OPEN QUESTIONS / NOT CLAIMS

RESEARCH QUEUE

Questions waiting for sources, experiments, interviews, calculations, code, or somebody stubborn enough to discover that the premise was wrong.

ASTROPHYSICS / QUEUED

Observation before spectacle.

What does a gravitational-wave detector directly measure, and what is reconstructed afterward?
Where do popular black-hole explanations silently switch between analogy and physics?
How do instrument noise, calibration, and model assumptions propagate into astronomical claims?
Which orbital intuitions fail most often when people move from circular examples to real systems?
How should uncertainty be communicated when the observation is indirect but the evidence is strong?

ROBOTICS + CYBERNETICS / QUEUED

Useful contact with the physical world.

Which failure modes dominate low-cost robotic hands under repeated real-world loading?
How much tactile feedback is needed before control performance meaningfully improves?
Where does adaptive intent prediction help myoelectric control, and where does it become frustrating automation?
How should graceful degradation be designed into assistive systems?
Which bench tests best expose pinch, thermal, runaway, force, and electrical hazards before human use?

AI + COGNITION / QUEUED

Measure the disagreement.

Does persistent role differentiation improve error detection compared with prompting one model to “consider alternatives”?
When does adversarial challenge reduce false confidence, and when does it simply create more text?
How should useful disagreement be scored without rewarding contrarian noise?
What changes when memory persists across repeated multi-agent interactions?
Which ablations isolate the value of routing, role persistence, cross-checking, and evidence-aware convergence?

SOFTWARE OPERATIONS / ACTIVE BUILD

Build the workbench, not just the website.

Which Cyberdelia utilities should become the first public releases: archive/provenance tools, network evidence utilities, firmware helpers, forensic inventory tools, or research-pipeline components?
What common package format should every Cyberdelia-built utility ship with for source, binaries, dependencies, tests, examples, hashes, changelog, and reproducible builds?
How should the Freeware Department classify upstream, mirrored, checksum-verified, locally reviewed, legacy/unverified, source-only, and quarantine/lab-only packages?
Which programming, debugging, reverse-engineering, packet-analysis, security-testing, firmware, forensic, privacy, automation, and scientific-computing tools deserve deep technical field guides rather than shallow recommendation lists?
What lab architecture safely separates unknown binaries, hostile samples, firmware images, and security exercises from production systems and sensitive data?
Which network measurements provide the most useful authorized attack-surface baseline with the least noise?
How should software articles preserve durable concepts while versions and interfaces change underneath them?
What compute governor can safely allocate CPU, GPU, memory, storage, and job time between interactive site work, research workloads, AI experiments, and future SETI analysis?

Open Software Operations

ENGINEERING + AEROSPACE / QUEUED

Serviceability is a performance metric.

How early should maintainability constraints enter conceptual design?
Which measurements belong on version one of a field prototype?
How do environment, contamination, vibration, repair access, and parts availability reshape architecture?
Where should safety margin live when weight, cost, and reliability pull in different directions?
What distinguishes documented resource-constrained engineering from improvisation that cannot be repeated?

INTERVIEWS + REFERENCE / QUEUED

Questions that produce durable knowledge.

What did the expert believe before evidence changed their mind?
Which failure taught more than the eventual success?
What constraint do outsiders consistently underestimate?
Which metric looks important from outside the field but becomes noise in practice?
What should a reader inspect for themselves rather than taking the guest's authority on faith?

QUEUE RULE

A question is not a claim wearing future tense.

Items stay in the queue until they have enough evidence to graduate into reference, lab notes, software releases, dispatches, or an interview. If the answer turns out to be boring, the queue survives the disappointment.

Open reference Software workbench Open interview desk