I don’t know what the rest of the world thinks when they use the phrase ‘power tool’, but for me it’s visceral, literal. My experiences using them and watching them transform my family’s work permeated my time building Puppet. These power tools aren’t little plugins to expensive frameworks, they’re large capital investments that dramatically change your job.
I grew up building houses with my dad. The worst task he gave me was trying to paint a set of louvre doors for a closet while in high school; I had to flip the doors over every 90 seconds to catch drips getting through the slats. After three days of misery, my father relented and rented a paint sprayer, with which we finished the job the same day, at a much higher quality.
Around the same time, my dad would rent a pneumatic nailer for big framing jobs. By the time I finished college a few years later, that critical tool went from borrowed to owned and traveled everywhere with him. Initially used only for large jobs, most contractors now have multiple nail guns to cover framing, trim, and every other use case, and the air compressor needed to power it is as important as electricity.
It might not be obvious, but both of these are examples of automation. You replaced a very manual process - applying paint, or nailing things together - with a machine. If this were a factory, these days you’d call those machines robots, but because it’s a construction site, we just call them tools.
And these tools were expensive. Even with how much faster we finished that painting job, I expect it cost more to rent the sprayer than to finish the work manually, because of how little he was paying me. (This does ignore the soft costs of listening to me complain, which were likely high.) Even today paint sprayers and nail guns are often rented rather than purchased, because good ones cost a lot of money and aren’t needed all the time.
It’s no surprise that discussions of tools and productivity are easier to understand from my experience as a carpenter than as a sysadmin. There’s plenty of room for arguments about what is or is not a software power tool, but when it costs more than a week’s wages, it trails a bright orange cord everywhere it goes, and it can nail your hand to the wall while you’re standing at the top of a ladder1? It’s a power tool.
There’s a common story about what robots and automation do to people like my dad (and both of my brothers, who followed in his footsteps): It steals their jobs and ruins their lives.
What utter poppycock.
If you think of your job as driving metal spikes into wood, then a nail gun is a mortal threat. But if this is your value add, your biggest danger was never automation. My dad never sold his ability to join raw materials together quickly; he sold homes, he sold the opportunity to enjoy your house and family more. How did these new power tools affect that?
They were awesome. Painting and nailing are classic examples of menial, low-value work, and yet we spent most of our time on them. All of the differentiation we offered to our customers was packed into a narrow slice of work, because implementation took so much time and money. As we were able to bring more powerful tools to bear, the menial work shrank and larger portions of our time could be spent on design work, customer interaction, and tuning our customers’ homes.
Interestingly, my father’s next career step was even more pointedly about experiences enabled by tooling. He took a job with a state hospital in Tennessee, fabricating custom furniture for severely disabled patients. Suddenly he was using industrial sewing machines for upholstery, and partnering with medical professionals to design multiple beds for each patient, enabling them to be happier and more comfortable (and also avoid bed sores, thus saving hundreds of thousands of dollars per patient). Given the tragically minimal budget allocation for this kind of work, every dollar saved through automation and tooling directly delivered health and happiness to his patients.
It’s no wonder I see the value in power tools, that I am more conscious of the benefit they can deliver than the loss of low-value menial work.
I had a similar experience as I was building Puppet. I would meet executives and salespeople (I don’t know why it was always them) who would say, “Oh, automation? Great, you can fire sysadmins!” No. Beyond the obvious reality that I was selling directly to my users, who would never buy on the promise to fire their coworkers, that was just not why we were valuable.
Puppet gave people a choice between lowering cost but keeping the current service quality, or keeping your costs flat while providing a much better service. “Wait, making things better is an option? I didn’t know that!” Most companies were aware that their IT sucked, but they only knew how to measure and manage cost, so that’s what they did. Once you believed in the power to make things better, power tools turned out to be great investments for both the user and the buyer.
By letting people spend more time on the parts of their work they enjoyed, the work that makes them special, we also delivered higher quality experiences for their customers and constituents. “Spend less time firefighting and doing menial work, and more time shipping great software.” If the heart of your skillset is clicking buttons or responding to outages, Puppet might have been a threat to you, but our users knew where their real value was. We helped them spend more time there and less time on the boring, low value stuff. The sysadmins hated the work, the customers hated to need it, and the executives hated paying for it. Great, done, don’t worry about it.
