PCs, Smartphones, and the Real Computing Bottlenecks

NPD Group now reports that year-over-year PC unit sales are down 6% for January, with similar declines appearing in prior months. This could have more than something to do with netbook shipments plummeting 34% because, well, netbooks suck. Now we have tablets, which are what netbooks should have been in the first place. Of course, it's not all grim news in the PC space. Mac unit sales are up 20% and Lenovo shows y-o-y growth over 31 percent. But if you're Dell or, God forbid, a local whitebox PC builder, then you're probably sweating buckets in 2011.

Of course, the writing has been on the wall for many years. When notebooks were just passing desktop PCs in unit sales, former Red Hat CEO Bob Young was already commenting to me, "The computers that people are buying now are being used for less and less sophisticated applications." He described eBay and Amazon as the killer apps of our generation, not Microsoft Office. Do you need a desktop or even a notebook PC for these apps? Of course not. You just need a smartphone.

Source: The NPD Group/Evolving Technology Trends: PC Activities on Non-PC Devices

NPD research clearly shows that the "killer apps" of the '90s are now the domain of modern phones. These tasks might be easier on a full-size system, but carriers aren't knocking hundreds of dollars off of new PC sales. In fact, Gartner ranked the devices people intended to buy during 2011 in this order:

1. Smartphone
2. Laptop
3. Desktop PC
4. Mobile handset (non-smartphone)
5. E-reader
6. Tablet

"Continued low retail pricing and widespread adoption of applications like Web browsing, e-mail, Twitter, Facebook, GPS and games will continue to stimulate consumer demand [of smartphones]," said Hugues de la Vergne, principal research analyst at Gartner, in a statement. Smartphone performance for these tasks is now "good enough" for mainstream buyers. Moreover, with speech-to-text capabilities a la Google's speech input, the old argument that phones are terrible for text entry is quickly receding. I can enter text faster by talking to my phone than typing either on-screen or with thumbs, even with the occasional editing to prevent ghastly gaffes.

Many people in the computing industry (especially those in manufacturing) keep telling me that there will always be a need for PCs. These advocates say that applications such as gaming, design, and modeling will keep requiring desktops. I don't buy it. Gaming is about latency and bandwidth while design and modeling are about processing power. Today, these assets still require local power for suitable performance, but all of these factors can be improved enough over time to enable cloud-based application serving. Ultimately, the real bottlenecks will be screen size and input efficiency, not MIPS and throughput, and there's nothing in those two attributes that requires a bulky desktop PC...or even a notebook.

Preview: Ray Kurzweil Optimistic on Averting Catastrophe

With less than one week until the debut of Architects of Tomorrow Volume 1, I thought I'd offer a piece from one of my favorite headliners in this collection. Ray Kurzweil was one of the most brilliant minds of the 20th century, but he sure hasn't slowed down in the 21st. As one of the most public voices in artificial intelligence and a pioneering inventor of many common computing technologies, he contributions to society can't be underestimated. My question, though, is whether the best is yet to come.


One of Kurzweil's favored research fields is human longevity...

WVW: As we see in your books, such as Transcend and The Singularity is Near, you believe we’re close to having the technology needed to radically extend human life. Even if similar technological feats allow us to solve issues such as energy shortages, there’s no getting around the problem of aging societies and overpopulation. Unless these longevity technologies come with a built-in zero population growth switch, how can we avoid this impending risk?

Kurzweil: The idea behind radically extending human life is to stop and reverse aging so we won’t be “aging.” As for overpopulation, the same technologies that will extend longevity will also vastly extend resources. We have 10,000 times more sunlight than we need to meet 100% of our energy needs. The total amount of solar energy we are producing is doubling every two years, and we are only eight doublings away from meeting all of our energy needs. The technology underlying this is the increasing application of advanced material technologies such as nanotech to solar panels. Once we have inexpensive energy, we can easily convert the vast amount of dirty and salinated water we have on the planet to usable water. We are headed towards another agriculture revolution, from horizontal agriculture to vertical agriculture where we grow very high quality food in AI-controlled buildings. These will recycle all nutrients and end the ecological disaster that constitutes contemporary factory farming. This will include hydroponic plants for fruits and vegetables and in vitro cloning of muscle tissue for meat—that is, meat without animals. Even PETA likes this idea.

