Tuesday, March 3, 2009

Why Microsoft's Future Vision Will Fail

Over the past year, Microsoft has produced a stunning series of “future vision” videos that express where new hardware technologies will take computing.

<a href="http://video.msn.com/?mkt=en-GB&playlist=videoByUuids:uuids:a517b260-bb6b-48b9-87ac-8e2743a28ec5&showPlaylist=true&from=shared" target="_new" title="Future Vision Montage">Video: Future Vision Montage</a>

At first glance, the underlying argument looks really sexy. The 20,000 view of the technology mix looks something like this:

Users will want unified communication between and more contextual interaction with devices.

This will be enabled by:
- Complex sensor networks
- Voice recognition
- Cross-device interactions
- Simple, idiot-proof security
- Cloud computing
- Constant geolocation of users

An Inconvenient Truth
Unfortunately, the amazing slickness factor of these videos (they are produced by Digital Kitchen), covers up an inconvenient truth: Microsoft believes we will be using tomorrow's high speed wireless networks, powerful batteries, sensing and display technologies to do what we do with computers now:

- Authoring, editing and organizing digital content

- Assembling customized solutions
- Obtaining and understanding information
- Communicating
- Meeting
- Working from anywhere
- Scheduling

The idea that businesses will spend on technology that enables workers to do the same basic tasks we do today is an insane argument.

The economics of hardware require a new platform every ~36 months, so we can rely on planned obsolescence on that timeframe. If all of these devices need to work together seamlessly, they will need to be replaced on that cycle or become outdated all at once.

Web 3.0 and Context Aware Computing need to create a MASSIVE increase in worker/material efficiency to jsutify the invesment. More of the same is just not a compelling argument to build the neccesary IT army to manage it. The cost and effort to constantly track and replace sensors, displays, routers and computers will be incredible.

The Scenario Makes 3 Wrong Assumptions
1. Tech manufacturers will play nice with each other
2. High-density/high-band width cloud computing will work.
3. Federated authentication and human tracking are good ideas.

Industrial Collaboration
The Microsoft Future Vision will require a closed hardware/software platform to work seamlessly.

The need for this goes beyond the performance bump of integrated technologies to the basic physics of the situation. IT managers simply won’t be able to cable all of this stuff. It will have to be wireless.

1-10k “Web 3.0” wireless devices in an office work group would create an RF density beyond anything you have experienced and beyond anything the FCC has ever considered approving. The sheer chaos of multiple vendors will scare off anyone who has any IT manager who has been burned by RF issues.

The resulting IP/System standards wars will turn off the few IT execs who aren’t scared of getting even more locked in to MS.

In any case, I’m unsure that systems lock-in is ever in anybody’s interest besides the vendor.

Limitations of Cloud Computing
At current growth curves, power available to datacenters in the US will run into a hard ceiling in 2012-13.

The energy usage required for constant cloud processing and delivery of realtime HD video for 300,000,000 consumers will be through the roof.

The amount of last mile wiring will require massive labor and egregious expense.

The materials used in optical switching limit the maximum speed of network
Early adopters of digital video are choking the system today. 300,000,000 people downloading HD video at the same time would and will kill it..

The materials used in CPUs also have physical limitations.
Multicore computing is an essentially inefficient technology solution. Building 3D circuits with Silicon Germanium may help solve this problem, but this is an unproven technology and the tooling will be through the roof.

Federated Authentication and Human Tracking

Federated Authentication: The idea that one organization manages your security across many networks is a disaster waiting to happen.
When a hacker gets access to one ID, they can break all of the others. BAD. BAD. BAD ... but maybe good for Microsoft, because everyone has a hotmail account.

Human Tracking: Privacy Is Important
Noone wants their boss, employees and colleagues to know exactly where they are all the time, something 50% of men have cheated on their wives. Let me restate that. 50% of executives have or are cheating on their wives. They REALLY, REALLY, REALLY don't want to be trackable.


  1. I think you are taking an unecessarily pessimistic viewpoint.
    I don't buy the argument that these videos just show people doing the same things they're doing today so therefore no one will pay to upgrade to this kind of system. I'm sitting here looking at my new OCS office phone. We just dumped our entire old phone system and went with new phones that integrate with office communicator. I love it. It makes my office life just a bit easier in several ways, but I can't actually do anything fundamentally new.

