Velocity Conference talk – DevOps and Intentional Emergence

I'm going to do some more writing on this topic but in the meantime, I've posted the slides from my recent Velocity Conference talk:


The main point of this talk is that the information age is having as large an impact on corporate organization as the industrial age did, and that the organizing principles of that era (e.g. bureaucracy) don't translate well into this one. 

The web grew up in the information age though, and therefore, the way web companies work and organize might make a useful model for the corporate enterprise in transition. The web has been emergent from the beginning, for the corporate enterprise to become emergent faster it's useful to make emergence a goal, to pursue it with intention.

• • •

Maybe it’s time for emergence on purpose

For ten years I owned the domain “” It seemed obvious to me when I bought it that the world had become post-modern and networked but that companies were still behaving like Napoleonic government bureaus. Their industrial age information designs out of synch with our complex and emergent era.

I was doing some work for the U.S. Army at the time and I saw first hand how a top-down, planned, hierarchical, and reductionist bureaucracy was ill-suited to deal with a fast moving, emergent, and networked environment. The common prescription to go through all of those serial bureaucratic processes faster faster faster simply had no chance in hell of working. I proposed that they put tools in the field that would encourage users of all kinds to field their own solutions rather than waiting on the center to provide things for them. I wanted them to intentionally create fertile soil for systems, organizations, and behaviors to emerge at the edge where they were needed. 


• • •

Silicon Valley and traditional corporate enterprises aren't just manifestations of different processes in action, they are different evolutionary branches, based on a completely different set of assumptions. Silicon Valley corporations aren't a different tribe, they are a different species. The only question is, are these two species going to mate and create hybrid offspring, or are some neanderthals about to disappear from the planet?

• • •

I’m not looking for an iPad shaped like a car

Last year at OSCON I heard an automobile manufacturer employee describe their product as a “mobile container for electronics, now it’s an iPad shaped like a car.” More recently I spoke with a chip manufacturer about their vision for compute power in the car – lots and lots of it powering everything from driver assist functions to computer vision and other computational intensive tasks. (Aside: are you ready for your computer to watch your face from a visor mounted camera and decide if you look too sleepy, or perhaps ragey, to keep driving?). Soon there will be enough high-clock-rate CPU and GPU in the trunk to make your car like a mini mobile datacenter, and for it to make a meaningful dent in the car’s energy consumption.

It’s fascinating to think what cars will be when their trunks are full of teraflops. I don’t think self-driving and social-connected infotainment are the only interesting outcomes (I sure as hell hope it’s not just the latter). I’m confident there will be some surprises ahead as engineers and designers have the chance to pull concrete improvements out of these new capabilities (Mini’s predictive shifting [pdf, section 5] hints at one new kind of possibility). I also think the moment that driving becomes an immoral act when the automated version is both safer and more efficient is going to be a difficult one for society (and me) to accept. But that’s the topic for a different post.

What I’m really interested here is the way automobile manufacturers and their suppliers are rushing to fill our cars with digital unnecessaria in an era when 60mpg still qualifies as a huge win. This post is about is about the basics. About cars as means of transportation, and the more efficient the better. I simply can’t understand why most automobile manufacturers seem so disinterested (although VW does seem to pop up a lot on the interesting side of the ledger).

If you are the least bit a gear head and haven’t read Jason Fagone’s Ingenious yet, pick it up. It’s a great read with the takeaway that the 200mpg car is difficult, but achievable. Right. Now. And while Tesla crows about its 89 mpge and .24 Cd, that thing is a fat-faced tank. I mean, it’s definitely a step in the right direction, and I’ll give Musk and crew a ton of credit for getting people’s heads wrapped around the idea of the viable electric car, but the fact remains that that car weighs a porky-assed 4,650lbs and its conventional layout makes sure that it has a big frontal working against its good Cd. 

Bits can only so much after all, in the world of atoms physics is still physics, even if your car has batteries and it’s own Android app. Automated cars will platoon and hyper-mile by default which will helpfully optimize our fat and heavy cars. But what we really need are lighter cars (by an order of magnitude), with small frontal areas and great Cd’s, have better power plants (when can I get my diesel electric hybrid?), and full of tricks like energy recovering suspension. In short, there is a lot of core R+D left to be done before automobile manufactures call efficiency and safety a solved problem and shift all of their focus to where to put all the OLEDs on the dashboard.

