Anderson’s knife can cut anything, taking a chunk from everything in the land.

Master Terrence’s strokes remove less with each cut.

Greg LeMond once said, “It never gets easier, you just go faster.” This quote has been a rosary for me, a source of contemplation and meditation over the years. I think of it every time a new problem reveals itself as an echo of something I thought I already solved.

When I started Puppet, everyone I worked with stressed the need for me to recognize my position as founder and CEO, to bring people with me. The summer before I stepped down, more than a decade into practicing this skill, it still showed up in every coaching session. Not because I haven’t improved - but because it is a life skill that I can improve at every year and still die with more work in front of me than behind.

I found it incredibly hard to hire those first few people. I second-guessed myself, was slow, looked like a fool, felt an even bigger one, and finally made bets based on too little information. When I rebuilt the executive team in my last year, I second guessed myself, went too slowly, made what looked like basic mistakes, and felt incompetent the whole time.

This isn’t a coincidence. It’s not a sign of my incompetence. It’s the exact opposite: It shows that I never let things get easier, instead I kept my intensity up and always focused on going faster.

LeMond knew how to win. You train, you work, you focus, and all that effort delivers leverage. What you did yesterday when giving 100%, you can do today giving only 97%. What do you do with that 3%?

If you want to win, you reinvest it. In what? Counter-intuitively, exactly what got you that leverage in the first place.

For each and every one of us, there are two kinds of skills: Those it’s worth devoting our lives to, and those it’s not. A given pursuit can be less important because it doesn’t matter to you, because you’ll never be good enough at it, or because you just don’t like it. Those that matter, though, are a perfect intersection of your love, your abilities, and your values. They will reward any amount of investment by making you better at what you care most about, and are best at.

You have to love it because you really can’t devote the time and energy needed to dominate unless you like the work. You need at least some ability. You don’t need to start with a natural talent, but you are most likely to be rewarded for investing in areas that come easier than when you’re starting at the bottom. And your investments have to match your values. Some people care most about skills that earn them money, others about having an impact on people’s lives, and yet others want to build. You have to believe that what you’re investing in will help you spiritually, not just materially. It’s the only way you can convince yourself to work as hard as you need to.

Yes, there absolutely are things you will never be good at that you have to invest in. When I started Puppet, I was an outright liability in some areas, such as business operations. I had to get better at those. But I was never confused into thinking I would be great at them. That’s what team building is for.

Given two skills, one where you’re 8/10 and the other where you’re 3/10, where do you invest? Some might recommend the skill you suck at.

Not me. If that skill mattered, you would never have gotten hired in the first place. The reason you have a job is because of your 8/10 skill. Get 10% better at your weakness, and wow, you’re at 3.3/10. Big whoop. Get 10% better at your strength? You’re at 8.8/10. That’s a huge difference.

But getting 10% better in your area of expertise is fantastically difficult. It’s your life’s work.

I am suspicious of people who say the hard things are easy. What I hear is not that it’s easy, but that they’re not trying very hard. Or they’re too embarrassed to be honest.

Greg LeMond was the fastest cyclist in the world, but never stopped trying to ride faster.

Miyamoto Musashi won more than 200 individual duels with a sword, but never stopped trying to cut faster.

Excellence requires perpetual intensity. You should be confident enough in your strengths to admit that it’s hard. It should be hard. It’s the only way to get better.

Reese’s keyboard produces truth of complexity and beauty, admired from a distance by all.

With a violent strike of Master Aaliyah’s pen, Jorge is enlightened.

I set out in January to study the realms of finance, software, and my own desire, but found myself adrift with no laboratory for experimentation. Programming was my testing ground for the hypotheses that led to Puppet, but my new questions cannot be answered in silicon. Clumsy studies provided two conclusions: Writing itself can and should be how I express and test my beliefs, but only if I get much better at it.

I started Puppet with core hypotheses, considered stupid by the experts around me. There was no point in talking about them; anyone can talk, and listeners similarly get to pick their conclusions. Only those I could prove had merit.

