A connected evidence frontier
Sources are connected to the exact capability question they can inform, while 99 identical multi-source bundle cases remain visibly flagged for scrutiny.
Titan Research Radar · living frontier map
The edge of human possibility.
Can people learn to navigate with sound, work with an extra robotic limb, communicate from a lucid dream or build a new directional sense? Titan follows extraordinary claims from first observation to reliable real-world capability—and shows exactly where the evidence stops.
A much larger frontier · mapped internally
Thirty-two families. Two hundred and fifty-six proposed powers—mapped, not endorsed.
The all-family programme now maps 256 proposed powers across 32 families—from sensory learning and disability-led adaptation to human–AI, collective and institutional capability. Each proposal has its own source-to-axis matrix and six separate evidence questions, so an intriguing signal cannot quietly become a promise.
Questions prepared, people unselected
The packets remain unsent. Seven review lanes and 163 required responses are staged across four waves without naming or contacting a reviewer.
Readiness is not endorsement
External review has not started. This work creates no review result, decision, payment, promotion, training permission or tutorial.
How Titan moves · five signals, not five ranks
A power can evolve without being promoted.
Better evidence, a safer boundary and a clearer public explanation are real progress. Titan records them separately from guide permission and independent validation, so no interesting discovery has to be either frozen or oversold.
Several signals can be true at once.
An internally complete record may still require external confirmation and forbid a guide.
- Research state
Internal evidence complete
Titan’s bounded evidence and challenge pass is finished.
256 / 256 mapped powers - Publication permission
255 ready · 1 held
PWR-078 permits a status-and-safety correction only; it has no public-explainer or operational authority.
255 explainers · 1 publication hold - Guide permission
Non-prescriptive guide frame ready
Bounded non-prescriptive guidance is eligible; only four G3 explorations are currently released.
22 scopes · 4 released G3 - Confirmation gate
External confirmation required
Specialist or affected-community authority is still necessary.
216 / 256 powers - Validation state
Independently validated
Reserved for real outside confirmation of the exact claim.
0 currently
No single progress bar: explainer-ready does not mean guide-ready; guide-ready does not mean independently validated; internal completion does not erase an external gate. The frozen Core and promotion count remain unchanged.
See the full evolution and closure model →Actionability is a permission—not an evidence grade
How far may a public guide go?
A fascinating phenomenon may deserve explanation while remaining completely unsuitable for self-training. These labels govern what Titan may publish; they do not rank the person, ability or research.
Hard boundary: no Candidate, Observatory, internal dossier or planned research wave can become public training merely because it is exciting, popular or technically measurable.
Nineteen public-safe windows into the frontier
The Research Radar.
Each card keeps the evidence state, tutorial permission and next research gate separate. The Radar is a public presentation of existing records—not a second register, a popularity chart or a claim of completeness.
Showing all 19 research windows.
Publication hold
PWR-078 status correction
PWR-078 remains held from explanatory publication and current operational use. Titan does not authorise this material as training, navigation or mobility instruction.
It does not establish safe independent navigation and does not advise replacing a cane, guide dog, live assistance or qualified orientation-and-mobility support.
Read the correction →No current module route
Adult pitch-class naming
This bounded research record remains separate from current operational actionability while its former parent module is conceptual/read-only.
Evidence boundary
- Known
- A bounded evidence record exists.
- Not established
- A current Titan assessment, session, retest or training route.
- Next gate
- An explicit non-echolocation task identity and governance decision.
No current operational route is authorised.
Operational module
Human–AI reliance calibration
The useful unit is person + model + interface + task + governance—not the person or AI in isolation.
Evidence boundary
- Known
- Human-only, AI-only and combined work can be compared on quality, time, defects and verification burden.
- Not established
- That adding AI necessarily improves performance or demonstrates judgement.
- Next gate
- Adversarial error detection, delayed fallback and model-drift testing.
Use non-sensitive, reversible tasks only.
Operational module
Assistive technology and person–task fit
Capability may belong to the configured person–support–device–environment system.
