Want skill? Build human connection.

We often think of skill as built via practice. The science is broader: skill unfolds via bonds of trust and respect. So skill's in trouble in our distanced, digital, genAI world. We have to do better.

“Now just focus on the two hands. Do variations there. One with one hand guiding another in some skilled activity. Another with four hands in the image. Another with two people working together, but emphasis on their hands.”

Think of your most valuable skill. The thing you can do reliably, under pressure - and that looks like magic to those nearby.

How did you learn it?

No matter who you are, no matter where you live, no matter the skill, a core part of the answer is practice. You stunk at it in the beginning. You tried and tried and tried, made many mistakes as you went, and, slowly, you started to get better. You put in the time, plain and simple. This just kept going, the more you leaned into it - maybe well past what you thought was possible. There’s a mountain of research to back you up here, and that explores the rich tapestry of experience required for more and less productive practice, when it comes to skill. If you wanted to start and stop at one person’s research to understand this, it would be hard to do much better than K Anders Ericsson, a psychologist who studied expert performance in a variety of domains, and literally gave name to “deliberate practice”, a technique that’s now standard operating procedure for elite athletes, musicians and their performance coaches the world around. Anyone who has been taught that visualization and mental rehearsal are a good idea, who’s read that practicing one thing at a time, with intention, will get you farther, faster - or anyone who’s actually read Malcolm Gladwell’s “Outliers” owes their insights to Ericsson.

I reencountered him - and discovered dozens of other researchers, across 13 disciplines - as I pulled together the world’s research on skill development for my upcoming book. They speak with one voice: the “how” matters quite a lot for practice. In the language of my book, we need healthy challenge - the right kinds of difficulty, experienced in the right ways - and healthy complexity - the right kind of engagement with the broader context for our tasks. But to put the details to the side, it’s clear: there’s no way to “git gud” without practice.

If that were the whole story, we could all build the skill we needed to handle the wild world of work we’re building for ourselves. We’d have the internet, YouTube, ChatGPT, and could get to work bootstrapping ourselves to readiness.

Yet we’d fail. In fact, we couldn’t even start.

Let me give you a window into the fuller - more human - picture of skill by honoring Doug.

RIP, Doug Michel - my competitor, mentor, and friend

Rewind time to three years ago, right in the middle of the pandemic. I got an Oculus virtual reality headset for my birthday and my wife happened to pick up an indoor spin bike from her gym that was going out of business (dang Covid). I *hated* cycling - I'm a runner and have never been an athlete - but right away I bought a bluetooth cadence sensor for the bike, one that would pair with the headset. Why? When I was living in Boston, I had tried a virtual reality cycling system from a startup called VirZoom. Turns out, they survived, found a market, and were doing fine. Cycling in a virtual race - or along roads in google maps - was the main draw for most VirZoom users. Not for me.

I played a *lot* of videogames as a kid, and while I toned it down as an adult, I never stopped. So I remembered that VirZoom also had simple, somewhat cartoony games. Those were fun. And my foot was hurt.

So I hopped on, and went straight for the fun - pedal and I got to fly as a pegasus, chase after bandits on horseback, zoom along a riverbed with a helicopter, or get in a tank battle. And boy, tanks were REALLY fun! Especially because you could compete against another human being.

I started to do it a couple times a week, huffing and puffing, getting shot at, and so on. Whenever I got a multiplayer match, I'd get trashed. People were so *accurate* from such a long way, and -wait, what- they'd always have some magical shield on right as I came over a hill and... dang. Game over.

But again, the fun kept me in. I certainly was working hard!

So then one day I realized there were leaderboards for all these games! And I was... in 119th place?! Wait a minute. Wait a MINUTE. Who's in first, and their score is... WHAT?

That's when I thought it was impossible. Literally. I saw a score that was like 27,000 for a 30 minute solo tanks run. I was getting say 11,000 working as hard as I possibly could? They must have changed the game, I thought. Like there was an old sensor that let you go 2x faster, or tanks were worth more, or whatever. So fine, I can't ever get into the top ten.

But let me try for the top 30. That's what I told myself. So I started putting in the time.

And that's when my scores started creeping up. No big revelation, no fancy tricks, just showing up four, then five days a week and working hard, watching and trying to figure out... WHAT. ARE. THEY. DOING.

There were no tutorials, no help. So I was improving pretty slowly. Matches were hard to come by - I didn’t know when anyone would be on!

So finally, in frustration, I reached out to Doug (his username was bakerbiker) on an old, janky user forum. Doug was in the top ten, and climbing. An avid player. He beat me by big margins, every time. So I was a little intimidated, but I had to know: how in the heck do you get that kind of score?

