Devices vs Cloud

Yesterday, on stage at an event hosted by our portfolio company Disqus, it was suggested that I was “trolling Apple” with the comments I made at TechCrunch Disrupt. I explained that I was not trolling anyone and that I attempted to honestly answer a question about the changes afoot in technology. I think there is a fundamental and important distinction between a device focused strategy and a cloud focused strategy.

Carlos Kirjner is an analyst at AB Bernstein who covers Internet companies. I was reading his analysis of Larry Page’s letter to shareholders this morning (the analysis is not a public document and I cannot link to it).

In Carlos’ analysis, he wrote this:

We believe … Larry Page’s discussion about the new mobile, multi-screen world …. is really about the importance of cloud services in that world. This is by no means a trivial statement and we believe goes against a more device centric model favoured, we believe, by Apple.

Many interpreted my comments as anti-Apple and pro-Google and I guess they were. But I was attempting to make a larger point. Which is that a device centric strategy is not a winning strategy in my mind. The big gains from technology in the coming years will come from things like machine learning and collective intelligence. Hardware and operating systems are important but to some extent a commodity at this stage of the game in mobile. Yes, we will see more sensors, better screens, better battery life, and more and more technology packed into these mobile devices. But I don’t think any one company has a lock on all of the device level innovation and I worry that one company, Google, is developing a very large and sustainable advantage in machine learning and collective intelligence that will be hard for anyone to compete with.

So when I look at which top technology company is best positioned for the next decade as I see it unfolding, well that’s an easy answer in my mind and that’s the answer I gave.

#mobile#stocks#VC & Technology#Web/Tech

Comments (Archived):

  1. awaldstein

    What’s your thought on how machine learning and collective intelligence will be monetized? If devices and OS are all commoditized (not as convinced absolutely as you though) then indeed the biz model of the future is________?Would like to think that as we evolve so will the models for monetization. So far this is not the case. New value, same old models.

    1. William Mougayar

      Your first question is exactly what popped in my mind. I agree with your statement Arnold.

    2. Barry Nolan

      Google is evolving from PageRank to PersonRank. The interaction with the ‘Person’ will increasingly rely on Personalization Technology.Netflix were talking about this yesterday. In their orbit, this involves replacing the grid of suggested shows with very few recommendations. Presenting viewers with just three or four choices is “a powerful possibility.”This future means less messages/ads, replaced by highly contextual ones. From our work @Converser, personalizing customer messaging has an order of magnitude impact on conversion.

      1. awaldstein

        You have my attention….PersonRank means what in context of search and discovery? Go deep and get geeky on this please.

      2. Richard

        There is no one model for personalization. Moreover, people are dynamic not static. As just one example, there are products that you already like that advertisers want to keep current, there are products that you don’t know about that advertisers want to introduce you to….knowing what ad to place, when, where and at what time is the challenge for advertisers. Personalized advertising is like personalized medicine. It sounds nice in theory but google ads are reality.

      3. Pete Griffiths

        I agree. Ubiquitous personalization is becoming the norm.

    3. fredwilson

      Attention based models (media and commerce)

      1. awaldstein

        I’m sure you are right. Personally–seamless integrated transactions is always something that feels like the future just around the corner.

      2. kenberger

        Yes, that’s a winner. This is the area that has my keen “attention” these days, looking to the future. There is some serious innovation to be done and to be gained by a far-reaching swath of companies beyond Google. My dev group is doing work here, including for Twilio customers.I now notice rapid sudden attention to this field breaking out from deeply geeky to approaching almost mass-adoption phase. Matt Turck of Firstmark runs a meetup called “Data Driven NYC” (I’m attending tonight), attendance is exploding– always a good early signal of a tech area to watch.

      3. JamesHRH

        Ugh – I am sure you are right also.Gary V is right: you are a media company, then a ____ company.This is totally my thing (media guy) and yet it makes me depressed to think that the internet will only really enable 2 basic activities:- direct data entry (which will never become seamless and universal – really think my medical & financial info will just be there in a secure cloud service that I manage? 2050?)- a exponential increase in the amount of and importance of social capital & the attending noise / BS / posturing / narcissism.How about this beaut from Jeff Bussgang this AM:http://bostonvcblog.typepad…Me Generation 2.0 – more aware, more sure, more dangerous than ever.

      4. SubstrateUndertow

        Is attention based models a synonym for empowerment based models ?

    4. Pete Griffiths

      The Google business strategy has already generated a new business model. The prevalence of free online services is enabled by ad targeting and retargeting.

      1. awaldstein

        I understand this Pete. Just saying that this ain’t near new and the world is changing and the models aren’t.AND–if media is the model, the threshold for startups is getting steeper each and every day as under 1M uniques a month ain’t really much of a media business any more to my quick calculation.

        1. Pete Griffiths

          The biggest potential shift in revenues remains to significantly onboard brand advertising. If it becomes possible to demonstrate that brand values can be effectively communicated in a personalized manner and brand advertising funds can be spent with real accountability and reporting the resultant wealth transfer will be profound.

          1. awaldstein

            I get what you are saying potentially but having trouble wrapping around the who cares of it.Are you simply saying that the top tier advertising companies will be more efficient or that efficiencies rolls out across the long tail on this?Gotta be the former.

          2. Pete Griffiths

            I don’t know how to respond to the ‘who cares’ point. I think the original question was about new business models. I’m getting lost. :)btw may be buying a place in NY – so there may be wine and cheese in our future 🙂

          3. awaldstein

            See ya in NYC then one of these days.