Evidence boundary
- Known
- Real-use comparisons can examine task benefit against setup, symptoms, privacy and abandonment burden.
- Not established
- That availability or novelty proves fit for a particular person.
- Next gate
- Long-term fit, fallback and accessibility evidence.
Stopping a poor-fit tool is a valid result.
Research Observatory
Human magnetoreception
Reported neural responses to geomagnetic rotation raise a fascinating existence question.
Evidence boundary
- Known
- Controlled work has reported neural responses.
- Not established
- A conscious, useful and trainable human compass.
- Next gate
- Replication plus conscious discrimination and functional-navigation evidence.
Never rely on an unproven compass for navigation.
Research Observatory
Wearable directional cues
A device can add a continuous directional cue that a user may learn to incorporate.
Evidence boundary
- Known
- Controlled work improved cardinal-direction and survey-knowledge performance; wayfinding effects were smaller.
- Not established
- Device-off retention, broad transfer or biological magnetoreception.
- Next gate
- Device-off retest, failure recovery and accessible-equivalent routes.
Never replace ordinary navigation aids.
Research Observatory
Two-way lucid-dream interaction
Researchers have exchanged simple information with some participants during verified lucid dreams.
Evidence boundary
- Known
- Two-way experimental communication has been demonstrated under laboratory conditions.
- Not established
- A reliable, safe induction method or broad functional value.
- Next gate
- Replication, sleep-cost measurement and adverse-event monitoring.
Study-design explanation only; no induction tutorial.
Trait / screening only
Super-recognition
Exceptional face-recognition performance can be measured as an unusual trait.
Evidence boundary
- Known
- High performance is measurable and relatively stable in selected people.
- Not established
- A reliable method that trains the general population to this level.
- Next gate
- Trainability, realistic transfer and false-identification burden.
No identity, surveillance, employment or punitive inference.
Research / regulated technology
Supernumerary robotic limbs and BCIs
People can learn to control and sometimes embody an extra virtual or robotic effector.
Evidence boundary
- Known
- Small studies demonstrate learning in defined person–decoder–device configurations.
- Not established
- Safe, low-burden everyday augmentation or transfer across devices.
- Next gate
- Long-term burden, decoder drift and failure recovery.
No DIY neural-interface or robotic-control instructions.
Phase‑2 packet prepared · unissued
Tendon adaptation under progressive loading
A bounded dossier asks how tendon properties change under a declared loading context.
Evidence boundary
- Known
- The internal evidence and hostile-review pass is complete, and a question-scoped reviewer packet is prepared.
- Not established
- Invulnerability, a public loading programme, external review or promotion.
- Next gate
- External biomechanics, safety, accessibility, retention and transfer review.
No hard-object conditioning, public load or progression.
Phase‑2 packet prepared · unissued
Quiet locomotion and gait acoustics
A measurement-first dossier separates gait acoustics from dramatic stealth claims.
Evidence boundary
- Known
- The internal evidence and hostile-review pass is complete, and a question-scoped reviewer packet is prepared; acute mechanical modulation is plausible.
- Not established
- Durable quiet-walking acquisition, broad transfer or functional stealth.
- Next gate
- Direct calibrated acoustics, delayed retention and cross-surface transfer.
Measurement explainer only; no hazardous stealth drill.
Phase‑2 packet prepared · unissued
Load-carriage endurance and recovery
A dossier examines performance and recovery under a defined load, duration and operating context.
Evidence boundary
- Known
- Experienced populations demonstrate task-specific performance; trainability evidence is partly supportive; a question-scoped reviewer packet is prepared.
- Not established
- Broad transfer, direct retention or safe self-prescription.
- Next gate
- Occupational, clinical, heat, foot, nerve and recovery review.
No public load, distance, dose or progression.
Specialist Phase‑2 packet · unissued
Cervical strength, endurance and head control
A specialist dossier separates neck performance from impact-tolerance mythology.
Evidence boundary
- Known
- The internal evidence and hostile-review pass is complete for a non-impact question, and a specialist reviewer packet is prepared.