Desktop nano factories will enable us to produce high quality modules to snap together high quality yet inexpensive housing. There is a Singularity University project devoted to this concept. Look at the cover of last week’s Economist magazine, it features a picture of a violin that was printed out using a contemporary three-dimensional printer. This spatial precision of this technology is improving exponentially. And there’s plenty of land available, just take a train trip anywhere in the world and you’ll see most of the land is not really used. We aggregate today in crowded cities because this was a good way to communicate, but the need to do this will dissipate as virtual reality becomes equal to real reality in resolution and realism.

Whither Intel? The Master of Mobile Answers

While composing my prior blog post on Intel's suckerpunch from Nokia and Microsoft, I was also waiting for a reply from David "Dadi" Perlmutter, Executive Vice President and General Manager, Intel Architecture Group, for our Architects of Tomorrow follow-up discussion. Tonight, I received Dadi's replies. Regarding the question of what happened to Intel in the ultramobile space, you might find this part of our dialog interesting:

Me: We previously discussed Atom and handhelds at a time well before Moorestown. Then I attended the Texas Moorestown briefing and was under the impression that Atom storming the ultramobile market was imminent—but the storm has yet to materialize. Why?

1. Perlmutter: The handheld market requires more integration and testing that traditional PC markets before a product can be released.  This is related to the integration of all the components on the device itself (namely apps processor SOC, cellular modem, power management and the whole SW stack) as well as testing of the device within the carrier networks.  Having great performing silicon is just the first step in a long process.  We are currently working with several partners to integrate our silicon into their devices and the carrier networks.  Although this integration work requires a steep learning curve, the good news is that the learning is transferrable to our next generation.  So, even though our Medfield silicon is just sampling now, we should have a phone shipping with Medfield this year (2011).

Red Hat's Founder on the Secret of Being a Successful Programmer

Bob Young founded Red Hat Software, built it up into a credible threat to Microsoft during the dot-com boom, and in 2002 left the company to go shake things up in another sector: book publishing. He started Lulu Press, which is now one of the world's top self-publishing services.

My discussions with Bob Young will be in the upcoming Architects of Tomorrow, Volume 1. He's a fascinating, clever entrepreneur with loads of experience and wisdom to share. To give you a taste, at one point I asked him what his first experience was with a computer, and what I got was a lesson in work habits.

Young: While I define myself as a salesman, my interest in computers goes back to my second year at the University of Toronto in 1973. I was taking a computer programming course, and the practical component of our course involved sitting at the keyboard of a big desk-like machine that punched what we typed onto 8" by 3" punch cards. Of course, the University’s computer department was seriously underfunded (some things never change), so there was a real shortage of these punch card machines, which resulted in long line-ups all day.

The most frustrating part of this course was that no matter how hard I worked, there was this group of kids who scored grades that were dramatically higher than the rest of us. Having spent most of a semester with them, I was convinced that they were not genetically superior, so I was curious. One day, standing in line at the punch card machines behind one of the smart kids in my class, I started whining about what a waste of time it was. Instead of sympathy, he sarcastically mumbled, “Well, then why don't you show up after midnight like the rest of us?”

Later that day, I did, and I learned one of the important lessons about computing I’ve never forgotten: real programmers don't sleep…which, I suppose, is why I never became a real programmer.

Thanks for You: Free Limited Time Ebook Download!

As a special thanks to you Behind the Lines readers, I'd like to offer you a free download coupon for my ebook, Strange Horizons Retrospective, Volume 1. Simply go to this Smashwords link, enter the "100% off" coupon code MG42K, and download the ebook in the format of your choice.

Feel free to pass around this coupon to anyone you like. The coupon will expire on February 21st -- the ebook is yours to keep. If you'd like to return the favor, a review on the ebook's Smashwords page would be most appreciated.

Best wishes, and again -- thanks so much for reading!

Nokia, Microsoft, and the Suckerpunch to Intel

Back in May 2010, I covered Intel's Moorestown launch for Tom's Hardware. At the time, Intel was dead serious about taking on the ultramobile handset and tablet field with its new Atom platform. At last, Intel was on par with its ultramobile CPU archenemies, ARM and Qualcomm, and it had the multimedia horsepower to blow its rivals out of the water. Moreover, as good as Moorestown looked on paper (and performed in the lab, as I witnessed first-hand) the follow-up platform, Medfield, would be far better.

So here we are in early 2011. Moorestown phones were supposed to arrive in force in the second half of 2010 -- and didn't. Intel and Nokia joined hands to work on MeeGo, a Linux-based successor to Intel's Moblin project and Nokia's Maemo. MeeGo represented a new shot for the cell phone heavyweight to shore up its flagging market share. Intel even launched a MeeGo app store.