          4. Pete Griffiths


        2. Salt Shaker

          Not sure 1M uniques is even the requisite hurdle these days. A start up that’s ad rev dependent is swimming upstream unless they’re somewhat of a game changer. So many parity products in categories that can best support one or two brands. Yes, there are more ad dollars migrating to digital and mobile but the real beneficiaries are the anointed few.

          1. awaldstein

            You are probably right but we agree on the core point.

      2. SubstrateUndertow

        The advertiser is a payment middleman that ultimately gets to hijack/subjugate the steerage of intent from end users whom ultimately pay for those media services.This business model has historically worked well to coalesce and focus financing for media services.But as we move into big-data-mining media-services that implement machine learning and collective intelligence, where the ultimate intent of end users is personal and collectively-distributive empowerment around unadulterated evidence-base-data, that ad financed model will becomes evermore counterproductive and unsustainable.The only question is how long before end users reach an awareness tipping point and insist on direct control over the steerage of such crucial personal/social capital.Yet another instantiation of:”History is a race between education and catastrophe”- – H.G. Wells

        1. Pete Griffiths

          With the ‘free’ business model the overwhelming majority of users more or less happily go along with giving up data on their interests for free services. I don’t see any contradiction between the success of that business model and its refinement with bigger data and better data mining. All that signifies is better targeted ads. And one can make a good case that more relevant ads are more valuable to the consumers enjoying the free services. So I don’t understand why this becomes ‘counterproductive and unsustainable’ any more than I can understand why better recommendation systems would be a bad thing. As for end users reaching some tipping point, any putative user rebellion against the steerage of their personal social capital will only be as successful as their willingness to wean themselves from the free services that they currently enjoy. So I am very skeptical of the likelihood of a peasants’ revolt any time soon.

    5. Matt A. Myers

      That intelligence alone isn’t valuable, it needs to be coupled with practical means.

    6. Michael Brill

      Selling things.If you assume that a “machine learning” world means that more things are done for you better than you could do yourself, many of those will end with you buying a product/service. Since advertising (and many forms of traditional marketing) don’t exist in a data-driven world, the $2 trillion that companies spend on consumer marketing seems like a nice pot of money to go after.Also I wouldn’t rule out subscriptions. If Nest2017 is completely integrated with my family’s location/schedule/etc and Google strikes a deal with power companies to help smooth out demand, and all that saves me $25/month, maybe I’d pay $10.

  2. jason wright

    and so is Google contradicting itself by pushing the Project Ara concept of the modular mobile device, or is it just trying to poison the apple tree?

  3. WA

    Did I just see a potential twitter seeking sound cloud deal on Bloomberg? Congrats…once again. Changing the world, one deal at a time.

  4. William Mougayar

    Well, until Google proves their revenues can be well diversified beyond search, they have a long ways to go before claiming they will own the cloud.And most of these intelligence services are not end-products; they are part of other products or services & many of them free.My prediction over the next 10 years? Google will continue to be Google, and Apple will continue to be Apple.

    1. awaldstein

      I’ve put this on my calendar to ping you my friend 😉

      1. William Mougayar

        You bet. Over natural wine & raw cheese.

  5. Seenator

    Would your view change if there was an iWatch? Because that is an important differentiating enough device that can lock you into the Apple cloud? Also- since the iWatch will generate enough data to feed the machine learning beast & maybe Apple could get better at it because they have a better data set (based on the assumption that data beats algorithms).Here is another great piece from a blogger on the iWatch, Device Uniformity & the apple cloud:…For everyone doubting Apple’s “closed” ecosystem, this is how Apple shines. The iWatch will be able to support the majority of Apple’s other product lines because it will use unfragmented and uniform software. An Android product that works like this potential iWatch will be extraordinarily difficult for developers and engineers to create. Uniform software and uniform hardware will allow the iWatch to perform its health and fitness purpose seamlessly across all Apple devices.And I just don’t see anyone other than Apple capable of pulling it off.=======FYI: I saw you at the event yesterday standing in line & wanted to come up & say hi but wasn’t allowed to enter since I didn’t register in time & apparently there was no space. Bummer 🙁

    1. SubstrateUndertow

      It all comes back to Job’s vision that Apple aspires to be at the intersection of:Liberal-Arts/Technology INTEGRATION.Which if we zoom out, for this conversation, translates to the intersection of:Collective-Mind/Distributive-Body INTEGRATION.orCollective-Mind-Cloud/Distributed-Device-Body INTEGRATIONTight-tight-tight INTEGRATION is the magic-mojo that breaths intelligent tour-de-force into the human Mind/Body INTEGRATION.It is that, “non-trivial-causal-spread”, that Mind/Body INTEGRATION that empowers our transparent indwelling sense of consciousness.Technology brings its true magic when it mimics/extends our sense of transparent indwelling within a technology’s extended tour-de-force. When I drive my car it feels like a transparent extension of my mind/body.Earth to SubstrateUndertow get back down here !The iWatch will no more be a watch than the iPhone was a phone. The iPhone was an integrative nervous-system extension empowering us to transparently indwell within an expanded organic world of collectively abstracted(discarnate) possibilities.The iWatch will likewise be a wearable extended nervous-system INTEGRATION-HUB, facilitating our sense of transparent indwelling within a larger personal/collective array of distributed device sensors and actuators(internet of things).In the long run tight-tight-tightCollective-Mind-Cloud/Distributed-Device-Body INTEGRATIONis an unstoppable gravity well.In the short run maybe Google does not have to lose for Apple to win and vice versa?Maybe both side of that collective-mind/body INTEGRATIVE cyberspace equation have a long ways to go before we hit indwelling critical mass?Things that grow together work together – long live Apple, long live Google !

