That’s because we’ve “hit the wall” for single-processor CPU performance, and we’re at the limit for CMOS processes, circuit performance, and instruction level parallelism (ILP).  New processors from Intel and AMD are “spreading sideways”, implementing multiple CPU cores (2, 4 or even 8 processors).  Future processors will have even more cores, and you can “rent” as many additional processors as you need, in the cloud, on the fly.

This is a fundamental change in CPU performance architecture, and it’s forcing software developers to think differently.  For decades, you could speed up your software by just waiting for the next (faster) CPU.  Now, that’s no longer the case.

This leaves us with many large, complicated legacy code bases (e.g. database engines, solid modeling kernels, etc.) that need to be completely redesigned to take full advantage of multiple cores.  That, in turn, will create new opportunities — someone will step in to build multi-core scalable implementations of this stuff.

This time, we bought an Epson (WorkForce 635), and it’s a refreshing difference.  The build quality seems comparable to HP, but the software is much, much more streamlined.

The whole experience underscores something interesting:  the overall usability, quality, and capability of many hardware devices is increasingly defined by software.  Yet many hardware companies fail to prioritize their software design.

Consider this thought experiment:  would you rather have an Android phone running iOS, or an Apple phone running Android?

Instead, I’ve been working on “bits and atoms”, or as I think about it:  the intersection of mechanical design/fabrication with the Internet ecosystem and Moore’s Law advances. The space includes:  CNC, digital fabrication, the “maker” culture, 3D printing, robotics, sensing, and automation.

I think we’re on the cusp (e.g. next 2-5 years) of some major disruption happening in the mechanical & electromechanical worlds, driven by the convergence of three things:

Internet-enabled collaboration.  Growing up in West Virginia, my technology sources were Popular Electronics and Digi-Key mail-order.  It took forever to build anything, and if someone else did an  interesting project, I was lucky to hear about it years after the fact.  Now, I can surf videos of cool CNC shop projects, and easily contact other makers to share project information.

It’s amazing how fast things can advance when information flows quickly and freely, and when it is easy to build on the work of others.  Consider how, in the span of only about a dozen years, we’ve transitioned from an expensive & proprietary software stack (OS, database, dev tools, etc.), to one that’s completely free, open, higher quality, and much more capable.  That only happened because of the collaboration the Internet enables.

Now, the tools & techniques pioneered by the software community are spilling into the mechanical world:  ”open source” designs, version control and configuration management, and collaborative projects.  GrabCAD is one example of a company working in this area. (Disclaimer:  I am an investor).

Moore’s Law advances in computing, storage, and sensors.  The aluminum and screws in a robot arm haven’t changed much in 10 years, but the control computer sure has.  At the top end, a modern Intel/AMD processor has a lot of horsepower for real time image processing, geometry modeling, and control and planning.  At the low end, a Roomba has more CPU capacity than the first desktop computers.  The same advances are happening with storage and sensors:  1TB now costs only $60, and video sensors are getting very cheap.

Many things that were “computationally hard” 10 years ago are now possible.

Modern 3D design tools.   I recently helped a friend rebuild his computing infrastructure after a flood:  $1,200 today buys a very powerful CAD workstation.   Add some parametric 3D design software like SolidWorks, and it’s absolutely amazing what you can prototype, design, test, refine, and stress entirely at your desktop.  When you mix in collaboration:  the ability to take someone else’s model, make your own improvements, and contribute it back, things really start to take off.

I don’t know where it all leads, but it certainly feels like something’s brewing.

Now, Business Insider is reporting that Dropbox turned down $800m from Apple.   I suspect (without data), that Dropbox is much less unprofitable than Groupon, but I’m still shocked.  iCloud isn’t perfect, but Dropbox’s long-term risk is that file sharing becomes an OS & app feature.

Don’t get greedy!

In this context, this means at least a few years experience in the domain, not “we’ve spent the past 3 months studying this market”.  It means you have:

• A clear, “inside baseball” understanding of the markets, customer behaviors, dynamics and existing players
• Personal relationships with key leaders in the market (if you called their cell, would they take your call?)
• An understanding of how the Internet will disrupt the existing market or create the new opportunity

On the last point, there’s a balance:  I sometimes find being in a domain too long can make it hard to think disruptively.