That’s partially why productivity has stagnated2. The world has not changed that much - some of the greatest improvements to productivity come from making large capital investments in tooling for your workers - but how we spend our money has. People balk at a $5k computer, when the Mac IIci would cost more than $13k in today’s dollars just for the hardware, yet was a powerhouse in desktop publishing. This is to say nothing of how the mobile app stores have driven down what people are willing to spend on software.
Yes, Adobe’s software is expensive, but it’s that price because it delivers so much value. If it didn’t, no one would buy it. Every large market should be so lucky as to have the collection of power tools that graphic designers get. It sounds crazy, but we’re suffering from not enough expensive software. Instead of building the most powerful software possible and finding customers who see its value, companies are building the simplest thing they can and trying to get everyone to use it.
There are bright spots in the industry, like Airtable and Superhuman. I’m hoping they help to shift momentum back to automating away the tedious work and enabling focus on what humans excel at.
More powerful tools improve your life, but they also make you happier even if you can’t buy them. They tantalize you, promising you great returns, if only you can come up with the cash. And they’re maybe just a little bit scary, warning you that buying them is not enough. You must master them.
A friend of ours managed to do this when working alone in the time before cell phones. ↩
Yes, I might be being simplistic to make a point. ↩
I did more writing in 2017 than I ever have before, probably more than I had done in my entire life. I stated at the outset my desire to write in order to learn, and publishing regularly has delivered. More importantly, it forced me to capture and share many of the results of my R&D. The writing topics were primarily venture capital, the software startup ecosystem, and founder optimization, but one of my weaknesses is a drive to the meta, spending time on the tools and systems rather than the work itself. Unsurprisingly, this ground is as fertile in prose as it is in code.
I have done all of my writing on Ulysses, which has been a good tool. I do miss much of the power of Vim for managing text, but the usability trade-offs are usually worth it. Plus, Vim is not exactly strong on wrapped text. Ulysses’s most important capability for me is that it transparently syncs between my phone , both of my iPads, and all three of my computers (yes, I know). I can open any document on any device with no worries (given my iPads are on LTE, and thus can always download updates). Second to that is the ease of getting data out. I was able to text a 5k word document directly from Ulysses on my phone to two different people, because it converts anything to PDF and can send out via the share sheet. It publishes directly to most platforms, so I used it to post everything to Medium. I do wish its backend storage was more visible, so I could version control everything in git, but you can’t have everything.
The only time I ran into troubles was publishing my travel writing - I was copy/pasting photos from Apple’s Photos app into Ulysses, then publishing into Medium, and it turned out that no part of this process downscaled the photos, meaning each was about 20MB. This is fine when you’re on solid wifi, but not so good from a campground in Yellowstone. The primary reason I did not write during the second month of my trip is it was just too painful to publish. Ulysses really fell down here, because when it works, it works great, but when it fails it is miserable. I’d just sit and stare at my iPad for ages, with no visibility of failure or success until the very end, and it would often manage to upload the text but not the photos. The only fix for this was to delete the draft from Medium and try again. Not so great.1
Over the year I became concerned, though. I heard about Andrew Chen, who built a following on successive publishing platforms (e.g., Blogger), each popular and well-funded, only to have the companies disappear (because that’s what Silicon Valley does to most companies). He eventually realized that the only way to consistently reach those who were interested was via email; it’s never going away, subscriber lists are easily portable, and of course everyone knows how to use it. He had to move from letting a platform own his audience to taking control directly, and email was the only real way to do that.
I think Medium is pretty good, and at least for now it has no shortage of funding. Even last year, though, it made well-publicized efforts to retool its business, a clear sign that whatever it was doing before wasn’t working. It seems unwise to bet that a follower on Medium will have any meaning in a few years.
