In grad school, I remember a conversation across the campus green with an visiting psychologist from Harvard. I don’t remember much about the conversation except that he introduced me to Isaiah Berlin’s notion of the Hedgehog and the Fox, and correctly pegged me as a Fox. I think I was a bit offended at the simplification, but time has proven him right. I’m certainly no hedgehog.
I got into a silly argument on twitter last night, about whether my looking to hire someone who I labeled (as job descriptions make us do) a “Coding Designer” was not just foolish (I’d seen the Unicorn references in my tweetstream already) but apparently a bad idea, because, so the ultra-simplified argument goes, you somehow can’t be both. And so I’ll use the energy to rant a bit about what seem to be prevailing attitudes around titleism and narrow definitions of “professionalism”.
We all need to define ourselves to others. It helps us be understood, and hopefully valued. Labels can be useful for that. We also, even more, like to label others. It helps us simplify our approach to them. If I can find a label for you, then I can rely on a prioris about how people with that label tend to think and behave, and I don’t need to actually get to know you too much. The more people we interact with, the more important these shortcuts are. Some roles are particularly subject to that (Recruiters, VCs, politicians, etc. — people who routinely talk to dozens if not hundreds of people a day). And the best at these roles are those who use a different labeling system than their peers. Recruiters who see the latent ambition or genius in a shy candidate; VCs who see the determination behind a stutter, or, conversely, the lack of self-confidence behind the bravado, etc.
Labels are useful and practical in the short term. And I don’t know how one could run a large HR department without them. But we should be careful to not take them too seriously, as in the long term, they can hurt. They hurt because people, especially interesting, worth-getting-to-know people, are much more subtle, complicated, confusing and hard to categorize creatures. Whether you take the label too seriously when thinking about others (e.g., refuse to see the valid opinion about a design expressed by a non-Designer) or about yourself (and limit your impact on the world because “oh, that’s not something that a mere ____ like me could say/do”), you’re not getting the most out of anyone involved.
As I write this, I realize that I feel quite strongly about this topic. Part of it is probably because I grew up in an educational system, which at least then believed way too much in labeling people and determining their fate based on that label. Much waste ensued. Part of it is probably because I can’t for the life of me figure out what my label should be, and if I can’t, then that must be bad. I’ve had a range of professional labels, from scientist to engineer, architect, team lead, vice president, CTO, CEO, blah blah blah. I’ve been called a designer, strategist, entrepreneur, boss, blah blah blah. None of those words will, I hope, be in my epitaph. And so I get cranky on twitter at night, because if there are people who strive to be both excellent at design and at coding, then by golly we should encourage them.
Titles are a poor approximation of a professional ideal, and a profession is a poor approximation of a human’s breadth, contributions, and talents. Embrace your inner fox, and if you happen to have both design and coding skills, can see a problem, conjure up a solution, prototype it, welcome challenges to your idea from peers, data, and users, apply.
I’ve been having a really stimulating few weeks, which is making me feel oddly optimistic, so I figured I’d spread the cheer.
It’s been somewhat quieter than usual in my little corner of Mozilla, as many of my colleagues were busy preparing for on our big announcements and demos at Mobile World Congress. This lull has allowed me to spend more time indulging in longer term thinking.
In that thinking period, I’ve had the privilege and distinct pleasure of spending a few hours of quality videochat with none other than Ward Cunningham, inventor extraordinaire. We’ve been talking a lot about his new baby, which he calls Federated Wiki (the name is misleading I think, but as with many new ideas, awkward names are better than long descriptions. The shortest analogy I can come up with is a space for personal writing that borrows the forking attitude from github, within a smalltalk-like conceptual model. — Confused yet?). Regardless of the name, it’s both refreshing and stimulating to talk to someone like Ward about topics we both see as important and interesting, with a shared commitment to improve the status quo. If you’re intellectually curious, willing to do some mental backflips and deal with fairly raw and unconventional ideas, I encourage you to watch the videos and play around with federated wiki. It’s a worthwhile exploration.
One of the surprising ways in which that exploration is changing my brain is that I’m finally starting to care about hardware. Until recently, I’ve been quite uninterested in hardware talk, both of the standard computer geek variety, as well as the internet-of-things variety. The former has felt no different to me than talking about fashion, and the latter always struck me as an industrial problem, not a human one. It’s made sense to me that companies like GE were keen on the Internet of Things — i just didn’t really care, and the odds of billions of networked computers being a net win for the world seemed pretty slim — I’ll admit to a bit of an head-in-the-sand reaction there.
What’s changed my thinking (and let me to acquire handfuls of raspberry pis and pcDuinos for the office) is a growing belief that the economics of hardware manufacturing, the maturation of the open source stack and standards, and the mainstream adoption of “the cloud” are, in concert, going to provide interesting opportunities for networked computation “by the people for the people”.