Here in February, we've had two critical bits of news. The one you probably heard about on February 11th was the new partnership between Nokia and Microsoft, wherein "Nokia would adopt Windows Phone as its principal smartphone strategy." This is a huge win for Microsoft, which hasn't been able to get the rush of buying interest for its new phone OS that it justly deserves. It's probably also a good thing for Nokia, which was obviously going nowhere fast with Symbian.

The second piece of news you may not have heard is that Nokia is essentially signing off on its MeeGo efforts. As per Nokia's February 11th press release, "MeeGo [now] becomes an open-source, mobile operating system project." That means that the budget for MeeGo will be whacked, and either Intel will have to pick up the entire tab (which would be stupid without a major phone partner) or the Linux community can poke at it for practice and giggles. Nokia promises that one MeeGo phone will arrive, but what sane developer would possibly commit resources to only one phone? Anyone who thinks the Nokia N901 will be the next iPhone is delusional.

So the doubled-over loser in this arrangement is Intel. Nokia wants/needs Windows Phone, and Qualcomm has a virtual lock on today's Windows Phone devices. Worse yet, when I was at that Moorestown meeting, Intel engineers clearly were not excited about the Windows Phone platform, ostensibly for performance reasons.

All of this adds up to very grim and likely unexpected tidings for Intel, which desperately needs at least a beachhead in the ultramobile market. Intel indicated last year in no uncertain terms that a large part of personal computing's future lies in handheld devices. If that's true -- and all signs indicate that this is the trend now afoot -- then the Nokia/Microsoft affair poses a significant challenge for Intel. If Medfield can't get the company into significant numbers of handsets and start reversing the ARM and Qualcomm tide, what possibly could?

Schlage Home Automation: How Cool/Hot is That?!

I just finished a piece on IP-based consumer home security products for Tom's Guide. In that article, I reviewed four basic products: a pair of D-Link IP cameras, a pair of (presumably higher-grade) Axis IP cameras, Logitech's new Alert powerline camera rig, and Schlage's LiNK home security kit. I went in thinking it all might add up to roughly 4,000 words. Turns out I was wrong...by about 4,000 words.

And would you believe that even at that egregious, distended length, I still cut material out of the story? One omitted element was the installation of the Trane thermostat that Schlage sent. Home HVAC control is a great component of the LiNK platform, but it didn't really pertain to security. (Unless you're trying to overheat your thieves to make them sluggish and so buy more time for police to arrive. No, that wasn't in the product literature.) Still, half of the articles purpose was to convey that home security setup isn't rocket science. Anyone can do it, and the same is true of home automation. Many people, including me until recently, have been leery of replacing their home's thermostats with something smarter and more integrated into their home networks. Since I expect we'll be seeing a lot more products like this in the future, I thought I could post the thermostat installation process here and show you a bit of what it can do. Check it out.

Here's the home's original thermostat. It's just your typical, low-end digital unit, with temperature scheduling that's about as easy to program as a 1987 VCR.

You can see how the old thermometer is held on with two clips along the top. With a push, a press, and a pull, the old unit neatly comes away.

If you love your wires, set them free. Then free the back wall panel with a little wrench.

Here we are, stripped down with four wires ready for action.

Yes, I had to make a couple more holes in the sheetrock, but that was far more preferable than seeing the white undercoat poking out from the new thermostat's left edge.

Note Trane's excellent labeling and the supplied sticker tabs to help correctly color code the wires.

The old thermostat only needed four wires. The Trane needs all five, so I had to uncoil the blue wire and strip off some of the end insulation.

Screw each wire into its marked place. The G (green) wire gets screwed down under the G screw, and so on.

This is what the Trane back panel looks like with all wires secured in place.

Snap on the Trane thermostat and -- that's it!

The user manual guides you through setup and whatnot. I particularly like the thermostat's graphing mode that shows daily heating hours over the past week.







After installation, you pair the Trane thermostat with the Schlage LiNK system's wireless Bridge controller. Both devices use Z-Wave wireless technology, and you simply pair them, much like a cell phone with a Bluetooth headset. With that done, the thermostat now appear in Schlage's LiNK interface, accessible through any Web browser. What you see looks a lot like this:

Slick, right? This means that you can control your home environment from any Web browser, including the one in your Android device or iPhone. If you don't feel like programming your thermostat the old fashioned way, try this:

With special adapter switches, you can have similar scheduling control over lights and appliances. Soon, I'll post some interesting material from Intel's Lorie Wigle and her thoughts about energy consumption and how technology will soon reshape how we manage our home electronics. For now, just know that "the future" of home automation is finally arriving, and it's both easier and more affordable than many might think.

