Skip to content

Facebook, Instagram and the Singularity

2012 April 18

Everybody was surprised by Facebook’s acquisition of Instagram, apparently, even Facebook.  Merely one week before Instagram closed a round of funding that valued them at $500 billion (and diluted their shares), so it’s clear that they were not in serious talks by that point.

Why would Mark Zuckerberg drop $1 billion on a company less than two years old with no revenues to speak of and no apparent business model?
 
There has been an avalanche of pundits giving all sorts of reasons why the deal makes perfect sense:  The user base, a foothold in mobile, their technology, development team, etc, yet I haven’t found any of those explanations comelling. I think they bought Instagram because they were scared of the singularity and with good cause.

Only the Paranoid Acquire?

It’s long been known that big acquisitions rarely pan out.  Companies usually pay too much and get too little.  More often than not, acquisitive CEO’s are driven by ego or just sheer boredom (adrenaline filled sessions with high-priced investment bankers are a great antidote for both).

Usually, there is some strategic rationale and talk of synergies gained from eliminating costs, but the fact remains that great companies are not built through acquisitions and cost savings usually fail to materialize.  More often than not, the costs of aligning two disparate cultures negate any savings on overhead.

However, as I pointed out in an earlier post, tech companies have an unusually good record with acquisitions (for instance, they had very small write-downs during the financial crises) and I think a lot of it has to do with the S curve.
 

 
While companies in the regular economy focus on established firms, tech companies tend to buy small firms in emerging areas.  So established companies at the end of their S-curves where returns are diminishing buy young upstarts in the breakthrough stage.  

There are many S-curves going on at once so large tech companies make dozens of acquisitions every year, most of which are incorporated into existing products. Siri, for instance, was not developed by Apple, but was acquired. Other times, small stakes are taken as a hedge (Microsoft, for instance, was an early Facebook investor).

What’s unusual about Facebook’s purchase of Instagram is that Facebook not only made a large investment in a brand they intend to keep separate, but they themselves are only eight years old. So the process has been greatly compressed and I think that is the most interesting aspect of the transaction.  In fact, it may be a harbinger of things to come.

The Exponential Digital Laws

I’ve written before about digital laws, the most famous of which is Moore’s law which states that processing efficiency doubles every 18 months.  However, exponential laws are the norm rather than the exception in technology.  Here’s one that Technology Review published recently, which goes back to the earliest days of computing:
 

 
Although the chart specifically shows the trend for computations per kilowatt hour, it’s a fairly good surrogate for what’s going on in technology overall.  The advancement over the last decade was roughly 1 billion times that of the 80’s.  The next decade will go 1000 times faster than the one we just passed. The rate of acceleration is mind-boggling.

Put another way, if you’re an executive who cut his teeth in the Reagan years, your thinking is a trillion times too slow.  If you’re one of those hot-shot digital natives who tweets more than speaks, you’re experience is still in an environment that moved 1000 times slower than the next decade will.

The Singularity is Near

That brings us to Ray Kurzweil’s concept of the singularity, where advancement becomes so fast that it’s almost instantaneous (the concept gets its name from the point of infinite gravity at the center of a black hole).

A lot of Kurzweil’s ideas are pretty far out and many of the predictions he made in his 2005 book, The Singularity Is Near, have proved too optimistic.  However, you don’t need to buy his entire theory to see that the basic narrative is undeniable.  Things are getting much, much faster.  Take a look at this chart:
 

 
While the trend isn’t 100% chronological, it’s clear that the S-curve is getting increasingly steeper.  Instagram grew to 30 million people in less than two years.  If you assume that decision cycles of large companies (and agile ones at that) run about 6 months, then management as generally practiced has a big problem.

What happens when development cycles become shorter than decision cycles?

The New Logic

I think it’s clear that Zuckerberg bought Instagram out of fear.  Historically, that has not been a generally accepted way to make decisions.  Nevertheless, times are changing and it’s easy to see how spending 1% of your total market value to stave off an existential threat has a certain logic.

However, it is a new kind of logic and certainly not the kind that they’ve traditionally taught in business schools or in statistics courses.  The calculated risk of Instagram was reasonably small, but the potential risk was enormous and possibly life threatening for Facebook.

We no longer live in a world of Gaussian statistics, but a broken order of uncertainty and emergence in which inaction is every bit as dangerous as wrong action and traditional decision cycles are no longer viable.

– Greg

14 Responses leave one →
  1. April 18, 2012

    Very interesting! Something to ponder; implications can be more than we can even comprehend now..

  2. April 18, 2012

    Yeah, it’s pretty far out. Kurzweil’s book is definitely worth a read.

    – Greg

  3. April 18, 2012

    Greg,

    Great post.

    I’ve seen you develop your ideas for this post over a while – through “apocalyptic” web squared type rates of change, the technium and finally the singularity.