It just so happened that I ended the year with an experiment in their new business model. My series on Venture Capital, in partnership with NewCo, was published behind Medium’s paywall. While the experience was positive overall, and I got paid more for those pieces than I have for, um, all of my other writing ever, earning money isn’t my real goal in writing. Yet, paying writers is exactly what Medium has to figure out in order to attract the content and audience it wants. I appreciated the extra attention working within their paywall provided, but I came out thinking it’s unlikely to be the right path for me in the long term. My goal is to maximize the reach of my writing, and it’s more about the arc of all of it rather than a couple of heavy-hitting pieces getting the most attention, which means our respective goals are orthogonal at best, and in direct conflict at worst.
While I have not written about it, I spent a lot of last year fascinated by email, inboxes, and how we consume content. I’ve been a deep fan of RSS from the Bloglines days (Unread is my reader of choice these days), but RSS is mostly dead (thanks, Google!). The kinds of sites that produced great feeds back in the day grew too big for a hobby. They became either wildly profitable for a small team and thus got destroyed with ads and shitty content (hi Boing Boing!) or people had to back away because it was just too much. Mainstream publications like the NYTimes might still publish RSS feeds (I literally have no idea), but RSS is a poor fit for how much content they produce.
When I look around, it seems that the RSS feeds of ten years ago are now awesome newsletters. See how often your favorite sites talk about their newsletters vs their feeds. I’m pretty confident I’ve never heard a podcast ask me to sign up for a feed, but many casually mention their newsletters. (Well, to be fair, podcast feeds are just RSS, so in that sense it’s survived, but I think we can agree it’s a very different use case.) The market has moved.
This shift is not all peaches and cream. The superiority of email as a publishing mechanism has not brought with it a superior reading experience. Long form content from writers I follow does not belong in the same inbox as requests for coffee meetings, yet that’s where we are. (I asked the Feedly CEO if they would please please add newsletter subscriptions to their platform, and he could only say they were considering it. If you want to build the newsletter reading app, hopefully with an index of great letters to subscribe to, count me in as an early contributor.)
Even with its downsides, I’ve decided to move my writing to my own site via a newsletter-first publishing model. From now on, everything will be published first to that site and the newsletter (sign up now!). There will be exceptions, when publications offer enough exposure that I give them some period of exclusivity. I will also indefinitely duplicate each piece to Medium on a delay, to take advantage of the audience I already have there.
There are enough examples out there that I’m pretty confident this is the right long term answer, but it’s early enough that it feels like an unstable experiment. Like most, I love and hate email, and I am a bit bummed about the extra infrastructure involved in this system.
If you’re still an RSS person at heart, you can always just subscribe to the feed, and if Medium is your bag, at least for now you’ll be able to see everything there, too. Of course I can’t promise how this plays out in the long run - else it would not be much of an experiment - but you can bet I’ll work hard to find the right way to talk with the people most interested in what I have to say.
One note before I go: I could not have made this transition without the help of Mike Julian. He helped set up each of the services, mediated the hiring of a consultant to modify the site, and connected all of the services together. He’s got a great book on monitoring, and is available for consulting on monitoring and observability. I can’t recommend him enough.
Thanks for following on so far. I’m excited about another year of writing.
It’s worth noting that Apple’s iCloud Photo Library did wonderfully here; I could upload every photo to my iPad, and it would sync when it had good data, and sit quietly when it did not. I love asynchronous protocols. ↩
I have been a passionate advocate of Trello almost since the day it came out. Visualizing the work is incredibly powerful, and its use of priority by ordering is so obviously better I could not imagine using anything else. I have brought Trello into companies, and caused hundreds of people to use it. Recently, though, I stopped using it in key functions, moving my personal tasks into Things and other work into the applications that own it. It started as an experiment, but quickly snowballed into a full transition as I realized how much better it would be for my core use cases.
I am a productivity junkie, from GTD to Deep Work, and while I know the tools aren’t the work, they’re one of the fastest and easiest ways to change a system, bringing new perspective and abilities.
I initially adopted Trello for managing my personal tasks. From there, it slowly spread to group and company use cases, and by the time I left Puppet in 2016 it was baked into our core business operations.
