Skip to content
All writing

Our View of Innovation

What separates a real revolution from rebranded software? Three tests, and a timeline.

Hongkai He 8 min read
  • #innovation
  • #technology
  • #framework

A technology only merits the term revolutionary if it changes one of three things — and the more dimensions it changes at once, the further the change cascades.

01 Cost structure of essential production inputs Carnegie steelmaking → Building materials $↓ Petroleum / ICE → Energy & power $↓ Computer → Calculation $↓ Internet → Communication $↓ Blockchain → Trust-building $↓ Note: human time is the limiting reagent in most cost structures. Most "cost-down" innovations reduce to "less demand for human hours." 02 Organizational criteria and principles ICE → Assembly line Automotive → Work-life separation Internet → Remote collab, crowdsourcing AI → Human-machine collaboration 03 Perceived spaces for innovation Automation Software, internet "Trustless" networks Generative AI

It either changes the cost structure of essential production inputs — Carnegie’s process for steelmaking, the internal combustion engine for energy, the computer for calculation, the internet for communication, blockchain for trust. Or it changes how we organize productive work — the assembly line, the urban office, remote collaboration, human-machine collaboration. Or it expands the perceived space of what’s possible — programmable software, the network as platform, generative AI as a general-purpose creative engine.

Most things sold as “innovation” check at most one of these boxes weakly. The interesting ones change two or three at once, usually because the underlying shift cascades from one dimension into the others.

How a revolution unfolds

A revolutionary technology doesn’t arrive whole. It moves through stages, and the kind of player who wins changes with each stage.

01 Foundation Research-driven Government, labs, large corporations 02 Building Block Engineer-driven SMBs, venture capital 03 Application Business-driven Operators with industry depth, SMBs, VC Where today's waves currently sit (early → mature) → PC & Internet Internet of Things Machine Learning Synthetic Biology Gene Editing Next-gen UI Quantum — Foundation — — Building Block — — Application —

The Foundation stage is research-driven. The economics aren’t commercial yet; the breakthroughs are. Government labs, universities, and a handful of large corporations spending on basic R&D do most of the work. Bell Labs and Fairchild Semiconductor for the transistor; ARPA for early networking; CERN for the web’s foundational protocols.

The Building Block stage is engineer-driven. Smaller companies — and the venture capital that funds them — wrap commercial form factors around the foundational science. Microsoft, Intel, IBM’s commercial computing era; Cisco and the early-internet infrastructure layer.

The Application stage is business-driven. Operators with deep industry knowledge apply the new building blocks to specific markets. Amazon and Booking on commerce; Salesforce and Tencent on platforms; the entire SaaS layer downstream.

Knowing which stage a wave is in tells you which player has the right risk profile to win — and which kind of mistake to expect. Confusing stages is one of the most expensive errors in venture and corporate strategy alike: chasing Application returns with Foundation-stage capital, or running a Building-Block business with Application-stage urgency.

The same diagnosis applies to today’s waves. PC & Internet is deep into mature Application. Internet of Things and Machine Learning are well into Application — real products shipping at scale. Synthetic Biology is at the late-Building-Block / early-Application boundary; the platforms are emerging. Gene Editing is in early Building Block, just past the Foundation phase — CRISPR is real and the first wave of platform companies is being built. Next-gen UI (NLP, AR, BCI) is still in late-Foundation. Quantum Computing is in mid-Foundation; the commercial apps are years away.

Reference: Perez, Carlota. “Technological revolutions and techno-economic paradigms.” Cambridge Journal of Economics 34.1 (2010): 185–202.