  6. JamesHRH

    The is a third rail – a company that provides solutions (cloud services & awesome devices).Neither seems able to get to that point. Apple web services have no reason to suck as hard as they do, other than no one at Apple is any good at all at thinking about them.If Apple bought Dropbox, & made Drew & Jony their leading design team, that would be scary.Who could Goog buy on the device end that would scare anyone?

    1. William Mougayar

      They got Tony with Nest.

      1. JamesHRH

        Yes!I wonder if the phone handheld is the search of the IoT?I know about Nest, have seen the product in a store, but it’s not on my wall yet. Not as urgent a need….?

        1. BlairMacGregor

          I think Google is definitely setting up something to act as a “universal hub” for IoT, whether it’s the existing Nest device or a new product that will come from the Nest folks. Which will of course be controlled by your phone like a remote control.As for the third rail, I think the reason you’re not seeing it is because it’s really really hard to get both software and hardware right. Apple’s come the closest with products that control the entire stack but as more services become cloud-based, their weaknesses become more apparent. Conversely, Google has failed as a hardware company so far. We will see what becomes of Nest but Motorola was basically a write-off. Microsoft has also never been a hardware focused company and when they’ve gone in that direction, aside from the Xbox, it has been a flop (e.g. Zune)It’s almost like you have to pick one or the other as your core competency. Apple has hardware, Google/MSFT/others have software/cloud. Both can co-exist in the same ecosystem. So stepping back, I think Fred’s point is totally valid that Google is probably better positioned to where the market is headed. It’s just that we’re still going to need quality hardware/devices for the foreseeable future and I don’t know who steps in to fill the void if not Apple.

        1. William Mougayar

          That’s funny, I’ve been looking at a Chamberlain garage door opener with the 1Q for the past month. Do you have one?

          1. pointsnfigures

            No, but I live in a building so I don’t need one. When I had a garage, we had Chamberlain. The company is based in Elmhurst, IL. I don’t think this is the last iteration of their technology.

          2. William Mougayar

            Cool. I’m trying to repair an old Stanley where the chain dislodged. If I can’t, a new Chamberlain is going in.

  7. William Mougayar

    Actually, the company that Apple might fear is not Google, but it’s Xiaomi. They are now neck to neck on smartphone sales in China.

    1. Pete Griffiths

      I agree. Whether it is Xiaomi or an as yet nascent Chinese company as soon as a category of devices becomes commoditized a Chinese manufacturer with a gigantic domestic market in a price sensitive region can become a major threat. These companies are on the verge of being a real danger not just to Apple but to Samsung.(check out the OnePluseOne flagship killer phone)

  8. JimHirshfield

    Sound bites. Every journo is looking for a sound bite.

      1. JimHirshfield

        I tweeted that out, hahahaha.Nice wrap up quote:Many titans of the media think the public is stupid. What we are finding out again and again is the public is infinitely smart. They make decisions for themselves better than oligopolies or bureaus do.

        1. pointsnfigures

          thx. Funny how the people in power don’t think so….

  9. Sebastian Wain

    But Apple is trying to compete in the productivity space with iWork for iCloud what are the non-obvious ways for Apple to compete in the cloud space? I mean, I don’t imagine an “Apple Social Network”, even G+ is a laggard.

    1. JamesHRH

      No, but an Apple designed fridge, stove, thermostat, security system w an app called iHomework might be something.

      1. pointsnfigures

        Google getting there with Nest etc.

        1. JamesHRH

          Except nobody trusts them.

      2. jason wright


  10. BillMcNeely

    My iPhone 4 is starting to age out. Apple has pinged me to upgrade. As most of my documents are in the close via Google Drive I may have to give more thought of how to best access them in the future.

  11. pointsnfigures

    Remember the 90’s? Microsoft was the future with software-and Apple was dead because it was a ho hum computer company. Apple came up with a game changer in the iPod, and had great marketing around the iMac, following through with the iPhone. Devices. Then, all of a sudden it’s a device world and Microsoft is wrong. Google has both devices and the cloud-Microsoft has a cloud, and so does Amazon. But, Apple has mobile, and Microsoft+Amazon don’t. I didn’t interpret your remarks as a knock against Apple, but more an observation. Apple has some really big challenges, because clouds scale a lot faster than devices.

    1. falicon

      Apple is the modern Standard Oil…Amazon the Carnegie Steel…Google the Ford motor company…Microsoft the Firestone Tire and Rubber company…all were/are profitable companies and important in a variety of ways through their history…

      1. LE

        Well Amazon is for sure no Carnegie Steel that is for sure.

        1. falicon

          I was thinking/meaning they are the “steel” inside of most of our (internet) buildings…

  12. Sudhir Mantena

    Device vs Cloud is more of a business model question then capability question. Apple makes almost all its money through devices, so continuing to invest more in that makes sense. While opposite is true for Google, which makes all its money from Services. Even if Apple were to build a better iCloud or iMessage or iTunes, it wouldn’t make them more valuable. But without those Google is dead.But is one more scalable than the other in the coming decade, I think both will continue to survive. Apple made more profit in 1 Qtr than amazon did in its entire history. But doesn’t mean Amazon AWS is less valuable or less scalable either.So I think each is its own self fulfilling prophecy.

