The nextdifficulty.
Solve a difficulty. Create a market.
Reveal the next difficulty.
A breakthrough creates
more than a machine.
It changes what people can do—and what they need next. The invention opens the door. Suppliers, manufacturers, operators, and applications build the market around it.
Enable a market.
Build the inputs, devices, networks, and services that make the capability useful.
Meet the next difficulty.
Scale brings new demands: access, discovery, coordination, trust, and control.
Fifteen shifts.
One recurring question.
What becomes possible—and who makes it work?
Meaning becomes durable
From human memory to reproducible records.
PrehistorySpoken languageCoordinate through shared meaning.
Community coordination, teaching, culture, exchange.
Shared language, customs, gathering places.
Human voice and hearing.
Families, teachers, storytellers, community leaders.
Prehistory · origin undated
c. 3300–3200 BCEWritingLet a record outlive its author.
Administration, law, trade, education, religion.
Scripts, literacy, clay, papyrus, parchment, ink.
Styluses, tablets, scrolls.
Scribes, schools, courts, temples, administrators.
c. 3300–3200 BCE
Antiquity / 1st c. CEManuscripts & codicesOrganize knowledge in a portable volume.
Scholarship, religion, law, education, government.
Writing materials, copyists, libraries.
Scrolls, codices, writing tools.
Scribes, monasteries, courts, universities, libraries.
Manuscripts from antiquity; codex appears in the 1st century CE
7th–8th c. / 1450sPrintingMake copies at a new scale.
Education, religion, science, commerce, public debate.
Presses, type, paper, ink, distribution.
Printing presses, typesetting equipment.
Printers, publishers, booksellers, authors, distributors.
7th–8th c. East Asia · c. 1450s Europe
Distance starts to collapse
Power, mobility, and mass transmission.
1712 / 1769Steam enginePut mechanical power to work.
Mining, manufacturing, trade, passenger travel.
Coal, iron, steel, workshops, railways, ports.
Stationary engines, locomotives, steamships.
Engine makers, railroads, shipping firms, maintenance crews.
1712 · Watt patent, 1769
1830s / 1844TelegraphySend the message without the messenger.
Trade, government, journalism, rail, finance.
Lines, poles, stations, power, operators.
Keys, relays, receivers.
Telegraph companies, news services, railways.
1830s · public message, 1844
1860s / 1876Internal combustionMove beyond fixed transport networks.
Retail, logistics, leisure, commuting.
Steel, oil, fuel, roads, parts.
Cars, trucks, motorcycles.
Automakers, repair shops, fuel suppliers, insurers, dealers.
1860s · four-stroke, 1876
1876TelephoneHold a conversation across distance.
Businesses, households, trade, public services.
Cables, switching, power, rights of way.
Handsets, switching equipment.
Telecom companies, network operators.
1876
1890s / 1920RadioLet one voice reach many.
News, entertainment, education, public information.
Spectrum, transmitters, antennas, studios, electricity.
Transmitters, receivers.
Broadcasters, networks, regulators, advertisers.
1890s · regular broadcasts, 1920
1920s–1930sTelevisionBring moving pictures into the home.
News, entertainment, education, advertising.
Towers, spectrum, satellites, cable, electricity.
TV sets, cameras, broadcast equipment.
Broadcasters, networks, cable and satellite operators.
Late 1920s experiments · mass adoption after WWII
Access becomes continuous
Computing moves into networks, then into pockets.
1940s / 1946ComputerAutomate calculation and data processing.
Manufacturing, finance, aerospace, government, research.
Electricity, electronic components, operating systems, and software; semiconductors underpin later generations.
Early electronic computers used vacuum tubes; transistors and integrated circuits followed. Mainframes, servers, and PCs widened access.
IBM, hardware makers, integrators, service firms.
1940s · ENIAC unveiled, 1946
1969 / 1980s–90sInternetConnect information across networks.