His (friendly) answer: I don't know! Just play. Get the pickups. Don't get hit. Hit stuff. Get good. So… yeah, not too helpful. That’s the way with so much skilled work - much of it just isn’t explainable in language, even if we tried. This is at the heart of Polanyi’s and Moravec’s paradoxes - a basic barrier to development in AI, not just human skill development.

But that was the beginning of him coaching me. We’d play. He’d win. We’d discuss. I’d ask. We’d schedule times to play. We chatted about other players.

And my scores kept creeping up. I got in a LOT better shape - maybe the best of my life, I started to realize - and I started to figure some stuff out. Like how to orient myself so I always had a clear shot, and that there was a pickup that let you fire faster.

And then one day I beat Doug! He warmly congratulated me. Said it was a great workout. Mentioned how much I’d improved. Said he’d become better as I had. Wow. I felt so proud, and he was delighted. It was back and forth from there on in. Then we started to reach out to other players. To assemble a dm group of devotees, eager to battle it out. Soon we had Crystal. Leigh. Superlith. Jazmin. Kathy. BillyDWolf. DrSteve. Paul. Even Eric, the CEO of VirZoom! We all chatted between matches - about the same stuff. Tactics. Goals. And through all that friendly competition I pushed myself to do better and better. The day I hit #1 in all solo-play categories, everyone congratulated me. Same when I topped the head-to-head combat charts. There was no trick. There was no man behind the curtain.

Except for maybe Doug. Beyond the bond that gave me a sparring partner and my motivation to improve, he was the whole group’s cheerleader, our glue. Always up for a match, welcoming to newbies, cheery and grateful for the workout, whether he won or lost.

Then we lost Doug.

His wife dm’d us via his account one Sunday late last year. With a heavy heart, she let us know he’d passed away the day before. I had played him that morning! We were all in shock. Devastated, really. By this point, we all knew each other well. We knew Doug founded and ran Columbia Bread, basically the first successful bakery that catered to people with gluten sensitivities. And he wrote and illustrated a wonderful and strange set of comics. Sue told us she wanted us to know how much the Tank Nuts (the name that Crystal gave us) meant to him. We sent her and her family a custom condolence package. A few of us still fly his flag in the game today (you can select a custom flag for your tank as you win more games). When a match is near over, we often fire a shot over each other’s heads - a salute that Doug invented. We talk about how we miss him.

We all got better because of Doug. As players and as people.

We need healthy connection to learn

Hit me with a few more. Make a choice between healthy and unhealthy connection clear. Center on this text: "That masks the fact that we simply can’t get healthy challenge and complexity without the kind of healthy human connection that Doug and I enjoyed. "

We often celebrate leaders who received The Devil Wears Prada–style mentorship - isolation, lack of direction, harsh critique - but these are typically the harbinger of poor skills development. And all too often, we think of skills in individualistic, egocentric terms: all you need is a good head and nimble hands. That masks the fact that we simply can’t get healthy challenge and complexity without the kind of healthy human connection that Doug and I enjoyed. So, if challenge and complexity are more about the “how” of skill, connection is more about the “why” - more to do with a rich, complex, and very human landscape built on warm bonds of trust and respect. So say the fields of psychology, education, and management, but recent developments in neurobiology and artificial intelligence lend even more credence to the idea, too.

This has an externally focused aspect that’s about practical opportunity: Want a chance to work at all, let alone get access to healthy challenge and complexity? Good luck getting it on your own. One or more people have the power to admit you to the party, let you stay, and lend support. Without my connection with Doug, I wouldn’t have even gotten a steady stream of practice opportunities. Zero chance I’d have improved like I did. Beyond simple access, there’s the question of advice and guidance. Remember, the first time I asked Doug how to get a high score, his answer was “I don’t know.” But he knew a lot of things, even if many of them were in-the-bones skill. He just didn’t know, like, or trust me yet - well enough to try to messily explain his experience, give me feedback, and describe a technique or two. I hadn’t earned the privilege.

But there’s also a more internally oriented aspect that’s about personal meaning: we get motivation for our work when it builds respect and trust with those who share our values. These are questions of the heart, like “Do I feel connected?” or “Do I feel motivated to gain the respect and trust from those I aspire to become?” We often dismiss these as unconnected with hard-nosed skill and results, when in fact they’re closer to the main event. I looked up to Doug. In the beginning, this was strictly numbers-based: his scores seemed stratospheric, and I couldn’t quite believe I had the good fortune of finding him on the user forum. At the same time, it was pretty clear we shared some core values: he was kind, a little wacky, inclusive, and of course really cared about the game. I wanted to earn his respect and trust. But he also told me later it was a real shot in the arm that I was so engaged with him - it meant a lot to help someone, to see them improve, and to grow a small community of practitioners.

The human connection was *why* we both engaged in the activity, really. Yes, we wanted to play a game, and get better. But the animating force - and the way it unfolded - was our evolving bond.