Unfortunately, what most of them seem to be saying with word and deed is “Yes, we may have hit peak oil, we may be warming the planet with our exhaust, and we may still be killing 35,000 of your countrymen, woman and children every year, but what we really think you’ll buy is the same old steel-clad beast, but now with a cool display screen on the dashboard and a disembodied digital assistant to read your tweets to you while you disinterestedly drive. And maybe we will, markets being democratic and all, so maybe it’s not them, it’s us.

• • •

A room with a (limited) view

I have always had trouble with focus and attention. When I was a kid they didn’t call it ADHD, they just said I was hyperactive. That term also works, to a point, but it implies a higher state of stimulation while failing to capture the dimension of internal struggle that comes from trying to stay on point. It’s difficult to ignore the squirrels in my peripheral vision. Or eventually, the absence of a squirrel, which sets my mind to looking for one. It’s like a timer in my head that starts at 60 seconds and at 20 makes me jumpy waiting for the next micro-burst of stimulation. 

The advent of the smart phone brought me to a crisis of attention and over the years I’ve tried various strategies to retain a modicum of presence and achieve flow when I have to have it. At one time or another I’ve implemented most of the strategies suggested in @jakek’s piece here. I’m working on a bunch of writing projects right now and too often the computer I’m writing on feels like a bar full of my friends, so his timely post gave me impetus to rebuild a framework for solitude when I need it.

I took his recommendations for my phone and removed email and twitter (those are my primary distractors) and already the twitch to reach for my phone in line at the coffee shop is receding. More interestingly, I decided that the phone wasn’t enough. I needed space for writing too so I re-built a spare computer into an isolation chamber of sorts. A digital equivalent of a room with window, a comfy chair and desk, book shelves and filing cabinets, and little else. I installed dropbox, writing tools (Pages, Scrivener, …), Evernote, and DayOne (I use it for stashing ideas) and nothing else. 

So far it feels like that first month of going paleo with my diet. Hard but rewarding with gradually improving focus. I hope that, like with diet, I’ll be able to build some longterm habits that make concentration and presence easier and more rewarding.

• • •

Emergence works best when it’s planned

In a moment of energetic ambition (or, maybe it was a lapse of judgement) I proposed a talk for the upcoming Velocity conference in New York. DevOps, and operations in general, isn’t my usual haunt, but I felt inspired to try to place DevOps into a bigger picture – an environmental context in which large organizations are changing dramatically with the diffusion of information age technologies.

I’ll say right up front that my thinking on this topic remains inchoate and incomplete. I don’t have a theory to promote. The talk, like this blog, is me thinking out loud.

One of the books that I’ve been re-reading lately is Science, Strategy and War; The strategic theory of John Boyd. Boyd is the originator of the OODA loop, perhaps the only term as haphazardly applied to business as Lean. But what the book is reminding me is that Boyd’s lectures weren’t just discussions of his theories, they were also expositions of his journey to them. He didn’t just want his audience to understand “fast transients” and the OODA loop, he wanted them to understand what he read and what he thought about to arrive there. In a sense Boyd was less a theorist, and more of a curator and synthesizer of an extensive reading list that ranged well outside the expected corpus of military strategic theory.

Thomas Osinga, the book’s author, points out that strategic theorists like Boyd live in a world with Godel-like constraints. Dealing with the complexity of war, society, and politics, their theories are by necessity incomplete. Theories of war are no better than theories of next week’s weather. However, like Eisenhower’s plans, even flawed and incomplete theories of war can still be useful, as is the intellectual development that leads to them. They prepare the mind to react to complex stimuli either intuitively, or consciously but faster.

Theories about business are like this too. Business also operates at an intersection of complex processes, psychology, society and evolving technologies. Which is why, with Boyd as an example, I’m comfortable pointing out well-sources of ideas that seem applicable, attempting some synthesis, and thinking about the connections out loud. I’m not claiming theoretical completeness or anything like it, I’m simply pointing out things that have made me say “huh.” 

The problem is that now that I’m trying to write the talk, I’m stalling out in an endless loop of “If still ambiguous, go more meta.” Before I get so meta that I can’t see that the turtle at the top of the pile way down there, I figured I’d just try to write down a few things for this blog. If reductionism is the pursuit of “turtles all the way down,” my project is rapidly turning into a kind of expansionist “it’s meta all the way up.” I’m hoping that writing a little bit out it will force me to get concrete and slow that process.

