in the 10s of dollars. Future might be in standalone devices that are multipurpose and have embedded processors/storage and connectivity. In home health for example, many of the people who need it dont have an expensive smartphone and can’t use a device that “fits in your pocket”Using Moves now ,gave up on connecting fitbit to my 5s

  52. Vasudev Ram

    I have to say that I hope that doesn’t happen – the combination of software and a smartphone *largely or fully* replacing hardware, that is. Choice and variety are good. Even if smartphones + software can make custom or dedicated hardware less needed, some things may be better done with the latter. And I don’t prefer total homogenization, something like what you said about Android and Apple recently. Of course, trends will be trends … Just my opinion.

    1. Vasudev Ram

      And I say that as a guy who has been totally into software, not into hardware, throughout his career though I did have a bit of a flair for (UNIX) hardware-related software troubleshooting some years earlier, when I used to work with it a lot as a system engineer for a large UNIX vendor.

  53. Vasudev Ram

    Anyway, after a quick scan of the whole thread, got to say it’s a pretty interesting post, Fred. You seem to have a talent for coming up with such 🙂 – as I’ve seen before too, and so have others. Going to read the whole thing over a day or two.

  54. Dan Goldin

    And the smartphone’s the central place for this! I had a similar thought a couple of months and wrote a quick post about it –

  55. Mike Post

    Using the smartphone as a platform to solve outdated hardware is nothing new anymore, it’s been like that since the first iPhone was launched. But I agree, it’s happening more.It’s also why it’s so important for the leading hardware suppliers (Apple, Motorola, Pebble, whoever) to keep on innovating. I’m sure there’s also developers out there who want to bring MRI scans to people’s homes, but are just “waiting” for that to be possible.One MAJOR problem with software though. Software doesn’t raise any money.Not when it comes to crowdfunding anyway. Crowdfunding isn’t the best investing strategy, but it has merits as an extension of bootstrapping, early on in a product’s life. People are willing to pay for hardware. Compared to hardware, people think that software, especially pre-release is worthless and should be free.This “free” perception problem, is a major deterrent for a lot of developers who genuinely want to innovate, but can’t justify it with not getting anything in return. I understand that crowdfunding and bootstrapping isn’t a priority on an investor blog, just raising the point though as a factor of innovation in general.

  56. Manish Gaudi

    Love what these guys are doing: I don’t think any of the options you listed are crazy – all a matter of time. Also, the ability to connect ancillary devices to a smartphone enables the innovative capabilities of the phone to spread to more complex sectors like this one.

  57. Alexander Peschkoff

    Where does that $35 dongle come from?.. Oh my, it’s h/w…Funny how VCs in general developed h/w-phobia over the years, although that’s where it all started… Yes, the incremental cost of an additional s/w user is zero (if we disregard user acquisition cost), but so is the cost of replicating one’s app/code (aka the barrier to entry)…VCs forgot what made Apple a great company. (End to end UX). As Steve Jobs put it, anyone serious about his s/w, should be serious about his h/w.Try doing Goggle Glasses without h/w… Or mass transit. It’s tempting to shoehorn it all into a smartphone, but the real world is not just about a piece of code.

  58. Alexander Peschkoff

    Strange, my comment vanished into the thin air… There we go again.Where does that $35 dongle come from?.. Oh my, it’s h/w…Funny how VCs in general developed h/w-phobia over the years, although that’s where it all started… Yes, the incremental cost of an additional s/w user is zero (if we disregard user acquisition cost), but so is the cost of replicating one’s app/code (aka the barrier to entry)…VCs forgot what made Apple a great company. (End to end UX). As Steve Jobs put it, anyone serious about his s/w, should be serious about his h/w.Try doing Goggle Glasses without h/w… Or mass transit. It’s tempting to shoehorn it all into a smartphone, but the real world is not just about a piece of code.

  59. Zach Dunn

    Hardware is a way for software to ask better questions. In that context, the IoT becomes more of a software problem than a treadmill of new devices. I wrote a bit more on this thinking here:

  60. paramendra

    Internet of Things. Obviously all of the above. No?