Proof is complicated. Few of us live in the world of math, where there’s enough overlap between objective reality and our language for discussing it that we can be truly sure. Computer science is itself lacking the unsentimental judge that the natural world provides to actual science. In truth, I should not discuss proofs, but rather disproofs. My real challenge with Puppet was to get rid of my false beliefs and let the remainders shine through, and there, programming was the perfect foil.

The best product idea in the world is worthless if it can’t be expressed in software in a sufficient timeframe. Merely building Puppet forced me to discard many convictions, and early rejects often became critical axioms undergirding the whole system. Of course, in the world of software, the compiler is only the first test. The real arbiter is your user, your customer. Once you’ve expressed your belief in a form someone can use, do they? I remember arguing for hours with a customer against a feature he requested, but the cold hard reality of his problem had me spending that night in my hotel room fixing it.

I am often asked if I miss programming. I do not.

What I miss is the laboratory it provided me, and the complete world that enveloped me, pressing in on me with its constraints, requirements, and complexities. I miss the hours I spent focused on one problem, and the complex systems needed to turn my ideas into a form usable by others.

I am thrilled and relieved to see that writing can be a new and even better laboratory for me. When I was programming full time, problems would haunt me, refusing to leave. I wore holes in my shoes walking back and forth between my keyboard and the local coffee shop, trying to line things up so they made sense. I’d scrunch up my face while drinking the coffee, drawing enqueries as to my wellbeing.

Writing is harder than that. The problems are bigger, and yet deeper inside, and harder to extract. I was never afraid of programming, but every idea I fail to write frightens me at least a little. On the far side, successfully expressing something is cathartic, a release, and it washes everything out with it.

It’s a better laboratory, because I have far more opinions about the real world than about the software world. Like those hypotheses that led to Puppet, I don’t know if they’re good. But I do know they’re different. I’ve bounced off reality enough times, asking for help from the crowd of onlookers and then getting back up, that I have a good sense of which of my beliefs are widely shared, and which are rare. For better or worse, little of what I believe, of what I think is important, is widely shared.

I’m excited by disagreement. Alignment is critical to execution, but diversity is needed for learning.

More than anything, I look forward to sharing my ideas, and as I did with Puppet, either finding them useful to other people, or having them cut to shreds to reveal new constraints, insights, opportunities. To do that successfully, though, writing is not enough: People must be able to read it.

Writing is no easier than coding, and in both, it’s ten times harder to write for others as for oneself.

One of my earliest optimizations in building Puppet was to prioritize my own productivity wherever possible. I explicitly chose to program in Ruby because I was faster there, even if it was slow and unpopular (Puppet was started before Rails was released). It took years for my programming style to reveal itself as an implicit personal optimization, highly effective for me but inscrutable to others.

My writing turns out to have a similarly mixed legacy.

All of the great shifts I initiated at Puppet began life as an essay, some extruded in a single sitting but more often sculpted over months and years. Once a hypothesis could defend itself, I then spent months and (more often) years translating it, turning it into something people could execute.

I’m proud of that work - how can I not be, when it helped build Puppet into what it is today?

But it’s fair to define good writing as something that doesn’t need a translator into its own tongue. By that measure, I’d failed.

Like a young programmer, I used to see people’s confusion as a sign of my genius and the value of my work. Now I look back with my designer’s eye and see it as a missed opportunity. We could have been there years faster if my beliefs had come with a better user experience. Kanies’s Razor is as applicable in prose as it is in code: Never attribute to genius that which can be adequately explained by incompetence.

I sit at the keyboard today as thrilled and scared as I was at the shiny benches of my science classes in college. Today I smell tea instead of banana oil and denatured ethanol, and I have no lab tech to stonewall me when I ask for easy answers, but the feeling of hope, of optimism, of a world to be discovered, understood, and fought through lifts me up just the same. I believe I can enjoy writing as much as I ever enjoyed programming, I can develop and test a wider range of hypotheses with a greater impact on those around me, and I can even learn get those ideas across to other people.

This, to me, is joy.

It should do more so you can do less

Originally published at Hacker Noon

Apple just released iOS 10.3, which for the first time puts one of their production platforms on their new filesystem, APFS. I’m pretty happy that they are finally replacing such an aging, obsolete piece of core technology, but I’m nowhere near as excited about it as John Siracusa is. All APFS seems to do is finally meet the basic needs of a modern operating system, and not even all of those. It doesn’t come anywhere near doing what I care about most in the filesystem, which is owning metadata so applications don’t.