    1. Sean Hull

      Well said Sudhir.

  13. JLM

    .We often set ourselves up for a strawman argument suggesting there is only one future and only one big winner. This is the classic strawman problem.I just got a Slab o’ Cheese phone. It is the Samsung Mega. It is BIG.Apps, sites and the old fashioned Internet render on this phone in a quantum order of magnitude superior to its predecessor.Sure, it’s a bit of a phablet but for me the future arrived in the form of a phone which simply makes the content more useable and enjoyable.It is going to be “both” “and” rather than “either” “or”.Convergence, it’s what’s for breakfast. [Well, actually it’s poached eggs but this is a metaphor.]Lots of winners out there, not just a few.JLM.

    1. JimHirshfield

      “Slab o’ Cheese”…for sure. That’s a muenster of a phone!

      1. JLM

        .Pretty damn………………………funny!JLM.

      2. Salt Shaker

        For sure this Disqus gig isn’t your calling. You’re much better suited for writing headlines for the NYP, or even better, The Onion.

        1. JimHirshfield

          No whey

          1. JamesHRH

            Curd you cut it at the Onion?

          2. JimHirshfield

            Sounds a peeling, but I don’t think I could camembert it.

          3. JamesHRH

            I’m bleu with laughter. Brie peeing my pants soon. Ummmm.

          4. Pete Griffiths

            Have merci.

        2. LE

          I was at the IC graduation 2014 a few days ago and the speaker was a 1989 graduate Chris Regan who wrote for 7 years for The Daily show with Jon Stewart.……Being in the crowd I didn’t see him I only heard him.It was amazing how not funny he was by my standards. Some people liked him but it didn’t seem to me by the applause to be anywhere near as funny as you would think it would be.If you listen just listen don’t watch and see if you agree (Nixon/Kennedy effect).I think Jim should offer to start to do some free writing for some upstate station that is short on cash to get into the industry.Or he can spend his time selling sugared water to children.That said I was thinking the same thing “NY Post headlines” before you made your comment.

          1. Salt Shaker

            Seth MacFsrlane was commencement speaker at Harvard a few years ago. He was very, very funny….check it out on You Tube.Separately, a very close friend of mine, and one of the funniest people I know, was also a successful sitcom writer for many years (and coincidentally an alum of Ithaca college). He’s also a great columnist and has written pieces for several national pubs. All that aside, he bombed doing stand up. It’s a diff skill set, where a comic’s personality, timing, delivery, audience engagement are all key factors. Speech writing similarly has diff skill requirements, especially if one is trying to both enlighten and be funny. Even comedy has specialists, sort of like dentistry, where there are surgeons, periodontists, endos, etc.

          2. LE

            You mention timing and that is exactly what he was not good at. As well as simple vocal inflections as well.I got a standing ovation at my daughters bat mitzvah and I’d never done anything like that before. I was just surprised that he wasn’t better trained and didn’t appear to have more skill than someone with no qualifications at all.I wonder if sir Jim Hirschfield can cut the mustard in front of an audience?Oh one other thing. When a writer writes what they write is vetted by others, right? So it gets morphed into what it is by a few people’s thoughts. For the commencement speech Regan probably read it for his wife and went with it.

      3. Nick Ambrose

        I see what you did there …..

      4. JamesHRH

        Gouda one is obvious retort.

        1. JimHirshfield

          OK, OK, let’s not get too cheesy.

      5. SubstrateUndertow


    2. SubstrateUndertow

      That’s the same breakfast you had the other day 🙂

      1. LE

        A JLM “file photo”.

      2. JLM

        .It’s the same picture but I often have the same thing.JLM.

    3. PhilipSugar

      Agree completely. Where you lose margin is where you go directly to the business buyer, but that’s not Apple. The further down the stack you go the more you are commoditized. That is why Apple doesn’t cater to the enterprise. They get in the enterprise through the consumer. Brilliant. Corporate purchasing department wants to shave 30 basis points off your margin? Too bad, so sad, we won’t sell it to you and our experience and brand is so strong that the employees will go around purchasing, starting in the C suite.

    4. LE

      Lots of winners out there, not just a few.For this phenomena you can totally blame the media and now bloggers. Normals don’t see the world as whether Sony sells more TV’s than Panasonic. They buy the TV set that they want at a price they want to pay. They don’t care at all about market share.I sometimes pay attention to that strictly as a marker for a better buying decision.For example it’s generally a safe bet, when time is restricted, to buy the car that sells the best in it’s class since the wisdom of the crowds has already figured out that it’s the best bet.The media otoh is out there (here we go!) all giddy with stories and headlines in order to sell more advertising. So they are constantly talking about things and making things important that if not put on the radar of people wouldn’t matter.It’s actually a wonder that people don’t hate the media and writers more than they hate lawyers. Actually it’s not because the media is the one doing the brainwashing and telling people what is important.

    5. Pete Griffiths

      Jees – I’d never even heard of the mega. I have a Note and this thing is even bigger. I’m impressed. It’s true what they say about Texas then – size does matter.

      1. JLM

        .Well, I did have to disrobe and take a prostate exam to get it. Thought that was a little odd but who knows any more?JLM.