No, this doesn’t mean ending wfh and genAI

I tend to be conservative on remote work. I happen to believe that it can’t possibly equal the productivity of in-person work… And then leadership and mentoring: I would not be where I am today had my bosses not been able to see me perform in some physical space. I’m convinced of that.
- Bob Iger, CEO, Disney; a16z podcast, Jan 6, 2023

Now that news of my book is out there, I’m giving more talks about it to groups of executives, professionals, policymakers, and so on. These are a delight, a privilege, and an education, all wrapped into one.

One lesson I’ve learned is that leaders are very worried about the loss of human connection associated with remote work. My findings stoke that fire rather dramatically - understanding the specific relationship between human connection and skill development gives them a new vocabulary to articulate what was only a creeping suspicion before. Then I wake them up to another central finding in my work: that we’re taking advantage of intelligent technologies like robotics and genAI in ways that allow experts to get better results with less help from novices. And that novices are building less skill as a result. That catches most completely off guard: not an issue they’d had on their radar.

This is about when they really start to speak up about fighting for human connection. This isn’t just practical, really. Yes, they see a dramatic, multifaceted threat to informal skill development throughout their organizations. That’s a dramatic issue for performance and adaptability. But it’s also a direct threat to their values. They came up through a system of collaborative struggle with experts. Looking up to them. Learning vicariously as they were included, bit by bit. Earning their trust and respect. And as they became experts themselves, paying it forward to novices. This feels healthy, right, and proper. A beautiful system. They react pretty violently to seeing it under threat. And many of them grab for Bob Iger’s solution pretty quickly.

As I was wrapping up a talk, one CEO summarized this neatly: “you’ve just given me all the ammunition I need to make everyone come back in the office.”

We all had a *very* interesting and valuable conversation after that.

It settled on the same reality we all have, more or less: remote work is here to stay. So are intelligent technologies. A wholesale return to the way we used to do things would be destructive, but it’s also not an option.

Finding positive needles in the negative haystack

“Okay, and here's the text from the last section - I need four images to represent or connect with it, please act like a world class artist with a distinctive style for each image. Review your work for creativity before you share:”

So the more useful question is how do we get the benefits of remote work and intelligent technologies while improving on the past? Not just get back to where we were. That would be great, in many circumstances, but let’s face it, pre-pandemic work reality wasn’t ideal. So all this disruption provides unique opportunity for a species that - collectively - prefers predictability and stasis. In the language of this post, how could we enhance bonds of trust and respect between experts and novices because of remote work and intelligent technologies like generative AI?

Answering that question in detail is the job of the back third of my book, but a simple answer is literally sitting right in front of you.

I never met Doug in person. Didn’t even know what he looked like for a year and a half. And that was the most important year and a half, when it came to our bond and my skill. In fact, I’ve never met *any* of the Tank Nuts in person. We live all around the world! Those bonds mean a lot to us, and have allowed us all to become far better players than we would otherwise. And none of them would be possible without the technologies behind remote work.

We are mostly failing at improving healthy connection because of technological disruption. Many of us have firsthand, painful knowledge of junior employees languishing for lack of mentorship in a work from home environment. But another core finding in my research tells you one thing for sure: somewhere, somehow, a small subset of us are finding systematic ways to enhance human connection in spite of these barriers. And in a few precious cases, technology is how they’re doing it. Often it’s the very technology that’s at the heart of the trouble for the rest of us!

Our urgent task is to find these systematic successes and learn from them. This is what I did in robotic surgery, and have done in every study since. Most of the time they’ll be partial. Other times they’ll involve unacceptable deviance. Nobody figures out the new best way in one go. The key is they have to be working in very different settings, with different specific technologies, different occupations, kinds of work, industries, geographies, and so on. The more broadly they appear and help people cultivate healthy connection, the more we can trust them.

In my strong opinion - built over more than a decade of research on skill and work involving intelligent machines - this will work if we anchor our search and redesign on what we know about how we build skill and get results. From the point of view of my book, we know we need healthy connection. The book provides a fine-grained checklist on what that looks like in action. We need to take a hard look at our situations and consider the facts: where and how are we handling technology in a way that gets results, but hurts connection? Where are we handling it in ways that helps connection, but compromises results? Where are we handling it in ways that improve both? And, if these are systematic, how can we share these unintentional gains productively with the organizations and billions of working adults who need healthy connection now, more than ever?

This is our urgent quest. It’s going to require hard work. Sacrifice. Investment. Inertia is in favor of productivity over connection, and that’s a death knell for skill, right at a time when we need it most. Let’s get to work.

ps: if now you’re curious about getting a great workout in VR tank battles, here’s a narrated YouTube introduction I recorded about two years ago. Let me know if you want to give it a try - the Tank Nuts will show you a good time!