Search, commerce, publishing, education, communication, cloud.
Telecom networks, routers, data centers, protocols.
Routers, servers, PCs, modems.
ISPs, backbone and hosting firms, web platforms.
First ARPANET hosts, 1969 · public growth, 1980s–90s
1990s / 2007+Mobile internetCarry the network with you.
Social media, commerce, navigation, communication, gig work.
Towers, spectrum, backhaul, semiconductors.
Smartphones, mobile network equipment.
1990s · i-mode, 1999 · smartphone surge, 2007+
The medium begins to act
From generating content to coordinating work.
2017 / 2022Generative AI & LLMsGenerate a first draft on demand.
Writing, programming, design, media, customer support.
Cloud, data centers, power, networks, training and inference data.
GPUs, CPUs, servers, accelerators.
Cloud providers, model developers, AI platforms.
Transformer architecture, 2017; ChatGPT public launch, November 2022
2023 surgeAI agentsTurn a request into coordinated action.
Workflow automation, data spaces, agentic commerce.
Cloud and edge, virtual machines, identity, permissions, tools, workflows.
GPUs, CPUs, servers, edge devices.
Agent runtimes and harnesses, model providers, orchestration, data-space operators.
Long research history · LLM agent surge, 2023
Different technologies.
Familiar opportunities.
Choose two shifts. Follow the difficulty each addresses, the market it supports, and the challenge that remains.
Comparison is interpretive. Similar commercial patterns do not imply identical technologies or outcomes.
When the medium acts,
trust becomes infrastructure.
An agent must do more than produce a plausible answer. It needs authorized access, usable context, reliable tools, and a way to check what happened.
“Get this done.”
A goal, a scope, and permission to act.
Coordinate the work.
- Models & tools CAPABILITY
- Identity & permissions AUTHORITY
- Data spaces & context KNOWLEDGE
- Checks & records ACCOUNTABILITY
“Here is the result.”
An action completed within scope, with evidence.
Opportunity thesis: dependable orchestration and governed access can support integration services, subscriptions, usage fees, and transaction businesses. Outcome pricing depends on outcomes that can be measured and attributed.
Look at what the breakthrough makes easy.
Then ask what becomes hard next.
That is where the next useful product, supporting service, or business model may begin.
READING NOTES & PROVENANCE
This essay develops the user’s technology and media table. Historical milestones are linked inside each entry. Entity links use DBpedia resource identifiers. Business models are representative examples across each technology’s development, rather than claims about its first year.
The “next difficulty” questions and the agent opportunity thesis are editorial interpretations. They are not forecasts of returns, universal laws, or claims of a single cause.
Dates distinguish early milestones from later adoption. Inventions developed across places and periods; the sequence is selective and not drawn to a proportional time scale. Spoken language cannot be assigned a precise origin date. Manuscripts precede the codex by millennia.
Generative AI predates 2017; that date marks the Transformer paper. The 2022 marker denotes ChatGPT’s public launch. AI agents predate 2023; that date marks a surge in LLM agent research.
This is a self-contained interactive entry. Links to external entities and evidence require an internet connection.
Trace the next opportunity.
Use the sequence to turn an innovation story into a testable business hypothesis.
- Name the capability
State what the technology lets people do that was difficult, costly, slow, distant, or unavailable before.
- Describe the difficulty
Identify the task or limitation the capability addresses. Keep it distinct from the technology itself.
- Map the complements
List required inputs, infrastructure, hardware, skills, and standards.
- Find the operators
Name the people and organizations that run, maintain, connect, or distribute the system.
- Identify beneficiaries
Specify the activities and groups that gain practical value from the new capability.
- Describe value capture
Explain how providers may earn revenue through sales, access, advertising, subscriptions, usage, or services.
- Ask what becomes difficult next
Look for a new bottleneck created by adoption, scale, reach, trust, control, or coordination.
- Test the opportunity
Treat the resulting business idea as a hypothesis. Check demand, costs, permissions, reliability, and evidence before calling it a market.