    There are two views on the singularity.

    One is the AI view – where “Skynet” becomes intelligent and AI “takes over”
    One is where the rate of change curve goes vertical.

    Its quite possible both will become reality but like you I am more interested in the rate of change view.

    I like O’Reillys concept of Web powers – Web 1, Web 2, Web squared – I think that Web Cubed may describe the singularity – possibly the combined AI-Rate of change singularity.

    “Predictions put the singularity at around 2030 but if exponential change itself becomes exponential (cubed) then it may come at us very much faster – I have a feeling that the 2020s will be very interesting indeed.

  4. April 18, 2012

    Martin,

    Yes, I’ve written about 2020 before (https://digitaltonto.com/2011/the-digital-world-in-2020/), we’re certainly getting to the end of a few key paradigms. I agree that you don’t need to believe in the Skynet scenario to take the singularity seriously.

    However, I don’t think the Skynet scenario can be dismissed entirely. The one thing that the apocalyptic vision of Termimator misses is that Skynet is one uniform intelligence. It’s unreasonable to think that if we have intelligent computers that they would all think the same way.

    – Greg

  5. April 18, 2012

    Greg,

    The question about AI and multiple intelligences opens up a singularity in discussion itself – a massive, massive topic.

    In the AI scenario I think that not only wil there be multiple “intelligences” but that they will indeed be Artificial – no reason why new AI should b anything like human intelligence.

    One reason I think that the rate of change and the AI scenarios will connect is that it may only be AI that can deal with the rate of change.

  6. April 18, 2012

    Yes, that’s true. Which is why augmented human intelligence is likely as well. Weird stuff.

    – Greg

  7. April 18, 2012

    Not to take away from your interesting analysis, but could it simply be that the trend has exploded for consumption of visual content, and in the scheme of Facebook’s coffers, it was a minimal investment for a popular service?

    We are in the middle of a visual content revolution as information competes for our attention. http://m.bbc.co.uk/news/business-17682294 Something I mentioned in November 2009, slide 36: http://www.slideshare.net/ehcamw/best-strategic-learning-investment-in-2010-x.

    If I were Zuck, I would also invest in the future of emotion measurement: http://www.scoop.it/t/angela-dunn-foresight/p/1616185534/the-science-of-smiling-the-future-of-emotion-measurement-technology-ubiquitous.

  8. April 18, 2012

    I’n sure that figues in.

    Thanks Angela.

    Greg

  9. Emily Knab permalink
    April 19, 2012

    Love this take – very interesting perspective! I think you meant Instagram in your title, though. 🙂

    Thank you for sharing!

  10. April 19, 2012

    Ooops! Fixed it.

    Well, anyway, I’m sure it will be Pinterest soon LOL:-)

    – Greg

  11. April 25, 2012

    Interesting Analysis. To think through your question, What happens when development cycle becomes shorter than decision cycles, I think, we are thinking again through the old mind-set. This instantaneous change requires us to rethink the very idea of decision cycles. Decision cycles comes from the military mindset where you are trying to orient your activities based on the movements of your competitors( prospective ones, in the case of an ecosystem). This mindset, as you pointed out, leads to paranoid behavior, as evident from this case. However, if firms could enhance their awareness, and look at each each decision as a starting node, whose steps would be driven by instantaneous feedback, I think it would make decision process synchronous with the development process, rather than being reactive, trying to catch up with the other.

  12. April 25, 2012

    Very good points Venky. I think you’re right that the decision process as currently understood is seriously problematic. However, businesses do need to be run day-to-day, so it’s not as if you can just turn a switch.

    Mintzberg’s concept of emergent strategy is interesting and seems to be able to work in some cases (i.e. Intel moving out of memory chips), but there are quite a few kinks to be worked out.

    – Greg

  13. May 1, 2012

    Love your take on this… I’ve been trying to figure out why the hell that acquisition had that price tag on it and the only reason I ended up with that made sense is facebook’s fear of the momentum that instagram has and the loss of picture momentum on facebook mobile… It would be interesting to visualize that if the data is available, so did instagram steal from facebook’s mobile image upload traffic?!
    And I am quoting your last paragraph “We no longer live in a world of Gaussian statistics, but a broken order of uncertainty and emergence in which inaction is every bit as dangerous as wrong action and traditional decision cycles are no longer viable.” as a possible trend in 2012 🙂
    Cheers,
    J

  14. May 1, 2012

    Thanks Jad.

    I don’t think Instagram stole Facebook’s traffic so much as they simply managed to gain momentum (there are several fairly equivalent services). More than anything else, it was probably worth 1% of Facebook to just not have to worry about it and move on.

    – Greg

Leave a Reply

Note: You can use basic XHTML in your comments. Your email address will never be published.

Subscribe to this comment feed via RSS