The default path
It’s fairly easy to see that the big players in computing (Apple, Google, et al) are deepening their stack and broadening their scope — from software or services to software and services to hardware to phones to tablets to TVs and other consumer devices, to cars, eyewear, physical networks, wireless carriers. The path ahead for them is clear. And I’d say that the consumer experiences that will follow promise to be stunning. But if nothing is done about it, it’s also clear that this will happen through market concentration. That concentration of power, economic influence and wealth is particularly problematic because it also implies a concentration in modes of thought, in cultural outlooks, in possible futures. As I’ve written before, I think the world would be poorer if that happened.
All conventional indicators (especially those concerned with things like “eyeballs”) are in that direction, of course — large amounts of money make it possible to make shiny attractive things, as well as to identify attention-grabbing things in a sea of uneven quality. Popular culture naturally gravitates towards superstars, be they pop stars, or youtube clips. We seem to be wired as people, societies, and markets to head that way.
Another architecture, and the economics that can power it
And yet I mentioned above that I was optimistic. That’s because while there are natural economic incentives for markets to create large companies, and for those large companies to concentrate attention on a few things which are safe, non-threatening, and profitable, the centralization forces are not the only ones at play, and not everything behaves in market-expected ways. Clay Shirky has written and talked about this eloquently — sometimes we don’t need to do what economists tells us we should want to do.
You can now buy $35 full-featured computers which, within seconds of being plugged in, are fully capable internet nodes. Within years at most the price will keep dropping and the capabilities will keep expanding. In addition, the open source software stack which powers them is the same which millions of web developers already know, and hopefully soon millions of amateurs as well. In other words, the means of production are being commoditized and democratized, and the network is ready to serve as a coordination layer for these devices. I am not predicting a utopian world of DIY hardware running bespoke software written by each user. Instead, I want to point out that the entire value chain for using the internet to do new things (from hardware to services) is now ripe for new, good ideas, even coming from single brains. The substrate for deeply personal web experiences is here. I should also acknowledge Allen Wirfs-Brock’s recent writings on the need for a bigger picture plan here as helping me to process the need to actively shape the architecture of the connected things around us.
Much has been written about the impact of Big Data for business, but internet-available Personal Data could be just as transformative for individuals and small groups. I’d wager that a Raspberry Pi and a cheap drive can store all of the words I will write and read in a lifetime, do meaningful computation on that corpus, and be available to all of my internet-connected devices. That feels significant. When people routinely have that kind of agency over their stuff, and if their digital agents are well coordinated through a well-decentralized network, it’s easy to imagine just as futuristic and impactful but less commercially driven outcomes.
When compelling experiences that work thanks to a decentralized architecture emerge, I’m confident that many kinds of business models will follow, including consumer devices, services, and others. Some of those devices will run FirefoxOS. All will have the web stack built-in. As important, these things (or virtual things) will be cheap enough that cost and profitability won’t be the primary consideration. Instead, you’ll think about what you, your family, your neighborhood, your community need. I could easily subsidize a fairly powerful computing service for all the children in my kids’ school. Larger foundations could subsidize the same on the scale of countries. Combining Shirky’s cognitive surplus and insanely cheap delivery systems means we can be freer from the stranglehold of short term economic thinking. Bring in network-enabled funding models like Kickstarter, and we can keep baffling economists and conventional investors for a long while.
The Product Challenge
The conundrum, then, is to find those compelling experiences which thrive on a decentralized architecture, and not just try to force decentralization on experiences which work well in a centralized model. As a “product person”, I’ll admit that it’s not easy to wean oneself from the dominant world view, which favors neatly bounded and hard, polished exteriors (metaphorical and otherwise), where control is critical to quality, and where complexity is managed through barriers. That’s why I’m so impressed with Ward’s thinking — he’s been starting with a decentralized architecture, and seeking to identify and facilitate outcomes that are easier, more natural in that world. We need more such thinking.
When I start thinking that way, my mind wanders towards deeply personal systems and deeply meaningful relationships. Private diaries. Family-centric stuff (family recipes, private calendars, snapshots, secrets). It is worth noting that Ward has seen his ideas resonate in tightly knit communities who need deep connections over ideas and data, but who aren’t seeking mass audiences. More generally, I suspect that this kind of architecture will naturally complement offerings from giant players by gravitating towards parts of our psychology which the giants aren’t built to serve. Our need for quiet; play space; safe space. Our need for a small number of deeply meaningful relationships. As the giants define what the “mall” style of interaction with the net is, I’m (finally) optimistic that we’ll see structures emerge that mirror many other social structures, from the self to the couple, family, school, workplace, church, interest club, support group, etc.