Questions readers ask.
Answers grounded in the table and its stated limits.
A difficulty is the task, constraint, or cost that a technology helps people overcome. The table treats it as an analytical lens, not a universal law.
A new capability creates demand for complementary inputs, hardware, operators, services, and applications. Their combination can support a business model.
Infrastructure describes enabling systems and inputs; hardware describes the devices and equipment people use. They often develop together but serve different roles.
No. Each period distinguishes selected early milestones from later adoption where the source data permits. The timeline is selective and not proportional.
Language is a foundational medium for sharing meaning. Its origin is prehistoric and cannot be assigned a precise date.
They mark different changes: recording language, organizing it into portable volumes, and reproducing copies at greater scale.
It follows practical steam engines and later improvements, then connects mechanical power to transport and industrial use.
The entries distinguish messages sent through coded signals from live voice conversations over a communications network.
It spans early electronic computers and later generations of mainframes, servers, and personal computers.
The period distinguishes early network milestones from broader public growth in the 1980s and 1990s.
Network access becomes portable, supporting location-aware services, app distribution, social media, commerce, and gig work.
No. Generative AI is a broader category. Large language models are one type of model used to generate language and related outputs.
The dates identify the Transformer paper in 2017 and ChatGPT's public launch in November 2022; they do not mark the origin of generative AI.
The entry highlights a shift from producing an output to coordinating actions, tools, and workflows. The 2023 marker signals a surge of attention, not the invention of agents.
The editorial thesis points to governed context, identity, permissions, workflow orchestration, and verification as enabling needs. These are opportunity areas, not forecasts.
No. It offers a way to ask where useful products and services may emerge; it does not predict returns or name a certain winner.
Terms for reading the pattern.
A small working vocabulary for the shifts, complements, and markets in this essay.
Difficulty
The task, limitation, or friction that the technology makes easier to overcome.
Business model
The way providers capture value from the capability and its complements.
Infrastructure and inputs
The networks, materials, energy, data, and institutions needed to make a technology useful.
Hardware and devices
The physical equipment that embodies or delivers a capability.
Operators and enablers
Organizations and people that run, connect, maintain, or distribute a system.
Application beneficiaries
People, organizations, and activities that gain new value from a technology.
Complementary assets
Inputs, services, skills, or infrastructure that help a technology reach practical use.
Media evolution
A change in how information is expressed, stored, copied, transmitted, or acted upon.
Adoption
The spread of a technology into recurring use across people and organizations.
Network operator
An organization that provides access to a communications or transport network.
Infrastructure
The durable systems and resources that support repeated use.
Hardware
Physical components used to compute, communicate, produce, or transport.
Cloud computing
Computing resources delivered over a network from shared data-center infrastructure.
Edge computing
Computing performed near the devices or people that produce or use data.
Large language model
A model trained to predict and generate sequences of language and related tokens.
Generative AI
Artificial intelligence used to generate content such as language, code, audio, images, or video.
AI agent
A software system that uses models and tools to pursue a goal through a sequence of actions.
Agent harness
The runtime and controls that connect an agent to tools, identity, data, permissions, and checks.
Orchestration
Coordination of tasks, tools, and participants across a workflow.
Data space
A governed environment for exchanging or using data under shared rules and controls.
Application
A use of a technology to meet a practical need.
Knowledge Graph Explorer 233 nodes · 720 links ¶
Interactive graph visualization derived from the companion RDF. Click nodes to resolve, drag to explore. Graph data embedded from companion RDF at generation time.
The Next Difficulty — RDF Knowledge Graph
KG Settings
Predicate Filters
Node Types
Literal Filter
SPARQL Workbench 3 sample queries ¶
Query this knowledge graph on URIBurner. The editor opens on the canonical SAMPLE entity-type summary (DAV named graph). Pick a recipe, edit freely, then run live or copy.