        1. Pete Griffiths


    6. sigmaalgebra

      She’s from Austin…

    7. Guest

      I opined on this very subject 4 years ago (…, in which I said:Apple produces a few versions of a great operating system (OSX, iPhone OS) that sit on devices it also makes (iMac, iPhone, iPad). It produces and allows others to produce applications for these operating systems on these devices (Mac-based software, iPhone apps).Google produces applications (Gmail, Calendar, Docs, Search) that work on any device with a modern web browser (i.e. in the cloud). They’re producing operating systems (Android, ChromeOS) that help advance devices to better connect with the cloud. And they’re supporting and further enabling others who produce applications in the cloud.With Apple, you’re locked into Apple hardware, Apple’s operating system, and Apple’s rigid processes. The environment is all Apple.With Google, you’re not locked into any hardware, any one operating system, and you can deliver your services without Google. The environment is open.Apple does walled garden well. That’s because they create superb, forward thinking experiences that are a combination of phenomenal hardware and software that is tightly controlled. They’ve radically simplified and enabled whole experiences because of it. People buy in to pricey Apple hardware and Apple makes money. Hardware leads.Google does open web well. They create superb web based software and experiences that are powerful and connected by virtue of the web. They’ve radically simplified and and enabled whole experiences as well. People choose Google software and Google makes money from ads. Software leads.Google’s betting on the open web. Apple’s betting on, well, Apple.

  14. Salt Shaker

    Devices vs. Cloud suggests the two paths are mutually exclusive. They aren’t and will increasingly be less so over time w/convergence. If Apple needs to pivot they certainly have the cash to do so via acquisition, although much of their cash would need to be repatriated. They’re also quite capable of creating cloud based solutions. iBeacon could be a game changer at retail. Apple certainly has options to evolve its biz model if need be. Many companies will be panning for gold…and several will succeed quite nicely.

  15. Dave W Baldwin

    You gave right answer. Too many are thinking magic niche and not seeing the bigger picture re seamless service to user.

  16. Pete Griffiths

    I profoundly agree.I think what perhaps confused people was contrasting a ‘device’ strategy with a ‘cloud’ strategy. I would suggest that a better terminology might be to contrast a ‘device’ strategy with a ‘data science’ strategy.In the future apps will simply be the front end to data science engines running in the cloud. and nobody knows more about that architecture and technology than Google. Indeed, not only do they know more already but they are pulling away.This fundamental difference in strategy could scarcely be better epitomized than by Google’s recent acquisition of the deep learning company ‘Deepmind Technologies’ and Apple’s current acquisition of the device company ‘Beats.’ Furthermore, even the companies Google has acquired they might seem device like (eg the Robotics companies such as Boston Dynamics) are machine learning driven.Once again I find my self recommening the Peter Norvig talk – ‘The Unreasinable Effectiveness of Data.’

  17. laurie kalmanson

    ? > google > apple/microsoft

  18. John McGrath

    I agree in principle, but I just switched from iOS to Android, and two weeks in it is still painful. Some of that is familiarity, but there is such an enormous gap in polish between the two, it seems like a real stretch to call the OS commoditized at this point.I am looking forward to disentangling the OS and hardware entirely, so I can have not just services but the whole OS in the cloud, able to be run on fresh cheap hardware of my choice. But I think somebody’s going to have to compete with Apple on design for that to happen, and at the moment nobody is close to being able to do that.

  19. Twain Twain

    Instead of Machine Learning (also Deep Learning) and Collective Intelligence per se, we can be more specific and say “Deep Intelligence” and “Contextual Intelligence”.Yesterday IBM Watson acquired Cognea:*…IBM Watson refers to Cognea’s depth of personalities.When I was in NY and met @jim hirshfeld I alluded to the nextgen of Machine Intelligence being about deep intelligence.It’s also interesting to read about FB’s intelligence wrt ad placing from Pando:*

    1. Dave W Baldwin

      Deep intelligence is where the big change is going to happen. This is where is put my $$.

      1. Twain Twain

        Which companies would you say are leading the way towards Deep Intelligence apart from IBM Watson?

        1. Dave W Baldwin

          Google would have to be top. They are bringing in profs from Stanford part time including Ng and you have Kurzweil going after the data gathering via books/media.

          1. Twain Twain

            Andrew Ng has left for Baidu:* http://www.technologyreview…@fredwilson shouldn’t worry about Google running the table on Machine Learning & Collective Intelligence.The article says: “However, Michael Mozer, a professor at the University of Colorado, Boulder, and a board member of the NIPS Foundation, points out that the core algorithms the neural networks use are much the same as those that triggered a surge of optimism about artificial intelligence in the late 1980s. Recent breakthroughs have come from finding “tricks” that allow these algorithms to be used at much larger scale, says Mozer. “The people that stuck with it are deservedly reaping the benefits now,” he says, but deep learning is not such a big leap forward for the field as it is sometimes made out to be.”

          2. Twain Twain

            And FT says “Big Data: big mistake?”*…A new paradigm may yet emerge for data and machine intelligence that Google does not have and is not about.Google has already declared its AI approach:*…I asked Amit Singhal of Google whether their Star Trek machine is going to understand our emotions.He

          3. Twain Twain

            He said, “No we’re focusing on facts and figures.”Meanwhile, Apple and IBM Watson are already moving towards emotional and contextually-aware AI.