    You're saying:
    1. Tech manufacturers will not play nice with each other? So I guess all the standards that run our current devices are a fluke? Why is it that my bluetooth headset can pause my mp3 player and switch to a call on my phone automatically, and then when I get into my car it all switches to the car's system? All these devices are from different manufacturers, so you think that level of cooperation won't continue?

    2. High-density/high-band width cloud computing won't work? Just like the broadband revolution didn't work? Were you saying that we'd never get the infrastructure setup for that 10 years ago? Obviously it will take a lot of work, but I see no reason to believe it's impossible.

    3. Federated authentication and human tracking are good ideas. I'm all for privacy, but I'm sure you realize you can already be tracked with your current cell phone. I'm not saying it can't get worse, but I don't see anything in these videos that implies a total loss of privacy.

  2. Hey Staph,

    Great to see you back! First off, the fact that you are working for an MS shop, using a snappy new MS phone (and are you using a Zune?) sort of proves my point. This future only really works when there is one player who holds all the keys.

    1. Tech Manufacturers Will Not Play Nice
    Apple, Microsoft and Sony are all busy building closed systems and focusing on integrating hardware and software engineering again.

    Why won't Zune play nice with iTunes App store? Why won't Surface run Flash? It's not hard or unprofitable to make these things happen. things don't work because the major players are trying to control the system.

    2. Cloud Computing
    I am more concerned about the hard limits of energy capacity in this country than anything else.

    I actually know way more than I should about this. In a previous life, I engineered a core technology that radically improves data center efficiency and my father is probably the world's leading expert in this space.

    Unlike the mysterious Peak Oil (which simply means that energy prices will go up, we will insulate our houses and maybe buy lighter cars) the power grid has an actual limited capacity.

    3. Privacy
    I'm not saying it won't happen. I'm saying that all of the players: Google, Microsoft, Apple, Mozilla, Yahoo and whatever upstart kicks their ass will fight for systems that lock the user into them. This is going to be a long an painful process...and it's a zero sum game. One of them will win.

    I don't think it's a good thing for the winner to own my identity ... or for all of my data, sensors and devices to be monitorable or retrievable with one code or from one device.

    I'd love more of your thoughts.


  3. Well, I think that federalized authentication won't ever occur. It would be realy comfortable but there are a lot of us who simply won't let one entity control all our data. There will be alternate open source solutions and private servers for that matter. Geek friends will buy old, outdated second hand PC-s and store their sensitive stuff there.

    I think as long human beings differ from each other there will be different solutions and some will use this and some will use that. And there will always be enough people who choose something else to have a healthy competition between the players. Flickr or Piacasa? It's a matter of taste really and both are flourishing.

    The energy supply is a much greater concern but I see new winds blowing there as well. I think a lot of the energy will be supplied locally from reusable sources. Small units utilizing the oceans movement will power head stations underwater. Accumulators and new generation solar cells on overground stations. Piezzo christal based generators under freeways and pavements harnessing the vibrations caused by cars and humans walking by. Small turbines on sound reduction walls next to freeways working with the winds caused by passing trucks(The piezzo generators and these turbines are actually being tested and isntalled right now at several sites). These are technologies that are recently developed or already in use. and they are getting cheaper and cheaper. Also new generation personal cpu-s use a lot less power than the old ones.

    The RF density is also a major issue, but as noise reduction and data compression algorithms evolve bandwidth requires less and less RF density. Having low power, high bandwidth base stations and utilizing new generation phones and PDA to route the - well encrypted - traffic of the others not so close to a base station will reduce RF density AND power consumption.

    The thing why MS future vision will fail is simply because they think in terms of closed, proprietary systems. There won't be a winner, there will be competing solutions. Part of the iPhone's success is the fact that virtually anyone can publish new software on that platform. This makes it very flexible and customizable.

    As for people tracking: as long as it's voluntary it's ok. I'd love a solution where they can find me if I'm hit by an avalanche but I could turn it off if I want to be alone. And I don't think lot of people would use anything where they can not turn these kind of features off.