- Not established
- An indestructible neck, impact tolerance or unsupervised trainability.
- Next gate
- Clinical review of population, bounded task, contraindications, retention and transfer.
No neck bridges, aggressive loading or impact conditioning.
Internal Wave B
Culturally situated navigation and ecological knowledge
Wave B examines capabilities developed within particular environments without extracting them into generic hacks.
Governance boundary
- Known
- Internal desk research, crosswalk and hostile challenge are complete.
- Not established
- Community permission, public teaching authority or ownership of the knowledge.
- Next gate
- Community-conferred authority, consent, attribution and benefit-sharing.
No culturally situated practice becomes a tutorial without community authority.
Internal Wave C
Disability-led alternative capability pathways
Adaptation, assistance, AAC and alternative routes are valid configurations—not lesser versions of a norm.
Governance boundary
- Known
- Internal research, crosswalk and hostile challenge are complete.
- Not established
- Affected-participant approval, external accessibility review or one route for everyone.
- Next gate
- Disability-led review, AAC authorship, consent and fallback evaluation.
No deficit framing or misattribution of configured performance.
Internal Wave E
Collective and institutional feedback closure
Can reports and reviews become verified corrective action rather than administrative theatre?
Governance boundary
- Known
- Internal research, hostile challenge and bounded reviewer packets exist.
- Not established
- External organisational review, pilot outcomes or a global institution score.
- Next gate
- Clinical, privacy, organisational and affected-person review.
Reporting volume or a meeting is not proof of learning.
Research / clinical only
Consumer neurostimulation for cognitive enhancement
Titan preserves the research question while refusing to convert device novelty into a DIY protocol.
Why Titan refuses this route
- Known
- Reported effects vary by protocol, population and outcome.
- Not established
- Reliable general cognitive enhancement from unsupervised consumer use.
- Next gate
- Sham-controlled replication, adverse-event evidence and qualified governance.
No DIY stimulation instructions, montage, dose or device recommendation.
Research / clinician gate
Voluntary autonomic or immune modulation
Measured physiological changes in intensive controlled practice do not create a safe public health intervention.
Why Titan refuses this route
- Known
- Small controlled studies report changes after intensive breathing, meditation and cold training.
- Not established
- Safe general self-prescription or broad health benefit.
- Next gate
- Replication, mechanism separation, adverse events and clinical oversight.
No hyperventilation, prolonged retention or extreme cold protocol.
Planned / unexecuted
Extreme, lifespan and creative horizons
Future waves will examine environmental extremes, lifespan development, creativity and frontier voluntary control.
Research boundary
- Known
- The scopes and governance questions are defined in the public roadmap.
- Not established
- Research completion, findings, external review, protocols or promotions.
- Next gate
- Execute each wave's discovery, evidence, challenge, authority and review sequence.
A roadmap position is not evidence or permission to experiment.
No matching research window. Try a broader term or return to All.
Dramatic language in · testable question out
Myth to Metric.
Titan does not have to mock an extraordinary claim or accept it whole. It can preserve the interesting question, replace the superlative with a measurable task and state what the result would still fail to prove.
Current authority: explanatory publication and operational use remain held.
Boundary: no training, navigation, mobility, assessment, validation or promotion route.
Measure: balanced twelve-class accuracy, delayed retest and untrained timbre.
Refuse: an effortless universal trait.
Measure: conscious direction discrimination and functional navigation.
Refuse: inferring a usable compass from a neural response.
Measure: accuracy, latency, errors and cognitive load with one frozen device configuration.
Refuse: safe everyday augmentation.
Measure: human-only, AI-only and combined quality, defects, time and verification burden.
Refuse: equating fluent output with human judgement.
Measure: protocol-specific adaptation, function, burden, retention and transfer.
Refuse: invulnerability or impact conditioning.
Measure: calibrated gait acoustics under declared surface, footwear and task.
Refuse: stealth or universal transfer.
Measure: specialist-assessed non-impact strength, endurance and head control.
Refuse: impact tolerance or DIY loading.