As anyone following the discussion of federal surveillance knows, metadata has become at least as important to our lives as our data. It’s not just the phone calls, it’s who you called, when, and for how long. It’s not just the phone, it’s which ones are your favorites and how they relate to each other. This metadata is as important to us personally as it is to the NSA, because it’s what brings simplicity, usefulness, and most importantly accessibility to the reams of information we create and rely on.

The reason I’m disappointed in APFS is that it missed the opportunity to promote metadata into something as important as the data, and as a result I’ve largely lost hope that computers will ever be as useful and open as they can be.

I’ve been trying for months to convert from Apple’s Photos to Adobe’s Lightroom for photo management, and the biggest barriers to doing so are metadata. It’s easy to get all of the photos - just drag the Masters folder into Lightroom. Oh, you organized your photos into albums? Yeah, that’s all gone. You had thousands of photos categorized as favorites so you could easily find the best? Oops, we ignore that.

Google Photos has the exact same problem. I can’t imagine how anyone who cares even a snit about photos could migrate to it, because there’s literally no way to do so that preserves any metadata at all. You get to keep the photos, but not how you organized them, and your only option is to rely on what they think matters to you.

Can you imagine switching email providers but losing all of the information about who the email is from or when it was sent? That’s as idiotic as what’s happening here.

If my filesystem owned this metadata, on the other hand, no conversion would be necessary. Any application could be written to read and write it directly, or I could trivially use any scripting language I wanted to do so. In fact, I could easily switch back and forth between different apps for different purposes, because they would not be able to lock me into their proprietary databases.

This shift is hard to understand, as we’re so used to thinking of applications and their metadata as somehow intertwined. For those in the programming world, there’s a clean analogy that helps to explain the horror of the modern filesystem:

For most of the history of software, applications have been written to directly access databases. They needed to understand the structure of tables and rows, and they had to write and manage their own queries for interacting with that data. Because of the complexity of duplicating this ability, applications tended to own their data entirely, and no other apps could get access.

We’ve recently instead seen a trend towards building very simple data services on top of databases, so that no app directly accesses them. This data service provides an API that understands how all of the data works, and also provides all of the key operations that can be performed on the data. This way, any application can access the data - for both reading and writing - without worrying that its use of the data is somehow incompatible with another consumer. You get the added benefit that any optimizations in the data service help every consumer, not just the one or two apps you remember to change.

Those data services are exactly what have enabled modern applications to scale to the level they do. The application developer can focus on their goals for the data, relying on the data service to provide consistency, portable operations, and everything else.

In the land of the operating system, though, we have no such luck. Everyone who talks to the filesystem gets about the thinnest interface imaginable: Get me a blob of data. Yes, there’s basic metadata (who owns it, when it was created, etc.), but none of it is or can be custom to the application, which makes it useless. Can you imagine how unpleasant it would be if we had to shoehorn both the subject of an email and the subject of a photo into the same keyword?

As a result, everyone turns that blob that the filesystem is trusted with into a database, and that’s where they put all of the metadata. Some even put the data there, even if it means huge database files and entirely non-portable systems. Really great app developers then go to the effort of building a full read and write API on top of this database that allows you to manage and port everything in the database. That’s enough work that most people skip it, though.

I would never have seen this missed opportunity in the filesystem if I hadn’t used BeOS. I went to Reed College, where Steve Jobs famously attended briefly (I made the unfortunate mistake of graduating, unlike him). This meant I was a Mac user all through school, which were the bad days of Quadras and MacOS 8 and 9. shudder I switched to BeOS because, well, I could, and it’s what really turned me into an OS junkie. There were many awesome things about it (many of which are now in OS X), but my favorite became how much responsibility the OS took for metadata. Applications became smaller, simpler, and in many cases, you realized your whole concept of an application was a way of organizing metadata, which meant you could rely on the OS entirely, and needed to write no code.

