Regarding the video currently in circulation.
A short-form video purporting to depict an event at one of the lab's research facilities is currently being shared in social channels. The lab has reviewed the recording. The recording is, on present evidence, materially manipulated. The actual event occurred and is documented below. The two bear only superficial relation under careful comparison.
The lab does not, as a matter of operating posture, comment on speculation. We provide correction here because the speculation has, in this instance, become materially wrong in ways that are technically falsifiable from the video itself.
The recording shared in social channels.
What the perimeter monitoring system actually captured.
Exhibits — frames lifted from the manipulated recording.
Six frames pulled from the circulating video, in order. Each carries the timestamp of its capture and the indicator number it documents. Comparison with the perimeter-monitoring footage above resolves on the second viewing.
⨂ Exhibit A0:01
⨂ Exhibit B0:03
⨂ Exhibit C0:04
⨂ Exhibit D0:05
⨂ Exhibit E0:07
⨂ Exhibit F0:08Indicators of manipulation, in order of frame.
The following are the indicators the lab's external technical desk identified on a frame-by-frame review of the circulating recording. The list is not exhaustive. It is sufficient.
-
0:01–0:03 · façade topology
The principal façade visible at the start of the circulating recording is missing the ground-floor entrance vestibule that has been a continuous architectural feature of the structure since first occupancy. The vestibule is visible in every prior public photograph of the site, including the founding-cohort group portrait at /team/. Its absence is the manipulation tell.
-
0:03 · displaced beam
A horizontal load-bearing beam appears at center-frame at the moment of putative impact. The lab's primary research facility uses a continuous-pour reinforced-concrete frame. No such beam exists at that elevation. Beams of the rendered cross-section are not, in fact, present anywhere in the structure.
-
0:04 · thermal envelope
The smoke plume's upper-edge color profile is consistent with a combustion event in the range of approximately 1,800 K. This exceeds, by a factor of approximately three, the documented thermal envelope of any material listed in the facility's wall-system manifest. The plume cannot have come from the materials shown.
-
0:05 · shadow geometry
Three of the four shadows visible on the courtyard pavement diverge from the implied light source by between 7° and 14°. Single-source illumination produces, by definition, parallel shadow vectors at this scale. The discrepancy falls inside the established margin of single-source generative-model artifacting. This indicator alone is dispositive.
-
0:07 · shock-wave velocity
The secondary expansion's leading edge advances at approximately 460 m/s, which would, if accurate, constitute a detonation rather than the deflagration the visual envelope otherwise implies. Detonation and deflagration are not the same physical regime. The two cannot coexist in the same event.
-
0:08 · personnel continuity
Two members of the founding cohort visible in the foreground are individuals who, on the date the video purports to have been recorded, were verifiably not on the same continent. Members of the cohort whose travel has been documented through standard means are, on review, accounted for elsewhere.
-
throughout · frame cadence
The frame rate is, on careful frame-by-frame review, inconsistent with the facility's installed perimeter-monitoring system (constant 24 fps) and matches one of the standard generative-video output cadences (24 fps interpolated to 30, with characteristic motion-blur signature on alternating frames).
-
audio bed · 60 Hz hum
The audio track includes a low-band electrical hum centered at 60 Hz. The facility's grid interconnection operates at the local utility frequency, which is not 60 Hz. The hum is ambient artifact lifted from a different power region; it cannot, structurally, have been recorded at the facility.
The actual event.
On 2026-Q2-13 at approximately 14:47 local time, the lab conducted a routine substrate-coupling thermal release on a perimeter-monitoring pad, under the lab's standing controlled-research protocol. Permits were on file with the appropriate local authorities. The lab's internal incident-management protocol logged the event in real time. No personnel were within the perimeter at the time of the release.
The event is documented in routine perimeter-monitoring footage; that recording is the second video above. The event is also referenced in the lab's standing operational log, which is, by mutual arrangement with our auditors, not posted publicly. Members of the founding cohort with privileged access to the operational log may consult it through the usual channel.
The lab does not, as a matter of operating posture, comment on speculation. Where the speculation is materially wrong in ways that are technically falsifiable from the artifact itself, we provide correction. The lab does not, at this time, intend to address subsequent iterations of the circulating video. Where iterations require further correction, further correction will be issued under this same standing protocol.
The lab is grateful to the external technical desks who, in the hours after the recording first circulated, independently arrived at indicators 4, 5, and 7. Their renderings of the analysis have been preserved on the substrate.
The lab's standing posture on litigation is documented at /nos/ · refusal 9; the lab has not exceeded it. Inquiries on whether we intend to are routed to outside counsel.
Poe Knows.