Week 2 Set: Tell annotation

Second set — turning the specimen shelf's rough, unlabelled stacks into timestamped tell-logs, using the vocabulary the screening just gave you

1 March 2027

Before the session#

Nothing to bring — the specimens are already on the shelf from week 1. Read the screening first; you’re using its vocabulary today, not inventing your own.

In the session#

Round one — swap and log (30 min). Pair up and trade specimens. For the clip you didn’t bring, produce a tell-log: a numbered list of timestamps, each naming one specific tell (uncanny hand, morphing logo, temporal flicker, the affectless narrator, the too-smooth cadence, or a tell the screening didn’t name that you can still describe precisely). “The face looks weird at 0:14” is not an entry. “0:14 — the left eyebrow’s position resets between two nearly identical frames” is.

Round two — check the log, not the clip (15 min). Return the log to whoever brought the specimen. Their job is to verify each entry against the actual frame, not to argue about whether it’s a real tell — if the timestamp doesn’t show what the entry claims, the entry fails, regardless of whether the clip is slop by some other measure.

Round three — compare to week 1’s working names (15 min). Pull up the informal stack-names from the specimen shelf. Where did the untrained sort from week 1 land close to the formal tell categories, and where was it wrong in an interesting way — a stack that grouped by subject matter (“cooking videos”) rather than by production tell, say. Log the comparison on the shelf next to the original entry.

Afterwards#

The tell-logs stay on the shelf. Week 6’s Slop Teardown asks for exactly this kind of timestamped, specific evidence, at greater length and against one artefact you choose yourself — today is where the habit gets built, on someone else’s clip, where getting it wrong costs nothing.