I had considered switching away from it for my personal tasks a few times, because it’s been obvious for years that it was not a fit for how I managed and processed my life’s flow of work. Where I had always taken great pride in my productivity system, I could no longer muster the discipline it required, which meant it was extraneous process and I was also failing to do my job. I sank so low that I began sending person to-do items to my inbox. shudder However, my life as CEO was busy, yet not really built around task execution, so it was never worth the cost of the tooling change. Or so it seemed.
While I abhor tooling changes for their own sake, sometimes you just have to experiment, and when I had a run of open space I decided it was time. Even if other places were worse, I was better off spending time elsewhere to get out of my funk than I was continuing to stare at this board that just wasn’t working for me.
As I looked to move away from Trello, I wanted to switch to OmniFocus, but the timing is wrong. I recently downloaded the current version, but it reminded me why I stopped using it years ago: It asks too much of me1. The biggest disconnects were the need for contexts, and the inability to set priority by dragging items around. It looks like the impending next version will be a great compromise for me, but I can’t use it until it arrives.
As I began moving tasks into Things, I was slowly reminded what it was like to have my task app be, well, an app. Before Trello I used Asana, at a time when its mobile experience was just a badly packaged web page. It’s been years since I had the pleasure of working with someone who bothered to build for the platforms I used. What a difference it makes.
Processing this small but meaningful change took over my brain for the rest of the day. I could not let go of the transition, but I also could not stop thinking about its consequences. “Thinking” is being too generous. I had not seriously looked outside of Trello for years; all of my habits and instincts are built within its constraints, its physics model. That meant wide open opportunity to construct whatever world I wanted, but also the need to do so, and embarrassing mechanisms for getting in and out. Creating a quick task in Trello is most easily accomplished by sending myself an email, which tells you how seamless the core experience really is. Telling Trello of work to be done in another application usually required custom applescripts to connect the two (e.g., creating message:// urls).
The keyboard’s change in utility keeps hitting me. I did not realize how rarely I used Trello shortcuts for their sheer disagreement with my hands. Well, partially, they were insane - ‘c’ archives cards, where I kept thinking it would create them, and repeating mistake kept my hands off the keyboard - but also, they were just different. I keep visualizing myself sitting at Things, fingers poised over the keyboard, in a world that makes sense to me. It’s pleasant.
I have a lot of work in Trello. Basically everything I’ve recorded in the last 4 years about what I could or should do is there, and it’s been the platform for organization all of that information. I’ve always been pragmatic, so even when I built it I would not have said it was the “best” answer, just that it worked at that time for my use cases.
Now I am questioning everything I’ve done recently - not to ask if I’ve made a mistake, just wondering what the world would look like if I had done that work differently. Are there projects languishing from sheer friction, that would suddenly become easier if migrated?
The answer in at least one other case is ‘yes’. I had a Trello board for managing all of my writing ideas, and I had a standard right-to-left workflow from idea to writing to done. I just never really used it. Instead, I did all of the management of writing in the same place I did my actual writing: Ulysses. I built a workflow from folders2 instead of lists3, and eventually I realized that the Trello board was hurting, not helping. The ideas I put in there languished, and Ulysses is just as good at capturing a sentence or two about a post as it is for writing the entire thing. The board is now closed, and my writing is in the same place as all of the work needed to manage it.
Tools do not make the difference, but tools can make a difference. A thing can go from undoable to tractable with a small change in tools, just as a math problem can go from impossible to trivial if you flip the equation around a bit.
I know Things won’t work for many of the use cases I have in Trello, partially because it’s so obviously a single-person tool, and much of my work does not have a home I can move the organizational work into. E.g., I have a list of lists in Trello of all of the day trips I could consider taking around Portland, and I can’t think of where I’d move that. I don’t know where those other problems get solved now. Maybe they don’t change.
But I know my standards have been reset. I know I made a mistake letting myself sit in one tool for years, unhappy but unwilling to invest in the change. It’s a wonderful feeling to be looking at the world through new eyes again.
To be fair, I also stopped using it in the days when it took 45 seconds to open the app on an iPhone, because that’s how slow the data synchronization was. ↩
I’m a tech founder and recovering SysAdmin. I helped found DevOps and grew up in open source. These days I am convalescing from a decade of physical and mental neglect while running Puppet.
Read more