In the last few years I’ve done a bunch of talks on “Corporate Evolution” and “Intentional Emergence.” What all of them share is an exploration of the challenges large organizations face as their environment becomes more networked and more complex. Most of them also explore the ways in which the emergence of the web can be applied to more traditional enterprise. Put another way, the information age is forcing bureaucracies of all kinds to transition. Finally. From a kind of feudalism to a post-modern post-bureaucratic form. One that is prepared for the complexity of a networked world, and that takes full advantage of its own networked body.

Corporations have always (or for a long time at least) operated in an environment with emergent properties. Price in a marketplace is an emergent phenomena after all. But what I’m talking about here is a recognition in corporate settings that the reductionist notions of bureaucracy, hierarchy, and planning are inadequate to many modern situations. The degree of connectedness and complexity is simply too much for the lazily looping feedback systems of bureaucracy to cope. Go through the loop faster is a poor (and impossible) prescription when the loop spans such wide swaths of corporate cruft.

As corporations become more internally wired shadow organizations emerge. The official org chart tells less and less of the reality on the ground. The idea of intentional emergence is to recognize this self-adaptation as a positive (rather than treat it as a cancer) and create policies that make emergence of unplanned organizations and organizational behaviors more likely. DevOps, Lean, and other tools of agility have a role here in tightening feedback loops and, by creating small cells of activity, making a fertile organizational substrate for emergent teaming.

Paradoxically, emergence in the corporate setting works best when it’s planned.

• • •

Bringing back Limn This

I’ve been writing at the Radar blog for the last few years and as you can see, have allowed this personal blog to languish. Maybe I’m part of a trend though, because it seems that there is a subtle renaissance building in personal blogging.

I’m reviving mine mostly for control. First, I just want to have control over what I write, and by owning the control, own the constraints. When I write on other people’s properties I absorb the limitations of their brand, or their editorial direction, their voice, or whatever. Self censorship follows close behind – a kind of self restriction that tamps down on creativity by narrowing the field of addressable topics, or one that whispers “You don’t know enough about that to write with any kind of expertise.”

I loved writing for Radar (who wouldn’t want to write for an audience that Tim O’Reilly built?). But with the audience came some expectations that were hard to fulfill. Tim is a great writer. He can go from amorphous idea to clean well-thought-out essay in one draft (and probably one hour). I can’t do that. I’m not a particularly talented writer and it takes me time, coffee, and an internet router with the plug pulled to find the space and clarity necessary to turn out something readable.

But it’s not just a question of focus. Sometimes it was hard to write for Radar because of the unspoken expectation that you were supposed to be like Tim and write some kind of world-changing essay every time you touched the keyboard. I mean, no one put that expectation on me, I did it myself by absorbing his audience’s expectations. 

By the time I started writing on Radar Tim wasn’t posting very frequently, but the things he did post were deep, essay-like, and frame changing. I might have one of those in me a year (if it’s a good year) but I like writing more than one post per year.

Anyway, since it’s hard to write The Definitive Essay, and with Twitter acting as an escape valve for partially formed ideas, the pressure rarely built up enough to get me to push the more complex pieces out the door. In short, I didn’t write as much as I wanted to, because of this unstated need to be making a definitive statement – the frame changing piece that would wake an audience up the next morning viewing the world through a different lens.

The funny thing is that, short of a few important pieces, Tim didn’t really write that way either, or at least not always. The Radar blog when it started out looked more like Twitter. Looking back at his Radar posts from 2006 this one seems pretty typical. Basically Radar back then felt a lot like Twitter with room for a bit more color. Frankly, it makes me wish Twitter was more like Google Plus…, but that’s another story.

So, back to Limn This. When I started this blog back in the way back, I named it that because I saw it as a place to ask questions, to find, not make, explanations. I imagined it would be a place where I posted partially formed ideas and where the feedback would help me make them better. Blogging didn’t always work that way, but I still see this as a place to think out loud and I hope that framing it that way will make it easier for me to post more frequently, to keep the “post” button threshold lower.

I should add that I was somewhat inspired to do this by reading Dan Hon’s newsletter. He’s been writing a daily newsletter to hone his writing skills and to find something to say every day. I’m more comfortable with a pull rather than push model for my own writing, but I like the idea of a space where I can write whatever comes to mind and not worry about whether it fits the remit of someone else’s site.

So, here goes…

• • •