Back Office · office.temerarii.xyz
One asset, all the way in — composition, the wireframe + storyboard, the output format stack, and the template, all read from the SAME content-index record. The expected output matches what /media surfaces for this post.
post social-W36-Sat-2kind threadweek W36date 2026-09-12campaign thread · Satpillar staff_trainingbeat Satasset videoduration 46.9sground blackscenes 9

Checklist the per-video bar — engine/sim

83.0/100
plain languagevo coverageno dead airuniquenesscaption fitcompletenesscleanliness
quantitative quality · weights learn from your reviews (engine.sim.memory review social-W36-Sat-2 good|bad)
⚠ 4 flag(s) — not yet ship-ready: copy_generictoo_complexdup_sequence_across_assetsdup_set_across_assets · see docs/strategy/VIDEO-CHECKLIST.md

Composition comp · template family · expected output

composition SceneReelfamily / template SignalFieldReel
9:16 Reelpending1:1 Squarepending16:9 Widepending9:16 4Kpending1:1 4Kpending16:9 4KpendingGIF (SMS)pending
render pending — silent master not yet on disk
expected output: 0/7 rendered — same matrix the /media preview surfaces for this asset.

Composition layer × scene 9 scenes · 46.9s · comp_id + rendered still + tier + the script

#Layer (comp_id · still · tier)BeatTimecodeMotionLogoAudioVO / on-screen / caption
1s1
matches intent
shared field
signature-3d		open0–5.5sspatial-parallaxicon·wireframe♪ bed_in
Machine learning used to sound like a club you needed a PhD to enter.
on-screen: ML sounded like a PhD-only club
expected on screen: black ground · knot hero in the shared Signal Field · Magister leads · node-graph · spatial-parallax · icon·wireframe logo · caption bottom-left
spec (the prompt): comp_id shared Signal Fieldvisual node-graphshape knotground blacktreatment wireframemotion spatial-parallaxpower summoninstrument summon→ML sounded like a PhD-only club
2s2
matches intent
shared field
signature-3d		hook5.5–10.5skinetic-buildicon·wireframe♪ node_lock
Most folks just nodded along and quietly hoped nobody asked a question.
on-screen: Most people just nodded along
expected on screen: black ground · knot hero in the shared Signal Field · Magister leads · mark · kinetic-build · icon·wireframe logo · caption bottom-left
spec (the prompt): comp_id shared Signal Fieldvisual markshape knotground blacktreatment wireframemotion kinetic-buildpower laser-lockinstrument laser-trace→Most people just nodded along
3s3
matches intent
NumberedList
template		teach10.5–16.0skinetic-buildicon·wireframe♪ node_lock
Here is the plain truth: a model just learns patterns from lots of examples.
on-screen: It's just patterns from examples
expected on screen: black ground · a NumberedList panel over a dimmed Signal Field · Magister leads · node-graph · kinetic-build · icon·wireframe logo · caption bottom-left
spec (the prompt): comp_id NumberedListvisual node-graphshape knotground blacktreatment wireframemotion kinetic-buildpower morphinstrument morph+laser→It's just patterns from examplescurate items, nodes
4s4
matches intent
RankList
template		proof16.0–21.2sreceipts-counticon·wireframe♪ node_lock
Feed it good clean examples, and it makes useful guesses on new ones.
on-screen: Good examples in, useful guesses out
expected on screen: black ground · a RankList panel over a dimmed Signal Field · Magister leads · receipts · receipts-count · icon·wireframe logo · caption bottom-left
spec (the prompt): comp_id RankListvisual receiptsshape knotground blacktreatment wireframemotion receipts-countpower receiptsinstrument spotlight→Good examples in, useful guesses outcurate rows, statsLabels
5s5
matches intent
BuildLog
template		diff21.2–26.4scrossfade-8ficon·wireframe♪ node_lock
But messy or biased examples teach it to guess wrong, with full confidence.
on-screen: Garbage examples teach garbage
expected on screen: black ground · a BuildLog panel over a dimmed Signal Field · Magister leads · code · crossfade-8f · icon·wireframe logo · caption bottom-left
spec (the prompt): comp_id BuildLogvisual codeshape knotground blacktreatment wireframemotion crossfade-8fpower morphinstrument morph→Garbage examples teach garbagecurate codeLines, lines
6s6
matches intent
ChecklistCard
template		teach26.4–31.9skinetic-buildicon·wireframe♪ node_lock
So the real work is cleaning your examples long before you touch a model.
on-screen: Clean your data before the model
expected on screen: black ground · a ChecklistCard panel over a dimmed Signal Field · Magister leads · node-graph · kinetic-build · icon·wireframe logo · caption bottom-left
spec (the prompt): comp_id ChecklistCardvisual node-graphshape knotground blacktreatment wireframemotion kinetic-buildpower morphinstrument morph+laser→Clean your data before the modelcurate items, nodes
7s7
matches intent
StatScoreboard
template		proof31.9–36.9sreceipts-counticon·wireframe♪ node_lock
Better examples beat a fancier model almost every single time, quietly.
on-screen: Better data beats a fancy model
expected on screen: black ground · a StatScoreboard panel over a dimmed Signal Field · Magister leads · receipts · receipts-count · icon·wireframe logo · caption bottom-left
spec (the prompt): comp_id StatScoreboardvisual receiptsshape knotground blacktreatment wireframemotion receipts-countpower receiptsinstrument spotlight→Better data beats a fancy modelcurate pillar, stats, statsLabels
8s8
matches intent
StepFlow
template		step36.9–41.9skinetic-buildicon·wireframe♪ node_lock
Try it: label fifty examples carefully by hand before you train anything.
on-screen: Label fifty examples by hand first
expected on screen: black ground · a StepFlow panel over a dimmed Signal Field · Magister leads · pipeline · kinetic-build · icon·wireframe logo · caption bottom-left
spec (the prompt): comp_id StepFlowvisual pipelineshape knotground blacktreatment wireframemotion kinetic-buildpower throwinstrument laser-trace→Label fifty examples by hand firstcurate stages, steps
9s9
matches intent
shared field
signature-3d		resolve41.9–46.9scoalescenceicon·wireframe♪ bed_out
Turns out the mystery was mostly careful examples and patient, honest labels.
on-screen: The mystery was just careful data
expected on screen: black ground · knot hero in the shared Signal Field · Magister leads · coalescence · coalescence · icon·wireframe logo · caption bottom-left
spec (the prompt): comp_id shared Signal Fieldvisual coalescenceshape knotground blacktreatment wireframemotion coalescencepower coalescenceinstrument coalescence→The mystery was just careful data

