Conceptual illustration of a person building tools with AI in a workshop that opens into sharing and collaboration with other users
AI-generated illustration of personal making leading to sharing and collaboration. It is not a photograph of actual people or a commercial transaction.

What should we learn for the AI era? Keeping up with models and services can help, but an earlier question needs to change: from “Which product should I buy?” to “How do I want to solve the problem I face?” When AI’s lower barrier to implementation meets a person’s own problem, consumers can move beyond selecting finished products and begin redesigning how those products are used.

Read together, UPMI Lab’s accounts of HomeKit integration, smart switches, classroom assistance and presentation production reveal this possibility. Each began as a small project at home, in class or at work. The shared question is larger: must users wait for suppliers to provide the functions they need? If they can build and revise those functions themselves, how far can their role in the market expand?

The person who knows the inconvenience can begin the design

Products tend to address needs common to enough people to support sales. Everyday life is more specific. A bathroom fan should run after the light goes off; a light switched off in an app should also appear off on the wall. Students listening to one lesson understand at different speeds and in different languages. A gap remains between a general feature list and a particular person’s needs.

Users encounter that gap first. The decisive choice in the smart-switch project was not increasing the button count to six. It was requiring the grouped control to recognize even one active light, and preventing the app and physical indicators from disagreeing. Continued, trustworthy use—not merely a successful connection—became the standard of completion.

The classroom assistant follows the same direction. Rather than moving between a translator and a separate chat window, the instructor arranged material-grounded answers and student-specific translated captions around the lesson. This does not establish improved learning outcomes. Its significance is that the tool’s user participated as the designer who defined the problem.

Prosumers produce the next use of a product

A prosumer—someone participating in both consumption and production—need not build an appliance factory or a large platform. Rules that coordinate purchased devices, tools for a particular class and document structures that remain editable are also things people produce. Experience adds a new use to what the market already supplies.

HomeKit integration sought to treat devices divided by manufacturer and apartment infrastructure as one living space. The presentation project needed independently editable text and shapes rather than a polished flat image. Neither was primarily about purchasing more features. They added user-defined standards—“Does this work together in my home?” and “Will the next revision be easy?”—to the supplier-facing questions of connection and initial appearance.

These activities predate AI. Open-source software, user communities and personal making experience provided the foundations. AI agents can shorten the distance between describing a problem and trying an implementation. Users explain aims and constraints; an agent helps translate them into code, settings or documents; people test the result and revise it. Existing knowledge and tools remain essential, but more routes to using them become possible.

Turning a personal solution into a market proposal

A configuration that works at home is not automatically a product. Others need understandable instructions, checks in different environments and a route to repairs when something fails. Solving one’s own inconvenience and supporting someone else’s use are connected but distinct capabilities.

With that additional work, a personal project can become a shared resource or reusable tool, and potentially a service involving teaching, customization or maintenance. Publishing project accounts in the Lab and completing a website with AI illustrate a starting point. AdSense approval permitted advertising on that public space; it did not establish demand or sustained earnings.

The prospect of prosumers moving to the front of the market is not a prediction that everyone becomes an entrepreneur. It means consumers can go beyond reporting dissatisfaction: they can demonstrate working alternatives, improve them with others and put new requirements to companies. Integration support, tools for user creation, verification and maintenance could become important competitive criteria alongside finished products. The possibility is a different relationship in creating value, not the disappearance of companies.

A cycle of noticing problems, building with AI and verifying use, then sharing, supporting and proposing solutions to the market
A personal solution needs sharing, explanation and support as well as implementation and testing to become a market proposal. This editorial diagram explains the argument; it does not measure market conversion. View full-size image ↗

Follow one small problem through, rather than chasing every new tool

One useful habit for the AI era is to stop overlooking recurring inconveniences. “Keep the fan running after the bathroom light goes off, but do not restart it if I switch it off manually” supplies clearer implementation and testing criteria than “Automate this.” Specific observation by someone who knows the setting provides a starting point AI cannot simply supply on their behalf.

A large service need not be the first goal. Set a narrow scope, compare expected with actual behavior and record why changes were made. One successful demonstration differs from reliable use over several days. Recording failure conditions makes experience an asset for the next project and potentially useful to others.

Learn what the work requires and judge where specialist help is needed. Agency does not mean doing everything alone. It means defining the purpose and completion criteria, then choosing what to build and whom to work with. AI capability should extend beyond writing impressive prompts to specifying requirements and evaluating results.

What the market needs is trust, not output volume

Easier creation does not automatically deliver accuracy or maintenance. Wrong answers, interrupted external services and changed device specifications return as operational problems. When others begin using a project, explanation and support take more time. AI costs and dependence on platforms remain too. A lower implementation barrier is not the disappearance of operating responsibilities.

As output grows, why something was made, what has been checked and whether it can be improved become more important. User experience is a starting point; trust that a solution helps others must be earned through repeated verification and feedback. Prosumers may compete through sustained attention to problems they know well, rather than simply releasing more AI-generated material.

We do not all need to establish companies. Nor must we remain consumers who stop when a feature is missing or wait for the next product. We can build, revise, share and, where appropriate, compete with other proposals. Strength in the AI era rests less on how many tools we subscribe to than on turning experience into something useful to others. Prosumers with that capability can move beyond commenting from the sidelines and help determine what the market should make.

Further readingExplore the original making and operating cases at UPMI Lab →
Sources & context

An editorial column based on the UPMI operator’s published Lab accounts and this publication’s related case analyses. Broader prosumer participation is an interpretation and prospect, not an empirically established market shift or earnings claim. UPMI Lab and New Epoch Journal are both operated by UPMI Co., Ltd.

AI assisted with drafting and editing. This article is part of our preview edition. Editorial standards & corrections →