This future feels more humane. That’s worth working on.
I’ve been considering switching hosting providers for a little while, as my current provider, while cheap, has been slower and slower, making me think that they’re not managing their hosts super well. And I figured it’d be good to get a feel for how that part of the market has evolved.
A tweet and some link-following later, I’ve come to a few tentative conclusions:
First, like everything else in the computer world, the capacities have gone up, but the price points haven’t changed drastically in the years since I setup ascher.ca. The dirt-cheap providers are a few dollars cheaper than before (e.g. Laughing Squid), but most low-end VPS offerings are $15-20/month — they’re just bigger defaults than before.
Second, there’s clearly interesting competition to reach out to l33t developers. Dreamhost, Media Temple (aka (mt)), Slicehost, etc. have been around for a while, and they seem to be doing fine, with many positive reviews to be found. All of these providers claim to be developers building a company for other developers, and their success clearly depends on that being a believable story.
Third, running your own VM is still a complicated business, requiring knowledge of Linux (absolutely no mention of any other OS, btw), configuration, system administration, security, etc. That’s been true forever, and I’m seeing some interesting attempts by some to differentiate in what must be a horrible race-to-the-bottom margin-cutting segment. In particular, I noticed:
Linode highlights their community, their library of useful docs, and their StackScripts, which could help automate setting up services on stock OSes (although I’d really want to see some security reviews on those scripts, which could easily p0wn the gullible).
No.de (from our friends at Joyent) isn’t a general purpose VPS, but has evolved the Smart platform into a node-led system which also includes various databases. I’ve played with no.de a bit, and automated deployment via-git-push could get really addictive.
Webbynode has a command-line rapid application deployment tool, which clearly has a Rails history, but is branching out into Django and Node. They also have an interesting github integration model where you point it to a github repo for both source and configuration of a package.
Overall, it’s good to see some innovation higher up in the stack. I haven’t decided who’ll get my account yet, but I’m likely to spend a few more dollars and subsidize someone who’s focusing on making my life easier, not just racing to the bottom.
We regret to inform you that the Zimki service is to be withdrawn and therefore will no longer be available from 24 December 2007. We would like to apologise for any inconvenience that this may cause.
One of the first things that will be disabled is the token usage section so unfortunately you will be unable to view your realms’ usage (code, bandwidth, storage etc.). You will still be able to create and manage your realms, and realms will not be suspended for exceeding usage restrictions during the close down period.
As of 24 December 2007 all applications and data remaining on the Zimki service will be deleted and the servers decommissioned. Users are advised to move their applications along with any associated data before the closure date. There are no plans to opensource Zimki.
We hope that Zimki has helped to increase your knowledge and understanding of the possibilities that utility computing and an online application framework can bring. We would also like to take this opportunity to thank you for your interest and support with beta testing the Zimki service.
The penultimate paragraph is funny, as AFAIK there’s no way to move the code, since it requires the Zimki platform, which they had promised would be open-sourced. Ah well.
For the first time my schedule is compatible with attending tim o’reilly’s foocamp, and I got one of the precious but hopefully less controversial invitations. It should be a great weekend (assuming I can work around my missing us passport..)
Bought a Dell laptop a year or so ago. Got the fancy warranty to deal with the inevitable juice spills. Juice spill happens. Tried to get a replacement keyboard. It is stunningly hard to get to the right part of the website (being in Canada doesn’t help). Tried calling. On hold for a half-hour. No wonder Dell’s stock is struggling. They forgot what got them successful in the first place…
So it’s finally live! I’m talking about the Google Summer of Code, a program whereby Google spends $$’s via open source foundations (such as our own Python Software Foundation) to students in exchange for code. I’m happy to see this project actually happen, and I hope it’s successful for all involved. I heard about it a while back when Chris DiBona was trying to figure out whether foundations would be willing partipants.
a compelling list of projects (I’ve started a draft on the Python wiki)
to rouse interested and interesting students (unfortunately, many of the best probably already have incompatible plans for the summer),
to rope in recruit mentors who’ll be able to help the students along if they get stuck.
Still, it’s a great opportunity to broaden the pool of actors in the open source world to include on a full-time basis those who need to pay the rent. I’m confident that the open source communities around each of these foundations step up to the plate, and we’ll end up in the fall with some code, sure, but more importantly some fresh blood with energy and a few scars to go with the experience gained.
How do you turn off accesskeys in Wikipedia? It’s a real bother, and a quick glance at the code doesn’t show a simple way to turn it off for the two most obnoxious uses, Alt-E (which trumps the Edit menu selection in Windows) and Alt-F (which does the same for the File menu).