Established 1974 · an independent research institution
Some discoveries
are not announced.
They are recorded, dated, and revisited until someone is ready to carry them. We study questions that take longer to answer than anyone is usually willing to wait.
- Founded
- 1974
- Stations
- 8
- Directorates
- 6
- Research areas
- 10
The premise
Most questions are abandoned for the same reason.
Not because they are forbidden, and not because they are beyond us. Because answering them takes longer than the institution asking is able to wait.
A university cannot pursue a result that arrives after the grant. A company cannot pursue one that arrives after the quarter. SENTINEL exists to hold the questions that fall through that gap — funded from principal, never from contract, so that no external party can end a line of inquiry by declining to renew it.
We publish when the Disclosure Board judges publication more useful than restraint. That judgement is recorded, dated, and revisited. Where we have withheld something, the Archive says so rather than leaving a silence.
The Reversibility Principle
“We do not deploy what we cannot withdraw. Any system we build must be able to be taken back out of the world by the people who put it there.”
It is the condition our founder attached to the endowment in 1974, and it has cost us at least four programmes that worked. We consider that the correct trade.
Current institutional state
What the public index admits
Index figures derived from the public catalogue at build time. Not live telemetry. Last synchronized 2026-09-05.
- Directorates
- 06
- Active stations
- 04
- Open cases
- 03
- Archival records
- 27
- Withheld
- 06
- Unresolved links
- 04
04 REFERENCES REMAIN UNRESOLVED. · inspect relationship index
Open investigations
What remains unresolved
Directorates I–VI
What we work on
6 directorates · 10 research areas
- Directorate I
Systems & Continuity
Quantum systems, distributed computing, and networking that assumes disconnection.
3 areas · e.g. Quantum Systems
- Directorate II
Alignment
Making capable systems legible to the people accountable for them.
1 area · e.g. AI Alignment
- Directorate III
Signals & Cryptography
Weak-signal recovery, secure communications, and applied cryptography.
3 areas · e.g. Signal Intelligence
- Directorate IV
Cyber Defence
Defensive engineering, detection, and disclosure done properly.
1 area · e.g. Cyber Defence
- Directorate V
Autonomous Systems
Machines that operate unattended, and stop when they should.
1 area · e.g. Autonomous Systems
- Directorate VI
Human Factors
How people behave around instruments, alarms, and procedure — with standing review authority over the others.
1 area · e.g. Human Factors
Stations
Where the work is kept
Discovery
Read carefully. Follow the numbers.
The catalogue is cross-linked on purpose. A document number, a station code, or a recurring interval may be more than filing. Some references resolve only as far as the public index allows.
Search the catalogue → · Relationship index → · Observatory →
We are not recruiting, exactly.
The Continuity Fellowship funds four to six people a year whose questions are too slow for conventional funding. No deliverable is required. We ask only that negative findings be written down.