The greatest irony is that the author of Be’s awesome filesystem, Dominic Giampaolo, is now a lead engineer on APFS. Through the work of him and the rest of the team at Be, I came away with a clear idea of the value of my metadata, and especially the benefit of the operating system putting it on center stage.

So while I’m happy that my phone is no longer running the oldest, crappiest filesystem around, and I’m looking forward to the days when my computers aren’t either, I’m not excited. Until it was released, I could always dream that Apple saw the same future I did.

Now that the new filesystem is here, that hope is lost.

Just because it’s working doesn’t make it right

Startups are in a constant life and death struggle, where a short succession of failures can destroy them completely, or a sudden string of successes can lead to growth and stardom. Exactly because of this constant risk of death, startups grasp at every flash of success, and their eventual failure can often be traced back to those successes they bet on, rather the roadblocks they hit.

This means what you decide to double down on is as important as the fires you decide to fight. This is true in any company - a specific form of this is covered in depth by Clayton Christensen in “Innovator’s Dilemma”, where successful companies over-solve their customers’ problems rather than finding additional customers. It’s especially true for startups, though, which are so risky that every potential customer, employee, investor, or partner can be the difference between making payroll and going bankrupt.

Steve Blank and Eric Reis have done a great job of teaching the world about the importance of quick learning and iterating until you get wins, but reality is more complicated. Say you’re in the middle of rapid product iteration and you get a rash of customers who sign up but are completely different from those you have been trying to attract. Is this awesome or horrifying?

How could new customers be such bad news, you ask? Easy: They could be the wrong customers. “Wrong” could mean many things. They could be from a very small market, such that success with them isn’t enough to build a scalable business; they could be unprofitable, in that they’ll be very high-cost to acquire and support, so every added customer loses you money over time rather than making it; and most commonly, they’re just not the customer your company strategy set out to find.

Eric Reis might say finding this kind of customer is the sign you need to pivot. That might be true. But it’s not automatically so. You were right enough to get this far, why give up so quickly on your definition of the perfect customer?

Because once you change your direction from focusing on the customer you want to customer you have - that is, once you start focusing on whatever seems to be working rather than your goals - you will quickly find yourself running a different company than you had planned.

This is one of the hardest problems for entrepreneurs: How do you decide when good news is leading to a better place, or a fatal compromise to your dream?

This came up constantly for me while running Puppet, in almost every part of my job. For instance, we did all of our high-growth hiring from Portland. Because of the small market and especially small tech sector, you have to be pragmatic about who you hire. You look up in 2 years, and you’ve got complete teams built off that first compromise hire. That first developer hire was a success, but changed your team’s direction. Now that that person has hired a team, or at least provided the template for the team, are you where you wanted to be, or completely off track?

You can argue that this isn’t a problem we would have had in the Bay Area, but not compromising when hiring is as pernicious a myth as true love. Even if you think someone is perfect, they’re still a real person, with opinions, behaviors, and habits that affect the destiny of your company, in good and bad ways. Again, the point here isn’t that you failed when you hired this person - the point is that the success you experienced changes your destiny in ways you can’t undo or predict.

That first customer started making feature requests and filing bugs, and two years later you’ve built something that can attract a lot more customers like them. Yay, or boo? There’s no right answer. I know one company that was almost destroyed by its biggest customer, and easily lost two years of roadmap progress focusing on them, but I know another company who built an amazing business by initially focusing on a couple of heavyweight customers.

When no decisions have been made, all life in front of you is opportunity. Choosing one opportunity inherently closes off others, and great decision makers know and accept this. If anything, sometimes a decision is great because of the opportunities it rejects, rather than those it selects. Running a startup is an accumulation of these bets, these selections and rejections. It’s hard to predict the consequences of even one of these big decisions, and it’s impossible to know what their legacy will be in a few years. You might end up exactly where you want… but you might be sowing the seeds of your downfall. That’s what makes the job hard.

The only tool I ever developed for managing this over the long term was to hold clearly to a specific set of goals, and to a specific strategy. When successes came along that didn’t fit our goals or strategy, we could decide to change them. But if we chose not to, then we essentially ignored it. It was a false positive. On the other hand, if you look up and a series of false positives starts to look like a trend you want to refocus on, then you must first rethink your goals and strategy, and only then switch your execution.

