Uncertainty and next research
The useful next step is to expand and verify the creative archive. This wiki describes published content; it does not establish a complete personal biography. The 66 recovered posts are a selective sample of a profile reporting 933 posts. Reaching the end of a third-party archive is not proof that Instagram was fully collected.
What is established, and what is still a hypothesis
| Claim | Current evidence | Limit |
|---|---|---|
| A recurring teaser design routes viewers to YouTube. | Persistent upper headline, central interview window and lower CTA in DaP4DIEp0CY, DdTnCXNpVfv, DZzYlGkJDwX. | The sample cannot establish its share of all historical posts or actual click-through. |
| Cars, travel and creator achievements contribute to public positioning. | DZuUWWCFzYE, Db3GRptlz85, DYHgP-Fp6cP. | A depicted possession or milestone is not audited wealth, a daily routine or proof of conversion. |
| Provocation and confession recur as hooks. | The forensic, sexual-health and deportation excerpts documented in narrative. | Full conversations may add nuance, challenge a guest or alter the apparent conclusion. |
| Collaborators are publicly associated with individual posts. | Publishing-account metadata and explicit coauthor headers in the inventory. | An overlay or coauthor handle is not a label identifying the nearest visible face. Nor does it prove a private friendship. |
| The sample addresses entertainment, status and creator-growth interests. | Interpretation of the contrasting interview, lifestyle and mentoring formats in the passport. | Actual audience demographics, motivations and purchases were not available. |
Corrections that changed this edition
Layout, not an edit. Inspection of the opening contact sheets for Db7_WdkJkSm, DZzYlGkJDwX and DdTnCXNpVfv showed persistent upper/lower graphic regions around a moving center. Earlier machine notes sometimes called those regions sequential title cards. Those descriptions were corrected. This matters for reproducing the actual composition.
Text attribution, not face attribution. The orest_murii promotion in DdYqtp0Jk5E does not securely assign that name to the black-jacketed participant. The participant description and the public coauthor record are now kept separate in people.
Metrics shown, not verified. DblT2qxpR4s contains a small performance screenshot over a night car selfie. Earlier notes described it as Instagram/Reels performance. The platform was not independently verified, so the wiki treats this as a displayed performance claim. The transcript resembles background lyrics and is not promoted to a statement by the creator.
Locators need semantic review too. Two proposed document/flag reference rows cited a nutrition reel. They were removed. The remaining reference catalogue passes checks for source, asset type and sampled timestamp; that check cannot prove every written description is correct. The catalogue remains a set of candidates, not a reviewed screenshot pack.
Where the evidence can mislead
- Sampling: a three-second grid with additional opening and final frames captures scene composition, but can miss a fast gesture, a brief prop or several cuts. It does not justify exact cuts per minute, gesture frequency or continuous movement instructions.
- Audio: 47 full automatic transcripts are available; a further video had no audio stream. Ukrainian/Russian switching, proper names, slang and music can confuse ASR. There is no reliable automatic host/guest diarization in this edition.
- Reused footage: multiple reels can originate from one interview or event. Repetition establishes a publishing template, not several independent real-world occasions.
- Promotional wording: captions and overlays establish how an offer, relationship or achievement is presented. They do not independently validate the underlying claim.
- Selection: the archive surfaces many interview promotions and lifestyle posts. Missing periods could contain substantially different content. No observed format ratio should be reported as an account-wide ratio.
- Generated notes: detailed source pages retain machine observations and explicit uncertainty. Their numeric confidence fields are model estimates, not calibrated statistical probabilities. Editorial pages carry the reviewed interpretation.
Research priorities
| Priority | Question | Evidence needed | Expected improvement |
|---|---|---|---|
| 1 | What is missing from the historical grid? | An authorized Instagram export/session or a lawful archive with stable post IDs and complete pagination. | Deduplicated history, calendar coverage and a defensible denominator. |
| 2 | Do teaser conclusions survive the full conversation? | Accessible full episodes corresponding to the forensic, deportation, hacker and sexual-health promos. | Argument context, speaker roles, answers omitted by the excerpt and more accurate topic summaries. |
| 3 | Which phrases are truly recurrent in the creator's own speech? | Listening checks of selected audio, reliable speaker labels, then phrase counts across independent recordings. | A defensible vocabulary rather than a mix of captions, guest language and ASR guesses. |
| 4 | Which references are suitable for generation? | Manual inspection of each selected frame at native resolution, with time/asset locators preserved. | Clean framing, fewer overlays, useful angles and a coherent outfit/set reference pack. |
| 5 | How does the creative style change over time? | More posts in sparse periods and a record of reused clips. | Real format evolution rather than archive-selection effects. |
| 6 | What actually performs? | Creator-authorized reach, retention, saves, shares and conversion data with dates. | Comparisons of formats and hooks without treating a promotional screenshot as an audit. |
Maintaining the wiki
Keep original downloads immutable in a private acquisition run. Correct analysis notes in version control and record why a claim changed; preserve the previous interpretation in the research log or revision history. A source ID remains stable when a description is corrected.
On each new ingestion, compare post IDs and media hashes, record unavailable items, refresh the coverage ledger and update only affected topic pages. Use ordinary Markdown links so the wiki works in Git and in its website. Run link and reference-locator checks. Promoting an uncertain claim requires new supporting evidence; absence of a contradiction is insufficient.