Current released set · four reversible G3 explorations
Try the method—not the hype.
These four—and only these four—are Titan’s currently released interactive G3 explorations. The wider map has 22 records eligible for a bounded, non-prescriptive informational guide frame; eligibility is not release. These four are not training programmes and not assessments; they are also not new validations. They use low-stakes tasks to make a process observable, promise no improvement, issue no capability level and store no data on this page.
Confidence-tagged quizzingSee whether confidence matches correctness on one low-stakes topic.
What this can show: task-specific calibration and recurring error patterns. It is not a global self-awareness, intelligence or character score.
- Choose a low-stakes topic with objectively scorable answers.
- Prepare a balanced set of at least twenty items before beginning.
- Answer before feedback and attach a fixed confidence band.
- Compare confidence with accuracy and classify recurring errors.
- After at least seven days, repeat on a new matched set without the original notes.
- Use the same method once on a related but untrained topic.
Direct: accuracy and confidence–accuracy gap.
Retention: new matched set after seven days.
Transfer: related untrained topic.
Burden: time, rumination and indecision.
Stop: if review becomes compulsive, paralysing or repeatedly reopens settled answers. Never use it to replace qualified high-stakes review.
Cue-anchored daily taskTest whether one observable cue supports one low-stakes future intention.
What this can show: completion after one declared cue and the burden created by reminders. It is not a general memory upgrade.
- Choose one low-stakes daily task.
- Link it to one observable event and define a completion window.
- Define the smallest next action and one fallback cue.
- For fourteen days, record cue occurrence, completion, fallback use and ignored alerts.
- Continue for another fourteen days without redundant prompts.
- Try the method once with a different low-stakes intention.
Direct: completion inside the declared window.
Retention: a second fourteen-day period.
Transfer: a different intention and cue.
Burden: ignored alerts, anxiety and checking.
Stop: if reminders increase anxiety or compulsive checking. Never use this as the sole safeguard for medication, childcare, driving, emergency or legal duties.
Task-duration forecastingMake the direction and size of planning bias visible.
What this can show: how estimates compare with actual active time across one task class. A forecast is never certainty.
- Choose at least ten comparable, low-stakes tasks.
- Define active time, waiting time, completion and invalidation in advance.
- Record an estimate before starting each task.
- Record actual active time without hiding overruns or abandoned tasks.
- Calculate signed error and one fixed absolute-error measure.
- Test an outside-view estimate on a different predeclared task class.
Direct: signed and absolute error.
Retention: a later rolling window.
Transfer: a different task class.
Burden: logging time and false precision.
Stop: if logging costs more than the decisions it improves or becomes self-punishment. Never turn an estimate into a promise.
Claim-and-counterargument decision memoMake evidence, assumptions, disagreement and possible updates auditable.
What this can show: whether one low-stakes decision states its evidence, opposition and crux clearly. It is not universal persuasion skill.
- Choose one low-stakes decision and state a falsifiable claim.
- Separate cited evidence, direct observation and assumption.
- Write the strongest opposing case in terms its supporters would recognise.
- Name the crux most likely to change the decision.
- State what evidence would change your mind and record confidence.
- Repeat after a delay on a different low-stakes topic.
Direct: source fit, opposing-case fidelity, crux quality and justified update.
Retention: a new claim after a delay.
Transfer: a different topic and sources.
Burden: time, defensiveness and conflict.
Stop: if the exercise becomes coercive, repetitive or harmful to a relationship. High-stakes claims still require qualified expertise.
What remains locked: Titan has released no G4 public training programme. The four G3 guides above inherit existing low-risk operational exemplars; their publication creates no research promotion or independent validation. PWR-078 remains on publication and operational hold; this correction-only notice does not require outside advisers, while any future explainer, training or navigation proposal would require its own higher-authority review. Read the correction →
The discovery loop
Find the signal. Preserve the doubt. Test what matters.
Titan can now absorb extraordinary claims without becoming either credulous or cynical: preserve the original question, identify the measurable capability, retain hostile evidence, govern actionability and reopen the record when better evidence arrives.