🧱 Building AI Brick by Brick
How my team culture is evolving to thrive in high ambiguity situations
👋 Adam here. Before we get to this week’s post, I’ll plug that I’m doing a fireside chat this Tuesday at 10AM Pacific. Leadership coach Nitin Garg and I will be discussing how to stay human as we make product and many of the themes of Mind The Beet will be on display: burnout, boredom, purpose, balance, intensity. Join us for a great conversation.
Imagine I am pitching you on my new business idea. I’m trying to get people to visit my mountain property.
It requires me to build towers that stretch thirty feet into the air to carry people in chairs where it is easy for people to fall off.
Once they get to the top, they travel under their own control at rapid speed - so much so that over the course of the day, enough people will sustain injuries that I’ll have to have a team available to go collect immobile people throughout my property.
The natural landscape isn’t suitable for me, so every night I’ll have to use giant machines to get it ready for the next day.
I’m subject to the whims of the weather and even in the best case, I’m only open five months of the year.
Oh, and I require my customers to wear the most uncomfortable footwear of their lives.
Pretty crazy, right? No one in their right mind would think this is a sane business idea - and yet there are almost twenty million people in the United States that went skiing last year.
The only reason you believe me that this is a viable idea is because of the existence proof. The industry couldn’t be exactly foreseen ahead of time, but instead it was built up over the years, with creative solutions to each of the unique risks and challenges that cropped up.
Why Am I Telling You This?
As our industry embraces the paradigm shift of AI, I’ve been looking for the right non-tech examples to explain to my team how to operate differently in the early stage of a new era.
Like many of us, for the past few years, I’ve led a team that was capturing value from a pretty well-defined flywheel born out of the transition to the cloud. Ambiguity was (relatively) low, a product culture of stability and predictability was an advantage, and there was high alignment between customer feedback and the right backlogs.
Now we find ourselves in a new situation:
There is no existence proof for the endgame. While we have strong intuition about the scenarios that AI will evolve to help with, we have no straightforward way of describing the end state. What will it all look like after we’ve gone through the dozens of incremental steps to get there? We are staring at a snowy hill with no idea of all the problems we will have to tackle to invent a modern ski resort. We need both vision and humility to tackle this.
“Anyone who is sure about anything in generative AI at this stage is a fool - perhaps a clever fool.”
It’s easy to focus on downsides. My description of skiing above focused on pitfalls and negatives - I didn’t talk about the thrill of the epic powder days, the kids who learn grit and agility on ski teams, or even the quotidian opportunity of ski lodge food service profits. Ambiguity breeds a focus on risks - it’s natural (and essential to understanding externalities and unintended consequences) yet an overfocus on it can be harmful to ideation and product making.
“Now is a time to ask What If? instead of Why Not?”
There are many more failure paths than success paths. We have flipped from “most incremental investments will pay off quickly” to “all plans inherently have more risk and the hit rate goes down.” Tech risk, design risk, business risk - all the classic forms of product discovery risks are on the rise.
“AI is anything that doesn’t work yet.”
Time To Act Different
So if the endgame isn’t clear, many roads lead the failure, and we naturally gravitate towards articulating risk - what’s the right way to lead?
Resilience in a business context like this is what it means to have a “tech” culture. Being in tech is less about whether a company writes software or not and more about their ability to pivot to embrace new ambiguous paradigms.
Like many, I believe that the future must be built brick by brick from smaller bets that cut through the ambiguity. It’s intentional that just last month, I reviewed The Geek Way, a book that frames leveraging science, openness, and speed. Here’s a few other ways I’m leading through the change:
Drive with more intensity. I’m spending more of my time with my LT setting ambitious expectations on outcomes and driving to a culture of ever higher performance. The beginning of something new naturally leads to a sense of needing “this team to capture this moment,” and it establishes a norm of problem solving instead of problem admiring.
Extreme awareness of context. I’ve upped the amount of “state of play” communications I’m sending to my leadership team. Staying informed is even more important when things are changing rapidly and it’s an easy way to create more safety among the team. I’ve always prided myself on being the glue.
Focus on opportunity cost. The sunk cost fallacy is top of mind for me as a leader. In particular, I’m asking more “What’s the opportunity cost of things we could be doing instead?” I’m framing how in hindsight it’ll look like a bigger mistake to leave folks working on an OK opportunity at the expense of rapidly tackling a great opportunity.
Resolve disputes with real code. I’ve found that one of the best way to reduce ambiguity is more hands-on time with product. Code wins. Using tech to solve real problems in your own life is clarifying. It’s been especially useful to clarity new conceptual models and UX patterns.
Resist the Urge to Promise Stability
As a leader, I often ask myself, “What would have happened anyway if I wasn’t around? What unique value do I provide over the team self-organizing?”
A couple years ago, I might have answered that question with a focus on providing stability, consistency, and nurturing a safe and stable place for people to learn and thrive at making product.
Today that’s evolving. I’m here to help the team embrace change and thrive in an environment where we are not afraid of evolving to meet new expectations. We’re proud that we are on the vanguard of the unknown. One leader on my team had shirts made for her team that say, “Sunny with a chance of chaos.” Love it.
I like the analogy. It might be worth stretching it a little to cover change velocity. Building / adopting AI brick by brick is very much like learning to ski. Everyone wants you to go fast but you should really start on the nursery slopes before attempting the diamond run. There is also that element of risk adversity that practice and confidence are needed to overcome.