  4. @wths
    Some really good points. But, I have to disagree with the idea that Apple has taught us any lessons about non-proprietary systems, especially with the iPhone. To me, it argues the exact opposite point. Apple's biggest success stories involve tightly coupling their hardware and software and controlling all access points. The iPhone can't load software that's not on Apple's store, and you can't write that software without OSX, which you can't RUN without an Apple manufactured computer. Yet they are a run-away success. Meanwhile, you've been able to run Windows on any hardware you want, and load programs on windows mobile from anywhere you want, since their inception.
    I'm not making any declarations of right or wrong here, but I think most people see Microsoft as the more open of the two companies. Evidence of their willingness to co-exist with competitors runs through nearly every product they make now. From Xbox and Sync talking happily with iPod, to MS Office outputting PDF, to IE8 finally becoming standards compliant.

  5. Apple an open system??? Have you actually read the iphone development and publishing contracts? I agree with Seraph about these points.

    Energy supply -- What's the timeline on scaling any of these technologies in the US? When will they break even. I smell the Nuclear winds a'blowin'

    RF Density -- While I agree with you to a point ... I haven't seen any evidence that the efficiency curve will match the demand curve. It would be great to get some if you have it.

  6. I did not intend to pose apple as an open system, I mereley pointed out that the fact that virtually anyone can create programs for the iPhone did a great part in it's success. A lot of my friends only bought it because the huge collection of apps. But I see now that the way I put it was missleading.

    Speaking of usage terms: they are not kept anyway. The people I've seen work with iphones usually use jailbreaked ones and just copy their stuff with scp on wireless. You CAN run OSX on a hackintosh and you can code for iPhone on a linux or unix system with some hacking around. The fact that you're not supposed to do so is an other thing. Theory and practice are miles apart in these cases.

    The thing is, that the iphone is *practically* a plattform open enough to attract developers who generate enough content that helps making the product itself more popular. The fact that there are patents on multitouch gestures seems to work fine now, since either Nokia nor Palm introduced multitouch gestures in their new phones. Apple has a huge headsart in this market, but as soon as some others unite and create a more open platform, be it a new symbian or android or whatever, where people can choose the phone itself from a larger variety and developers can reach a larger community with their products their success will lessen. They'll of course try to come out with new products again and again, but there will be others as well. They still ride the popularity wave they keep fueling since the introduction of the first iPod. But that also creates a counterculture.

    Windows mobile is not less successful because it's "more open" in my oppinion. It's less successful because you don't have that central store for applications and the flame around those apps. The Xbox live arcade marketplace is the same business model and it is a success. An other thing is, that windows mobile is simply a bad platform. A friend of mine owns a samsung omnia. He's actually relived if his phone is not frozen when he tries to make a call. Sometimes it takes up to 10 secs for the start menu to open...

    But the main point is, that the new business model for applications is what makes the differences. I think it won't be that far, that you'll get cut rates for your internet connection if you donate you're computers idle time. With pcs like the new mac mini that is supposed to consume 13 Watts when idle cloud computing can be made much more effective.

    RF density: I dont have such curves these are more hunches from my side. Right now I got less optimistic myself as I learned that the 60 GHz wifi is coming in order to stream *uncompressed* HD... Such a waste of energy.

    Concerning timelines... Piezzo generators can be installed for every road or pavement with heavy traffic as the'y are renovated which occurs continuously on different roads. It can also be installed on railways, runways and heavily trafficed pedestrian zones, eg.: shoping malls. It produces around 500 kWh energy per kilometer and hour so about 800 kWh per miles and hours. It's return rate is 6-12 years. Which is quite good. And these are first generation devices.

  7. Hi,

    Well thought out piece BUT i strongly disagree. We are already seeing elements of those videos appear in apps we build for clients.

    MS are gambling on this future and they are making the right moves to enable it.

    Tooling (WPF/Silverlight)
    Cloud services (Azure)
    Partners (Intel/all the major hardware vendors)
    Research (Entertainment & Devices devision, Live)

    Who says that this future that MS paints is gonna cost an arm and a leg for businesses. A completely new model may emerge that will fund thise evolution.

    I see this as a viable future indeed!