Intelligence research produces many kinds of evidence: psychometric studies, longitudinal studies, clinical observations, educational assessments, occupational data, public datasets, intervention studies, task analyses, open science records, and lived-experience signals.

Too often, that evidence is scattered across journals, proprietary tools, institutional reports, private datasets, disciplinary silos, and inaccessible language.

The Observatory would help organise what is known, what is uncertain, what is being tested, and what should not yet be claimed.

It would allow readers to see the difference between established findings, emerging evidence, proposed frameworks, active studies, validated instruments, restricted materials, and open questions.

The Observatory may track evidence about relatively stable cognitive profiles alongside changing cognitive states, including fatigue, stress, grief, health, illness, medication, readiness and motivation, as well as development, ageing, learning, support use, AI assistance, task demands, environmental conditions, repeated performance episodes and outcomes.

At the person level, it may help research understand variation across time: how sleep, illness, recovery, support, learning, or environment changes what someone can do.

At the population level, it may help identify patterns: which tasks create widespread burden, which supports appear useful, which systems exclude people, and which claims need stronger evidence.

At the research level, it may track the status of instruments, protocols, datasets, preprints, peer-reviewed publications, replications, corrections, and open questions.

The Observatory should make public knowledge easier to inspect without making sensitive information careless.

Open outputs may include evidence summaries, versioned definitions, public dashboards, study registrations, non-sensitive aggregate findings, literature maps, instrument status labels, and links to OSF or Zenodo records.

Restricted materials may include identifiable data, sensitive participant information, test items, scoring keys, proprietary measure mappings, practitioner-only materials, and unfinished instruments that could be misused if released too early.

The point is not to make everything public. The point is to make the reasoning public enough that people can understand what is known, challenge weak claims, and contribute responsibly.

The Observatory receives evidence from the broader research infrastructure.

The Human Task Atlas contributes structured knowledge about the tasks people face.

The Environmental Registry contributes knowledge about the conditions that shape performance.

The Performance Lab contributes repeated evidence about what happens when people act under specific conditions.

Contributors across health, education, work, daily living, public systems, technology, research, and lived experience help identify what should be observed and what claims need testing.

The Observatory should eventually support public evidence briefs, research maps, validation status pages, glossary records, page-specific reference updates, consortium reports, open questions, correction logs, and summaries of what changes as new evidence appears.

This would allow How Intelligence Works to remain current without pretending that developing research is already settled.

The Observatory is how the project can be open, cumulative, and corrigible.