          4. Dave W Baldwin

            Thanks for update and article. Having the focused research in a growing number of companies is going to make a big difference.Only thing in article I question is the sweeping 70% accuracy on the cats project. When NG did the condensed project via parallel, the accuracy went up.

  20. Brad Lindenberg

    The problem with Apple is that everything they do has to be 1000% perfect. Their branding, product, UI/UX, and all experiences inbetween. Apple is like a supermodel that has to be emaculately dressed before they can leave the house, and in order to reach this level of perfection they can only do very few things very well before they start spreading themselves too thin. You never see Apple ‘experimenting’ with things like self driving cars or beta hardware like Glass. It’s not in their DNA because it’s hard to perfect this up front. Google on the other hand, very early on, made Beta a part of their DNA. Remember how long Gmail was in beta for? They can get out of bed and wear whatever then want and nobody judges. This means they can experiment and release new adhoc services and people don’t judge if they are not perfect. The problem Apple has is that without ‘beta experimentation’ in their DNA, they can easily miss many huge markets by trying to do everything too perfectly, and that’s a real threat over the long term as scrappier competitors like Google can try new things faster.

    1. Michael Brill

      This is a great point. Perhaps in Jobs they had a leader that enabled them to deliver on perfection but no longer. Add in the need to only focus on multi-billion dollar opportunities and that really seems like a recipe for never shipping.

      1. LE

        People see this as perfection but it’s really not. It’s really the pursuit of having a clue and realizing that all members of the organization need to think before they act from the customer’s perspective.I was at the bank this morning and was taken right when I arrived. But it still took the person who was helping me 1 hour to figure out how to get certain paperwork in order and print things for needed signatures. Totally ridiculous. Obviously that organization is not driven by a “jobs”. And unlike an airline (who doesn’t make money so they have to scrape by with mediocre customer service) this institution makes plenty of money and should really have clue full people in charge who make the customer experience way quicker.

    2. LE

      You never see Apple ‘experimenting’ with things like self driving cars or beta hardware like Glass. It’s not in their DNA because it’s hard to perfect this up front.I would say that Apple doesn’t do this because there are adults in charge and those are actually foolhardy pursuits that have only a small chance of paying off.Not to mention the fact that they are (iirc) driven by the personal interests of the google founder(s) or top execs. Or their boredom. And they take those chances because all the money they get is driven by the one trick pony. So they have to throw some shit at the fan in order to get the next blockbuster drug. Because they don’t want to end up like the newspapers (they must have been reading business books or studied business history about getting fat, lazy and complacent).Apple on the other hand creates things that are really way closer to what people need. That pursuit of perfection that you speak of has allowed them to take products that were already out there and improve them in a way that made them way more valuable. Google has done nothing of the sort.I don’t think anyone can seriously argue that the contributions that Apple has made are vastly more real than any of the google initiatives. Specifically google glass and self driving cars. While many of their other offerings are a complete and total clusterfuck in user interface. God knows you can’t even talk to a google employee if you have a problem with anything they offer (unless you are at some higher tier as a customer).

  21. LE

    it was suggested that I was “trolling Apple” with the comments I made at TechCrunch Disrupt.Saying someone is trolling is the lazy person’s way out of presenting a cohesive counter argument. It’s so easy to say that someone is a troll. Costs nothing to do so. Not to mention that most normals don’t even understand what (the fuck) trolling is anyway. (Otoh everyone knows what porn is and what an “asshole” is.)

  22. Liban Mahamed

    Fred another excellent topic. The first part of your thesis, the demise of the device economy is already here. Margins are collapsing, plus market saturation in many developed countries. The growth of smartphones will be in developing world, they can afford at most $100 phones. It is very likely Apple will be a niche player in mobile devices just as it was in personal computers before Iphone. Sort of the BMW of mobile phones, small company with luxurious brand and tiny market share.I am not sure about Google in the future either. Google is very vulnerable to any changes in internet ad market. Over 80% of Google revenue comes from search engine advertisement. Any change in the internet advertisement will basically shake their business model.On the revenue and profit front Google is one product company, true they dominate this industry but it is very precarious situation. Smaller search engines such as DuckDuckgo etc could one day develop cheaper more efficient way to do internet advertisement.In the next 5 years Google faces more danger than Apple. The emergence of one search engine can endanger their only revenue source. This could be fatal for Google. Advertisement money is the main source of revenue.Google’s investment in cars, Android, artificial intelligence etc, are all bets in the future, they may succeed , they may not. As of now they do not generate any substantial revenue.They both face critical problems but Google is more vulnerable in the short run due to heavy dependence on one single service Internet search and advertisement.Apple will survive as a small niche player, the BMW of mobile phone and computers.

  23. stefano zorzi

    Well, it seems to be quite easy to be accused of “trolling” (what that means then…)I think you have a good point especially if we consider “relevancy” of tech companies/brands in the future, but this could be a very different discussion than revenue/profitability. It is not automatic that a not-so-relevant company will also be a low market cap company. After all apple could still have 20+% marketshare and keep charging a premium on its devices based on design and brand. I don’t expect hardware to disappear and, as fashion clearly shows, the fact that something is a commodity doesn’t mean a few brands are not charging for luxory

  24. Aviah Laor

    There is at least one device that is missing: a portable, powerful tablet that enable to produce content and art and not just consume and “follow”. Powerful enough to run the heavy software required for graphics, animation, multimedia books, comics, with capabilities of a good touch pad, (now it still expensive) , with the relevant software packages re-imagined.

