Checklist the per-video bar — engine/sim
86.6/100
quantitative quality · weights learn from your reviews (
engine.sim.memory review social-W32-Sun-1 good|bad)⚠ 3 flag(s) — not yet ship-ready: copy_genericdup_sequence_across_assetsdup_set_across_assets · see docs/strategy/VIDEO-CHECKLIST.md
Composition comp · template family · expected output
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 · 56.0s · comp_id + rendered still + tier + the script
| # | Layer (comp_id · still · tier) | Beat | Timecode | Motion | Logo | Audio | VO / on-screen / caption |
|---|---|---|---|---|---|---|---|
| 1 |  matches intent shared field signature-3d | open | 0–6.3s | spatial-parallax | icon·3d-extrude | ♪ bed_in | The old way to teach a machine was to write every rule by hand, and pray. on-screen: The old way: hand-coded rules |
expected on screen: white ground · cone hero in the shared Signal Field · Augur leads · node-graph · spatial-parallax · icon·3d-extrude logo · caption bottom-left spec (the prompt): comp_id shared Signal Fieldvisual node-graphshape coneground whitetreatment 3d-extrudemotion spatial-parallaxpower summoninstrument summon→The old way: hand-coded rules | |||||||
| 2 |  matches intent shared field signature-3d | hook | 6.3–12.2s | kinetic-build | icon·color | ♪ node_lock | Machine learning flips it: you feed it examples and it finds the rule for you. on-screen: Machines learn from your data |
expected on screen: white ground · cone hero in the shared Signal Field · Augur leads · mark · kinetic-build · icon·color logo · caption bottom-left spec (the prompt): comp_id shared Signal Fieldvisual markshape coneground whitetreatment colormotion kinetic-buildpower laser-lockinstrument laser-trace→Machines learn from your data | |||||||
| 3 |  matches intent NumberedList template | teach | 12.2–19.2s | kinetic-build | icon·color | ♪ node_lock | Here is the move: label your data, then split it into a training set and a test set. on-screen: Label data, split train and test |
expected on screen: white ground · a NumberedList panel over a dimmed Signal Field · Augur leads · node-graph · kinetic-build · icon·color logo · caption bottom-left spec (the prompt): comp_id NumberedListvisual node-graphshape coneground whitetreatment colormotion kinetic-buildpower morphinstrument morph+laser→Label data, split train and testcurate items, nodes | |||||||
| 4 |  matches intent ComparisonTable template | proof | 19.2–26.5s | receipts-count | icon·color | ♪ node_lock | You train the model on most of it and hide the rest to check it is not just memorizing. on-screen: Train on most, hide the rest |
expected on screen: white ground · a ComparisonTable panel over a dimmed Signal Field · Augur leads · receipts · receipts-count · icon·color logo · caption bottom-left spec (the prompt): comp_id ComparisonTablevisual receiptsshape coneground whitetreatment colormotion receipts-countpower receiptsinstrument spotlight→Train on most, hide the restcurate colA, colB, rows, statsLabels | |||||||
| 5 |  matches intent TerminalRun template | diff | 26.5–32.8s | crossfade-8f | icon·color | ♪ node_lock | That hidden set is the difference between a model that learns and one that just parrots. on-screen: Memorizing is not learning |
expected on screen: white ground · a TerminalRun panel over a dimmed Signal Field · Augur leads · code · crossfade-8f · icon·color logo · caption bottom-left spec (the prompt): comp_id TerminalRunvisual codeshape coneground whitetreatment colormotion crossfade-8fpower morphinstrument morph→Memorizing is not learningcurate codeLines | |||||||
| 6 |  matches intent ChecklistCard template | teach | 32.8–38.699999999999996s | kinetic-build | icon·color | ♪ node_lock | Start with a plain model like logistic regression before you reach for a giant network. on-screen: Pick simple models first |
expected on screen: white ground · a ChecklistCard panel over a dimmed Signal Field · Augur leads · node-graph · kinetic-build · icon·color logo · caption bottom-left spec (the prompt): comp_id ChecklistCardvisual node-graphshape coneground whitetreatment colormotion kinetic-buildpower morphinstrument morph+laser→Pick simple models firstcurate items, nodes | |||||||
| 7 |  matches intent KpiGrid template | proof | 38.7–44.6s | receipts-count | icon·color | ♪ node_lock | Half the time the boring model wins, and it runs on a laptop for free. on-screen: Boring models often win |