Or at least, that’s what I did. In truth, though, I don’t think I held the line enough. I think I was too willing to listen to people who were tactically great but didn’t buy into our strategy, and I was too willing to believe someone else’s expertise should take priority over my opinions. That doesn’t mean these were all mistakes, but it does make my time at Puppet easy to see through the lens of compromise and regret rather than the great success it really is.

In the end, when I look at how Puppet evolved, my sharpest pains come from compromises in strategy, and my brightest joys shine from a flawed execution in service of a pure vision.

We’re seeing a rapid separation into SaaS companies and suppliers to SaaS companies

Software as a service is the most important technology business model innovation my lifetime, and before too long all important technology will be either provided as a service, or be an expert tool purchased in order to provide a service.

Software as a Service (SaaS) will not just restructure the entire economy of technology but dramatically improve it, because the rate of learning in a SaaS company is orders of magnitude better than for on-premise software (and oh my god better than for hardware). A really good on-premise software product is released and upgraded quarterly, meaning its fastest learning cycle is a 3 month cycle. A good hosted product is released tens of times a day. Twenty releases a day, times 5 days in a week and 12 weeks in a quarter mean that a SaaS company gets 1200 releases out in the time an on-premise company gets a single release.

Every release is a learning opportunity, which means a SaaS company can learn, and thus improve its user experience, more than 1000x more quickly than even a high-performing traditional software company. With this kind of difference in the rate of improvement, before too long anyone who isn’t a SaaS company is irrelevant. Literally - you can just ignore them.

Except of course that’s not true. If you’re a SaaS company, you can’t rely only on other SaaS companies; there are problems that can’t be solved off-premise, complexities that can’t be abstracted into a service, and technologies needed to provide your service that can’t be procured as a service. The most obvious example is that it all has to hit silicon at some point, and someone somewhere does actually have to buy CPUs, hard drives, and memory, with enough duct tape and baling twine to hold it all together and enough fans to cool it all down. While this is one obvious exception, there are plenty of others.

Thus, even with the learning rate of SaaS companies, and the resulting obsolescence of most other kinds of tech providers, there is still room for those who sell to the SaaS companies. This is somewhat akin to the ‘49 gold rush, when there were gold miners and those who sold to gold miners. Both were valid pursuits, and tightly entwined, but they were very different businesses.

In the early days of the software world, and especially the early days of the enterprise IT department, every company evolved to be mediocre at any kind of technology, and they all looked pretty similar. They could do just about anything, and they sold to pretty much everybody, but they couldn’t do anything particularly well. This impending world of SaaS companies hires only experts; they are absolutely fantastic at everything they care about, and they have an equally fantastic service partner to whom they can outsource anything they don’t care about.

In some ways, this world is even more dangerous to the status quo of the technology industry. Our entire industry has evolved to sell to generalists: We build pretty decent stuff, and sell it to people who are ok at managing it. Our stuff can’t be too great, because great stuff requires expert users, and we can’t demand too much of our users because they have to be competent at nearly everything within a very large field (“networking”, “hardware”, “applications”), so they can’t invest in being particularly excellent at any one thing. Also, great stuff would be really expensive, and companies don’t buy expensive stuff, because no particular investment is all that important to them.

In this new world, there exist only experts, whose entire mandate is to specialize in the specific tools and technologies that will allow a SaaS company to excel, to win. For the investments that can make a real difference, no cost is too high, and no expertise is too narrow. For investments that can’t make a real difference, of course, we just find a service partner who cares so much about it we don’t have to think about it. Thus, as a SaaS company, I only invest in technology when I can afford to invest in experts, and I immediately reject any technology built for generalists because I can’t achieve outsize results with tools built for generalists.

Here we reach a future where there exist only SaaS companies, and companies who build expert tools for narrow use cases that exist within SaaS companies. Anyone else is fiercely riding a bicycle while watching a train recede into the future. The pundits of today are correct that the cloud is the most important trend of the modern era, but they’re wrong in thinking that AWS or Cloud Foundry hold a candle to the restructuring that SaaS will wreak.