  25. SS

    FWIW, it is worth mentioning that past performance is not indicative of future returns. Go figure.

  26. Semil Shah

    There are some differences in hardware that will make iOS devices less of a commodity. The M7 chip is a good example of this. This will envelop some current wearable offerings and attract users who want certain apps…which relates back to app development, where it will remain much more efficient for small startups and individual developers to build for iOS first. Beyond M7, the network effect for iBeacons or even AirPlay or AirDrop protocols could be massive and not possible even if some of the higher end Android devices are close to hardware parity.

  27. Teddy being Teddy

    Fred – good post. if you wanted to get into the beer biz, would you rather own one beer brand, or own the distribution of a bunch of brands?Between a device and a cloud, which of the two is the distribution channel and which is the product being distributed? Does it matter?

  28. sigmaalgebra

    > The big gains from technology in the coming years will come from things like machine learning and collective intelligence.> I worry that one company, Google, is developing a very large and sustainable advantage in machine learning and collective intelligence that will be hard for anyone to compete with.I’m not too sure just what ‘collective intelligence’ is although maybe one case is some case of ‘user contributed content’, maybe get the data for ‘page rank’ from links entered by humans who in a sense have evaluated the pages; so, humans are the real sources of the data. Okay.For “machine learning”, maybe something good, useful, and valuable can be called that, but so far from ‘machine learning’ it appears that we are getting some quite old, applied statistics done poorly and some parlor tricks. Maybe some money can be made doing such things, but also I have an uncle who could sell ice to Eskimos in winter.One reason Google stockholders might be disappointed and why one might not expect much more from Google is in their recruiting as in:At…is Why Google doesn’t care about college degrees, in 5 quotesAt…is Part 1 How to Get a Job at Google FEB. 22, 2014 Thomas L. FriedmanAt…®ion=FixedLeft&pgtype=articlewith How to Get a Job at Google, Part 2 APRIL 19, 2014 Thomas L. FriedmanSo, the Google head recruiting guy seems to have his idiosyncratic ideas on what a good college education should be. Bye bye Ivy League!A key point is that the HR recruiting guy seems to believe that a bright guy without much in formal education can ‘take it on’ and ‘get it done’. Would the recruiting guy really want to have such a bright guy take out his daughter’s appendix?Sorry, HR guy: There really is some rock solid, quite powerful stuff out there where, without quite a good start in formal education, just never but never will get to this stuff — won’t get there, won’t understand it, won’t be able to apply it, “the secret sits in the middle and knows while they [the bright guys without a good formal education] dance ’round and ’round and suppose.”.E.g., so far sometimes ‘machine learning’ does ‘optimization’ and stands to do more of it in the future. E.g., let’s have some ‘machine learning’ to run an oil refinery in Houston! Well, one of the pillars of optimization is the Kuhn-Tucker conditions. There, just for a little review test, argue that, for problems in functional form, the Zangwill and Kuhn-Tucker constraint qualifications are independent. You are welcome to use Google and to follow any links you find there. How about 24 hours for this one?To see farther than others, one way is to stand on the shoulders of giants, and for some of the giants quite a lot in formal education is necessary to get to their shoulders.In the third reference above, the HR guy said:“I took statistics at business school, and it was transformative for my career. Analytical training gives you a skill set that differentiates you from most people in the labor market.”Gee, the guy got his “transformative … analytical training” from a B-school stats course! Just one course? Must have been quite a course! The mind boggles!Since Google likes questions, here’s one: What does it mean to say that sample mean and variance are ‘sufficient’ statistics for an independent, identically distributed sample from a univariate Gaussian distribution? What does it mean to say that order statistics are “always sufficient”? What is the connection between sufficient statistics and the Radon-Nikodym theorem? Explain the Radon-Nikodym theorem. What do you remember about von Neumann’s proof of the Radon-Nikodym theorem?Here’s another: In statistics and many of its applications, including ‘machine learning’, a key tool is hypothesis testing. Then explain the Neyman-Pearson result.For one more, in statistics, why do we care about the law of large numbers? What is the difference between the weak and the strong laws of large numbers?Uh, don’t expect people who didn’t go to college to do well on such things.The HR guy seems to be making a classic mistake, the ‘golden rule’, “He who has the gold makes the rules.”. In particular, since he’s doing the recruiting, he believes that he already knows what qualifications and what kinds of qualifications leading Google workers need for the next several years. No he doesn’t; there’s no way he could. And he wants to say that the Ivy League doesn’t know as well as he does. No way for him to be right there; wrong again.Google stockholders should conclude that this HR guy is putting a concrete cap 20 feet thick on the potential of the Google brain collection and is some of the best news for Yahoo, Facebook, Microsoft, etc.Since I’m not a Google stockholder, I’m not complaining, you understand!

  29. Ciaran

    I haven’t seen your speech so can’t comment on that am worried that something you have said which, I assume, would be well reasoned and argued (even if it wasn’t necessarily something that everyone agreed with) is classified as trolling.Actual trolling is childish, destructive and to be abhorred. Well reasoned argument is something that the world and particularly the tech bubble is in great need of. The misuse of the word trolling is as annoying a trend as the similar misuse of the word hater to describe people who simply happen to have a contrary opinion.