Format stack 3 aspects · same scenes[], re-cropped

9:16
1080×1920
Stories · TikTok · YouTube Shorts · Reels
1:1
1080×1080
LinkedIn · Facebook · Instagram
16:9
1920×1080
X/Twitter · YouTube · LinkedIn video

Channels 9 destinations

LinkedInX/TwitterYouTubeInstagramFacebookThreadsTikTokPinterestBluesky

Social captions supplemental published copy · per channel (comp_id level)

tiktokMachine learning, demystified: a model just learns patterns from examples. Here is the real move: messy examples teach it to guess wrong with confidence, so clean your data BEFORE the model. Better examples beat a fancier model. (AI-assisted) #machinelearning #ai #datascience
instagramML demystified. A model learns patterns from examples. Good data in, useful guesses out. Messy data teaches it to guess wrong. Better examples beat a fancier model. #machinelearning #ai #datascience #aitools #tech
linkedinMachine learning used to sound like a club you needed a PhD to enter. The plain truth: a model learns patterns from examples, so good clean examples produce useful guesses and messy ones produce confident mistakes. The takeaway: the real work is cleaning your examples long before you touch a model. Better data beats a fancier model almost every time.
xMachine learning, plainly: a model learns patterns from examples. Messy data teaches it to guess wrong with confidence. Clean your data before the model. temerarii.com
facebookMachine learning used to sound like a PhD-only club. The plain truth: a model learns patterns from examples, so good data makes useful guesses and messy data makes confident mistakes. Clean your data before the model. See it at temerarii.com.
threadsMachine learning, plainly: a model just learns patterns from examples. Messy examples teach it to guess wrong with confidence. Clean your data before the model. Better data wins.
pinterestMachine learning explained simply: why data beats the model. A model learns patterns from examples, so clean your data before training. Better examples beat a fancier model. Evergreen guide to machine learning basics, data quality, and AI fundamentals.
blueskyMachine learning, plainly: a model learns patterns from examples. Messy data teaches it to guess wrong with confidence. Clean your data before the model. temerarii.com
youtubeMachine Learning Demystified: Why Data Beats The Model Machine learning used to sound like a club you needed a PhD to enter, so most people nodded along. The plain truth: a model just learns patterns from lots of examples, so good clean examples produce useful guesses and messy or biased ones produce confident mistakes. The real work is cleaning your examples long before you touch a model. Label fifty examples carefully by hand to start. More at temerarii.com.

Cross-links every lens is a view on this one record

→ Wireframe · scenes→ Calendar week→ Media (audience preview)→ Storyboard hub→ Day view→ Channel view