So, how do entrepreneurs avoid the lottery?  There are a number of ways (this is the first in an ad-hoc series of blog posts).

One way is to have a real asset that’s core to the business, but is hard to copy.  Amazon’s product reviews are a great example:  competing with them means competing with the fact that many buyers go there first to check reviews.  Amazon built their own review database, but I think there are many data & content assets that entrepreneurs could buy or license (exclusively) as the basis for a new business.

Sometimes, IP can be the core asset, when there’s real technology with high-quality, issued, in-force patents, and relevant expertise in the team. Given the time it takes to get patents issued, this means the startup is licensing or acquiring patents to get started — pending applications usually aren’t worth that much.

(Note that software/code is rarely a core asset.  There are lots of smart developers, and software is usually pretty easy to copy.)

The next lottery-avoidance topic:  deep domain expertise (to be continued).

… I am concerned that he [Zuckerberg] sets an example of meteoric success that virtually no one else will ever be able to repeat. But wannabes are trying to copy him, and consequently squandering their careers on false hopes.

For pure-software/Internet/mobile ideas, the low barriers to entry have created a very crowded, “weedy” ecosystem.   It’s a great time to be a consumer (e.g. look at the number of mobile apps and free Web sites), but a very tough time for entrepreneurs in these segments.  More and more projects look like lottery tickets:   long odds, with a slim chance of payoff.

Why do entrepreneurs do it?  It’s partially the psychology of poverty, where entrepreneurs hungry for success (financial or otherwise) make irrational decisions. Worse, a startup career commitment, unlike a lottery ticket, has an especially high opportunity cost.   The other element is a belief that skill can influence the outcome.  This is generally true, but is much less so in these segments:  competition is fierce, copying is rampant, and success often comes from quirky combinations of factors that are difficult to plan.

We will have more big Facebook-scale lottery outcomes, and that will spurn more entrepreneurial career bets.  But I think the smartest entrepreneurs will avoid the lottery.

It’s a well-executed simple app offering virtual music “rooms”.  You can just listen, or be one of 5 DJs selecting songs for that room.  It’s a fun way to share music with friends and co-workers.  It’s taking off fast.

I think we’re entering the next chapter for music, following the iTunes era, and it’s all about social.  Apple’s Ping service has the right social buzzwords, but isn’t quite right. Turntable.fm is dead on for one aspect of social music, and Facebook has big plans for music.

(And if Facebook is smart, they acquire Turntable.fm now or just copy the feature. Unfortunately, I don’t think Turntable.fm lasts independently & competitive long-term; it’s so obvious that it should be part of Facebook).

It will be interesting to see how Apple reacts.  Is the iTunes era ending?

It’s amusing how this seems to have come full circle.  Many users started out on AOL, a relatively closed ecosystem.  Then, the open Internet came along and AOL’s relevance faded.  Now, we’re cycling back from open to Facebook’s partially closed ecosystem.  (I do think Facebook will settle into a hybrid of open & closed that AOL never seemed to achieve:  they’ll control the ecosystem “backbone”, while providing open APIs for apps.)

Also, we’re starting to see Facebook opportunities analogous to what’s existed for the Internet.  For example, we have content management systems (CMS) and other tools to manage Web sites, and we’ll see analogous tools (and associated apps) for Facebook page content.   Other things to expect for Facebook:

• Analysis tools for the social graph, analogous to Web site analytics tools
• Tools to manage marketing communications (status updates, new content, direct messaging, etc.), analogous to email marketing / campaign management tools
• Individual user analysis (using social graph data) and profiling, analogous to ad targeting/profiling systems
• Systems to manage communications with individual users, analogous to existing Customer Relationship Management (CRM) tools

In some cases, existing Web tools will evolve to include Facebook-specific functionality. In other cases, the Facebook ecosystem will be different enough that we’ll see new tools emerge (I’m expecting this to happen around the social graph).

For example, consider ideas around DVDs:  they’re still very popular, but their future seems clear.  With bandwidth and storage advances, movies will be like music is today:  downloaded on demand.

These kinds of ideas aren’t necessarily bad, but they require a shrewd assessment of how big the window of opportunity really is.  Trying to build a company is hard enough; doing it while the market is moving in the other direction is virtually impossible.

But at least you should know going in:  is your idea oriented in the “right” direction?  Do you have enough time?