  30. sfrancis

    Maybe I’m looking at this wrong, but I see Apple as not a device centric company, but an experience centric company. Hardware is a key part of the experience, but so is how the hardware and software interact, and the software itself, and the services in the cloud that back it up. But the overall experience is more important than any of these things on their own.So if I’m picking between the experience company and the cloud company, I’d put my money on experience (note: experience and cloud are, obviously, not mutually exclusive).

    1. Dave

      If Apple were experience centric, it would be cloud centric and device agnostic. Many of Apple’s services barely work across their own devices and are regularly glitchy. Some of its key services, such as iTunes, are generally regarded as complete junk to use with a worse user experience than 1980s DOS computers. It is a company that so far has proven an ability to deliver an unmatched experience across its own devices. But little aptitude for delivering cloud services so far. Try accessing your iTunes or iCloud information from another computer. This said as someone who currently owns 12 apple devices for various purposes.On the flip side, I log into my Google accounts and my information is there. Regardless of device.

      1. sfrancis

        You have a difference of opinion about what constitutes experience. Apple’s POV is that experience includes how the object feels in your hand, its size/weight/heft/texture. Google assumes hardware isn’t part of the experience and doesn’t matter. Different world views. Because Apple thinks devices are part of the experience, they’re never going to invest as much in the experience on Someone-Else’s-Device. BTW, my iTunes and iCloud and contacts and other “apple” cloud services work great across my apple devices… even in a browser on someone else’s device (as far as the iworks/icloud stuff goes). Not arguing they don’t have a long way to go but there are signs that they’re sorting it outI can get to my Google services easily as well (and I’m a big google services user as well) but the experience is raw and unpleasant most of the time. Sad but true. That recent update to maps has made it harder to use. Email has been polluted first with buzz, then with G+… but the spam filtering is awesome 🙂

  31. Matt Zagaja

    I think hardware is becoming less commoditized. Chromebook Pixel, MacBook Air, and the new Microsoft Surface are all hardware products that could not be more different. Apple is increasingly fabricating its own parts and components. People used to not care and just use their default earbuds or headphones, then when the iPhone 5 came out Apple redesigned its earbuds to improve sound quality and now it seems like Beats is one of the biggest growing brands. Back in the day you’d just get a computer monitor whatever it is, now people want fancy retina and 4K displays. Nobody used to care about the quality of their cell phone camera, now it’s one of the most important things and where a lot of money and innovation is being pushed.

  32. stevanpopo

    Clayton Christensen talked about a similar fear for Apple last year – specifically mentioning that “their propriety design will soon become commoditized as the architecture of the system opens”. He doesn’t mention the cloud but he certainly questions Apple’s business model, particularly how heavily it relies on hardware.From 8.05 –



  34. davidnotik

    I opined on this subject 4 years ago (…, in which I said:Apple produces a few versions of a great operating system (OSX, iPhone OS) that sit on devices it also makes (iMac, iPhone, iPad). It produces and allows others to produce applications for these operating systems on these devices (Mac-based software, iPhone apps).Google produces applications (Gmail, Calendar, Docs, Search) that work on any device with a modern web browser (i.e. in the cloud). They’re producing operating systems (Android, ChromeOS) that help advance devices to better connect with the cloud. And they’re supporting and further enabling others who produce applications in the cloud.With Apple, you’re locked into Apple hardware, Apple’s operating system, and Apple’s rigid processes. The environment is all Apple.With Google, you’re not locked into any hardware, any one operating system, and you can deliver your services without Google. The environment is open.Apple does walled garden well. That’s because they create superb, forward thinking experiences that are a combination of phenomenal hardware and software that is tightly controlled. They’ve radically simplified and enabled whole experiences because of it. People buy in to pricey Apple hardware and Apple makes money. Hardware leads.Google does open web well. They create superb web based software and experiences that are powerful and connected by virtue of the web. They’ve radically simplified and and enabled whole experiences as well. People choose Google software and Google makes money from ads. Software leads.Google’s betting on the open web. Apple’s betting on, well, Apple.

  35. andyidsinga

    yesss! cloud focus is the winner.not so sure that google will lock up machine learning and intelligence for the long haul. there are some great open source tool boxes in machine learning and statistics ( R, ipython & pandas, weka etc) ..and these are starting to break out of the realm of specialization into more common use by software times!

  36. Chris McCoy

    I see a world where it projects to screens from the cloud. Hardware not needed.

  37. awaldstein

    ;)Personalizing content especially ads in a social context that no one wants is not progress. It’s actually a step back.

  38. Barry Nolan

    “that no one wants” – whilst I’m in agreement with the sentiment, the problem with that statement is that social/demographic/contextual targeted ads convert exceptionally well. Just look at the success of Facebook’s app install ads. Where you see conversion, expect deeper investment.

  39. JimHirshfield

    I see google leveraging my search history on mobile all the time. Whatchyoo talkin’ ’bout Wllis?

  40. falicon

    The car market has been a ‘devices’ market forever…and people still care about the maker/brand there…as the hardware becomes less distinguishable it becomes all about the design, how it gets marketed, the story behind it all that people buy into…

  41. awaldstein

    Data please.

  42. pointsnfigures has had some wins with text ads on mobile.

  43. falicon

    Good points (as always).The “Falicon 500” device will be available on Kickstarter within the coming year…it will automatically program drones, block incoming calls and dump them all to email, play a slew of time-wasting games, help you coach your kids, search the internet, alert you on trending conversations, and even slice bread (if you opt for the handy-dandy, knife add on for just $49.95)…stay tuned! 😉