Chuhee LeeIndependent research
All publications
Chuhee Lee · Independent research · Method

How this research is made

How topics are chosen, where facts come from, how AI is used, how claims are labeled and scored, and how reports are refreshed and corrected.

01Independent and personal

These are personal research projects. The views are the author’s own and do not represent those of any employer or organization. No company covered sponsors, reviews or endorses this work, and product names and trademarks belong to their owners. Reports make observations, not company-specific recommendations.

02How topics are chosen

The site follows one thesis: time and attention on the move. People move every day, to work, to visit, to travel, and spend that time reading, watching, learning, working or resting. Mobility, travel and what people do on the way are treated as one connected space.

Topics sit on a map of four layers across four domains (Shared, Mobility, Travel, Entertainment):

  • Forces: macro trends over 10–15 years, shared by all domains.
  • Shifts: technology and consumer changes over 1–5 years.
  • Signals: benchmarks of what is actually shipping.
  • Opportunities: a product lens on what the evidence opens up.

Six lenses on the person on the move help find topics: time, attention, state of mind, social context, space and transitions. Topic selection is the author’s.

03Where facts come from

Every factual claim is sourced, and every report lists its references. Sources are used in this order of preference:

  1. Official and primary sources: company releases and filings, government and statistical agencies, regulators.
  2. Recognized studies: research firms, academic work and established surveys.
  3. Specialist trade press.

A source that reports another organization’s figures is marked as secondary. Each reference carries its publication or as-of date, and dated anchors are flagged as dated. Where sources disagree, the report says so rather than picking one silently. Figures that could not be confirmed against a primary source are dropped and the drop is recorded in the report’s flags.

04How AI is used

Research, source verification and drafting are carried out with several AI models. Claude (Anthropic) is the main research tool; Gemini (Google) and ChatGPT (OpenAI) are used for smaller supporting references and cross-checks. Each report names the models used for it. Topic selection, scores and conclusions are the author’s.

AI-generated images, audio and video may appear on the site. Each carries a visible label where it appears, and the request behind each one is recorded in the site’s prompt log.

05Fact, interpretation, hypothesis

Summaries label each takeaway by the status of its main claim:

  • FactStated by the cited source, reported as the source states it, with its date.
  • InterpretationThe author’s reading of the cited evidence: scores, frameworks, rankings and what a finding means.
  • HypothesisExpected from the evidence but not yet observed. It is a claim to test at the next refresh.

Figures inside any takeaway are always sourced facts; the label describes the conclusion drawn from them. The labels are used in each report’s “The report in 60 seconds” summary.

06Scores and scales

Scores are the author’s analytical judgment from the cited evidence, not measured results or sourced figures. Each report states its own scale:

ReportWhat is scoredScale
Mega Trends Outlook 2040Each trend on impact, certainty, speed and cross-industry breadth. The criteria were set for this report, not taken from a validated method.1–5 each, total out of 20
In-Car AI Assistant BenchmarkEach assistant on seven capability dimensions, plus who controls the stack. Calibrated across the set; no vehicle was driven or instrumented; pre-launch entries scored on announced specification.1–5 per dimension
Mobility Product Outlook 2026–2031Fifteen technologies on daily-life impact and mobility relevance, with timing and certainty.1–5; timing Near (by 2030), Mid (2031–2036) or Far

07Refreshes and corrections

Each report and benchmark shows its data-as-of date and its next scheduled refresh (one-time briefs are not refreshed). Long-range outlooks are refreshed at least once a year; the benchmark is refreshed quarterly. A refresh re-checks the report against its sources, and every change is recorded in the report’s change log. Updates are reviewed by the author before they go live.

Corrections are welcome via LinkedIn. To follow new editions, use the RSS feed.

08License and citation

Unless stated otherwise, reports are licensed under CC BY-NC 4.0: share and adapt for non-commercial purposes with credit. Each report gives a suggested citation in the form: Chuhee Lee, Report title, edition (data as of date), URL, CC BY-NC 4.0.