expected on screen: white ground · a KpiGrid panel over a dimmed Signal Field · Augur leads · receipts · receipts-count · icon·color logo · caption bottom-left spec (the prompt): comp_id KpiGridvisual receiptsshape coneground whitetreatment colormotion receipts-countpower receiptsinstrument spotlight→Boring models often wincurate kpis, statsLabels | |||||||
| 8 |  matches intent NodeGraphCard template | step | 44.6–50.5s | kinetic-build | icon·color | ♪ node_lock | Score it on the hidden set, fix the weak spots, retrain, repeat until it holds. on-screen: Score it, retrain, repeat |
expected on screen: white ground · a NodeGraphCard panel over a dimmed Signal Field · Augur leads · pipeline · kinetic-build · icon·color logo · caption bottom-left spec (the prompt): comp_id NodeGraphCardvisual pipelineshape coneground whitetreatment colormotion kinetic-buildpower throwinstrument laser-trace→Score it, retrain, repeatcurate hub, nodes, stages, steps | |||||||
| 9 |  matches intent shared field signature-3d | resolve | 50.5–56.0s | coalescence | icon·3d-extrude | ♪ bed_out | That loop is all machine learning is, and now you can run it yourself. on-screen: Learning beats guessing |
expected on screen: white ground · cone hero in the shared Signal Field · Augur leads · coalescence · coalescence · icon·3d-extrude logo · caption bottom-left spec (the prompt): comp_id shared Signal Fieldvisual coalescenceshape coneground whitetreatment 3d-extrudemotion coalescencepower coalescenceinstrument coalescence→Learning beats guessing | |||||||
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, minus the mystery (AI-assisted edit). The old way was writing rules by hand. The new way: feed it examples, split your data into train and test, and let the model find the rule. The hidden test set is what tells you it learned instead of memorized. Start with a boring model. Half the time the boring one wins. #machinelearning #aibasics #learnonTikTok
instagramMachine learning without the mystery.
Feed it examples. Split train and test.
The hidden set proves it learned, not memorized.
Boring models win more than you think.
#machinelearning #ai #datascience #buildinpublic #learnAI
linkedinThe clearest mental model for machine learning we use with clients:
The legacy approach was hand-coding every rule. Machine learning flips that — you supply labeled examples and the model derives the rule.
The one habit that separates working models from demos: split your data into a training set and a held-out test set. Train on most, score on the part the model never saw. That hidden set is the honest answer to "did it learn or just memorize?"
And start simple. A plain logistic regression often beats the giant network and runs on a laptop. Takeaway: complexity is a cost, not a credential.
xOld way: hand-code every rule.
ML way: feed examples, split train/test, let it find the rule.
The hidden test set is what proves it learned instead of memorized.
Start boring. Boring usually wins. temerarii.com
facebookMachine learning, in plain words: instead of writing every rule by hand, you feed the computer examples and it works out the rule itself. The trick that keeps you honest — split your data, train on most of it, and test on a chunk the model never saw. Start with a simple model before the fancy one. Want the full walkthrough? It is on our site.
threadsMachine learning without the mystery: feed it examples, split into train and test, let it find the rule. The hidden test set is what proves it learned instead of just memorizing. Start with a boring model — it wins more often than the fancy one.
pinterestMachine learning explained simply for beginners: how to label data, split into training and test sets, and tell whether a model learned or just memorized. A plain-language guide to supervised learning, model evaluation, and why simple models often beat complex ones.
blueskyML without the mystery: feed examples, split train/test, let it find the rule. The hidden test set proves it learned instead of memorized. Start boring — boring usually wins. temerarii.com
youtubeMachine Learning Explained: Train, Test, and the Trick Most People Skip
The old way was hand-coding every rule. Machine learning flips it: you feed labeled examples and the model finds the rule. In this short we cover the move that keeps you honest — splitting data into a training set and a held-out test set so you know whether the model learned or just memorized. We also make the case for starting with a simple model before reaching for a neural net. Run the loop: score, fix, retrain, repeat.
More plain-language AI breakdowns at temerarii.com.