Note: Special thanks to Joe Jones for his help with the Schlage installation and photography for this posting.

From Left Brain to Post-Human

Synchronicity can be fun. It can also scare the hell out of you.

I've been listening to Daniel Pink's A Whole New Mind, and I'm at the chapter in which he looks at how developing nations, such as China and India, are rapidly catching up (or have already caught up) to the U.S. in terms of being able to handle left brain-oriented work tasks, meaning those which rely on consistent processes and linear thinking. An increasing number of large companies are outsourcing their accounting work, for example, hiring equally qualified Indian labor for less than 20% of the cost of American equivalents. 

I've been thinking a lot about such trends recently. I've been pushing my children (now ages six and nearly nine) hard to master math skills on Khan Academy and introducing them to broad science concepts. Some people think you're a tyrant just for pushing your kids to be a year ahead. I want my boys four, five, even six years ahead. Why? Because the rest of the world is hungrier and more motivated than we are. If kids abroad can be trained at this pace, they will be -- and they are. Those kids will emerge from school competing against my kids for all of those left brain jobs, and those other kids will be able to apply their accelerated educations for a fraction of the wage my kids will expect. In short, if my kids do nothing more than "meet benchmarks," if they even manage to keep pace with the kids in countries currently trouncing America in secondary education, they will still lose. There is no way my sons will be able to compete for geographically unrestricted work at $2 an hour, especially after racking up a $100,000+ college debt. The implications for American unemployment and our general economy are grim.

As if the outsourcing situation wasn't bad enough, it gets worse. Pink continues his argument, noting that software developers are creating applications able to replace even cheap foreign knowledge workers. It's the inevitable result of Moore's law. When computing becomes so powerful and inexpensive that TurboTax can effectively replace a mid-level CPA, not even Indians can compete. Pink points to Appligenics and the fact that we now have software "smart" enough to create new software. Computers are the ultimate left brain endgame. How long will it be before computers can perform contextual analysis and automatically generate the press releases some companies pay me hundreds of dollars to write?

Enter synchronicity and my in-progress interview with eminent inventor and futurist Ray Kurzweil. I interviewed Kurzweil in 2002 for my very first Q&A column in CPU magazine. Today, I'm doing a follow-up to this piece for my Architects of Tomorrow book. Kurzweil currently has a just-released movie called The Singularity is Near, based on his 2005 book of the same name. "Singularity" is a coalescing of accelerating technology trends. As Kurzweil's site puts it, Singularity is "an era in which our intelligence will become increasingly nonbiological and trillions of times more powerful than it is today—the dawning of a new civilization that will enable us to transcend our biological limitations and amplify our creativity." He's talking about the blending of humans and computers -- the chip in your head or hand or wherever it becomes advantageous. Because isn't that what it will boil down to: advantage? Some people will get such implants for enjoyment. I suspect most will get them out of economic necessity. Today, my kids need to access more knowledge in order to stay competitively relevant. My grandkids will face the same challenge on a higher order of magnitude.

When it comes to forecasting technology trends, Kurzweil has a rather amazing track record, and he has the resume necessary to avoid being taken out of hand as a quack. He believes, and the trend undeniably indicates, that computers are only a few years away from being more intelligent than humans. Both he and Pink point to the defeat of chess grandmaster Garry Kasparov by IBM's Deep Blue as writing on the wall. Kurzweil's chief area of research for the past decade has been artificial intelligence (AI), and if anyone is in a position of authority and knowledge to make such grand statements, it's him.

Kurzweil dismisses the oft-discussed risk of computers rising up against and overthrowing their human masters, as in The Terminator series, and I suspect he may be right. After all, if we bind biology and computing together, doesn't the technology become a sort of integrated prosthetic? It's not us against the machines. It's just a new, different us. And if we're bending genes in the same way we're building circuits, then it's not really so much about prosthetics as it is about changing the human form. We are racing down the path, well along the way toward turning ourselves into something else. It's pointless to pass judgment on whether that something else is better or worse. It's just different and more complex. 

I don't want to agree with Kurzweil. I find the idea repugnant. But it all makes terrible, logical, and perhaps inevitable sense. If Pink is right, then we're engaged in a left brain arms race that seems bound to lead us to Kurzweil's Singularity. The remaining question is whether computers will be able to master creativity and emotion, as well, thus migrating into the "right brain" half of our humanity. And if Kurzweil's time tables are right, if we only have 20 to 40 years left for these technology tracks to play out, what does this mean for my children and grandchildren? What can I do except continue to drive them toward a successful tomorrow...and pray?