I've Seen This Movie Twice Before

Hey there,

Roy Amara got it exactly right in the 1960s:

I've lived through this exact pattern twice before – desktop computing in the late 80s/early 90s, and the internet in the late 90s/early 2000s. Here we go again.

The pattern is always the same: Massive predicted impact (correct), ridiculous timeline expectations (wrong), followed by organizational chaos as people try to force unrealistic adoption speeds.

Desktop computing: "This will revolutionize how we work!" (Correct.) "Every office will be paperless by 1995!" (Wrong – took another 15 years, and we're still not there.)

The internet: "This will change everything about commerce and communication!" (Correct.) "By 2000, everyone will be shopping online!" (Wrong – even today in 2025, less than 20% of retail sales happen online.)

AI: "This will transform every job!" (Probably correct.) "It's happening right now!" (Probably wrong.)

What I learned from watching these transformations unfold is that we consistently misjudge the speed while getting the direction right. The breathless "everything changes this quarter" articles. The counter-reaction of "this is all overhyped." The organizational whiplash between urgency and disappointment.

Fortunately, there is familiarity about this AI moment that we can learn from: the same chaos patterns are playing out exactly as they did before.

Organizational resistance from people worried about displacement? Check. I watched secretaries worry that word processors would eliminate their jobs. I watched retail executives in the late 90s dismiss e-commerce because "people need to touch products before buying."

Wild experimentation with no coordination? Check. Remember when every company had seventeen different websites because each department built their own? Now every company has seventeen different AI tools because each team found their own solution.

Competitive scrambling versus strategic patience? Check. The companies that survived desktop and internet transitions weren't the ones that moved fastest – they were the ones that moved most thoughtfully. Remember Pets.com and eToys versus Amazon? Or Lotus 1-2-3 versus Microsoft Excel? Speed without strategy creates expensive lessons.

But here's what's genuinely different this time, and why this transformation feels more unsettling: Previous transformations enhanced what humans were already doing. AI is forcing us to decide what we want it to do instead of us.

Desktop computing made us faster at calculating and creating. The internet made us faster at accessing and sharing information. But AI is doing something that feels distinctly... not human. Synthesizing across domains, recognizing patterns we can't see, generating ideas we wouldn't have had.

That's why the resistance feels different this time. It's not just "this is hard to learn" or "this is overhyped" – it's "this makes me question what my unique value is."

The productivity data backs this up. Ethan Mollick's research shows AI can lead to 20-80% productivity improvements across job types. By contrast, when steam power was put into factories – the technology that created the Industrial Revolution – it improved productivity by 18-22%. We're potentially looking at something bigger than anything we've seen.

And unlike previous transformations that primarily affected blue-collar work, this is the first wave of automation that broadly affects the highest-paid professional workers. Knowledge workers are experiencing what factory workers felt in the 1800s.

First, plan for the reality that AI transformation will take longer than the hype suggests. The organizational changes, the cultural adaptation, the discovery of what actually works – that's measured in years, not quarters. Don't burn out your team on false urgency.

Second, invest in building capabilities, not just experimenting with tools. The companies that thrived in previous transformations didn't just adopt new technologies – they developed systematic approaches to leveraging them. They built processes, trained people, and created sustainable competitive advantages.

Third, recognize that this transformation requires a new kind of judgment. Previous transformations asked us to learn new tools. This one asks us to decide where we want to be enhanced versus where we're comfortable being replaced. That's a fundamentally human decision that requires thoughtful leadership, not just technical adoption.

The familiar part: We'll overestimate the speed and underestimate the impact, just like always.

The unprecedented part: We get to choose what we want the future to look like.

Your assignment this week: Stop asking "How fast should we adopt AI?" Start asking "What work do we want to keep, and what are we comfortable delegating to AI?"

The transformation is inevitable. The timeline is longer than you think. The choices are yours to make.

Break a Pencil,
Michael
www.breakapencil.com

P.S. If you're ready to build systematic AI capabilities instead of just experimenting with tools, my next "Build an AI-Confident Product Team" cohort starts July 21. This is exactly the kind of thoughtful, long-term approach that separates sustainable transformation from expensive theater. [Learn more here.]

P.P.S. I'm working on a follow-up piece about the specific frameworks for making these "enhance vs. replace" decisions. Hit reply if there's a particular aspect of this transformation you're wrestling with – your real challenges make my writing better.