When (AI)lectricity Disappeared
There was a time when electricity was the headline.
Newspapers wrote about it as a miracle. Investors chased it. Entrepreneurs built companies around it. Engineers debated AC versus DC as if the future of civilization depended on it. In some sense, it did.
Factories reorganized around electric motors. Cities extended working hours beyond daylight. Entire industries restructured. Electricity was not incremental improvement. It was a new layer of capability.
And then something strange happened.
Electricity disappeared.
Not physically. Economically. Culturally. Strategically.
Today, no serious company claims competitive advantage because it “uses electricity.” No board approves a budget because the firm has successfully “adopted power.” Electricity is assumed. It is infrastructure. It is a platform upon which value is created, not the value itself.
The transformation was real. The conversation faded.
We are watching the same pattern unfold with AI.
The Illusion of the Tool
In the early phases of any general-purpose technology, people mistake it for a tool.
Steam power was a tool. Water wheels were tools. Horses were tools. Electricity looked like a tool. Early adopters plugged it into existing processes. They replaced steam engines with electric motors without redesigning factories. Productivity gains were marginal.
The real shift came later.
Electricity allowed decentralization of power inside factories. Machines no longer needed to be physically arranged around a central shaft. Layouts changed. Workflows changed. Management changed. Entire industries reorganized around the flexibility electric power enabled.
The technology did not simply improve the old model. It made new models possible.
AI is at the same stage many factories were in 1905: swapping steam for electricity without rethinking the factory.
“We use AI for customer support.”
“We integrated AI into our workflow.”
“We automated parts of marketing.”
This is the electric motor bolted onto a steam factory.
The deeper shift is not automation. It is cognition as infrastructure.
From Power to Cognitive Power
Electricity provided physical power everywhere. Cheap, reliable, distributed energy.
AI provides cognitive power everywhere. Cheap, scalable, distributed intelligence.
Not intelligence in the human sense of wisdom or judgment. But pattern recognition. Prediction. Language generation. Optimization. Classification. Simulation.
For most of history, cognitive labor was scarce. Skilled analysis required trained humans. Scaling thinking required hiring.
Now cognition is becoming ambient.
When a capability becomes ambient, its economics change.
Electricity reduced the marginal cost of energy close to zero in relative terms. That enabled refrigeration, computing, telecommunications, and global manufacturing networks.
AI reduces the marginal cost of certain forms of cognition close to zero. Drafting. Summarizing. Translating. Coding. Designing variations. Running scenarios. Generating alternatives.
The question is no longer: “Do we use AI?”
That question will sound as outdated as “Do you use electricity?”
The real question becomes:
What value can you create because cognitive power is ubiquitously available?
The Strategic Misunderstanding
Calling it an “AI strategy” already reveals confusion.
No company ever had a serious “electricity strategy.”
They had manufacturing strategies. Distribution strategies. Cost strategies. Scale strategies. Electricity was embedded inside them.
If your competitive advantage depends on “using AI,” you do not have an advantage. You have access to a commodity.
The strategic work lies elsewhere:
Which constraints disappear because cognition is abundant?
Which bottlenecks remain scarce?
Which business models become viable only when thinking is cheap?
When electricity spread, illumination was not the final value. It enabled nightlife economies, refrigerated supply chains, elevators that allowed skyscrapers, data centers that enabled the internet.
Electricity was not the end product. It was the substrate.
AI is becoming substrate.
The Plug-in Fallacy
Many companies treat AI as a plug.
“We have integrated AI into our CRM.”
“We use AI for forecasting.”
“We deployed AI chatbots.”
This framing assumes AI is a feature layer.
But general-purpose technologies do not stay features. They reshape system architecture.
Electric motors did not just replace steam engines. They dissolved the architectural constraints that steam imposed. Central shafts disappeared. Flexible layouts emerged.
AI dissolves cognitive bottlenecks.
Previously:
Analysis required analysts.
Prototyping required designers.
Coding required developers.
Research required time-consuming synthesis.
Now these tasks can be generated, iterated, and stress-tested instantly.
That changes cycle times. It changes experimentation economics. It changes decision quality. It changes the boundary between exploration and execution.
If cognitive iteration is cheap, then the bottleneck shifts.
The constraint moves from “can we generate options?” to “can we judge them well?”
Judgment becomes the scarce resource.
The Real Scarcity
Electricity did not eliminate scarcity. It shifted it.
Once power was abundant, coordination became critical. Supply chains expanded. Logistics complexity grew. Organizational design mattered more.
Similarly, AI does not eliminate uncertainty. It shifts it.
Paid membership includes
Weekly deep-dive (paid-only)
Full archive
Comment access
Founding Decision Circle ($499/year, 10 seats) includes
1 structured decision submission per month
Written response within 5 working days
Priority queue
