Inside the AI rock-slope video: visual effects and real monitoring

A roughly 51-second first-person hiking clip turns a rock face into an interactive display. Colored outlines mark possible footholds while labels show grip, slope and candidate positions. In a post on X, Poonam Soni describes it as AI using cameras and LiDAR to identify dangerous rock slopes.

Tracing the source leads to a more accurate description: a creative visual-effects piece with a game-like interface. In the original Instagram post, @thebigflatnow playfully credits a neural model called “Brain™” and labels the work as creative coding, VFX and motion design. It offers material for an interface discussion, but does not establish a deployed AI safety-monitoring system.

Watch the clip

Video shared by Poonam Soni on X; original creative work by @thebigflatnow. The video is served by this site with the footage and source attribution preserved.

How the interface presents a decision

The overlays identify several possible footholds on an uneven surface, then distinguish options through color, short labels and percentages. This makes an abstract comparison spatial: viewers can see which locations are being considered.

Those numbers and outlines are visual elements in the clip. The available material does not provide a sensor configuration, model identifier, raw point cloud, error measurements or field-validation report. It therefore cannot establish what the percentages measure, or whether a live model produced the apparent selections.

Ideas for games and interactive systems

Three design ideas stand out:

• Show the candidate actions. Visible alternatives make the next choice easier to understand.
• Place explanations beside the object. Local labels reduce the need to look back and forth between a scene and a separate panel.
• Make state changes legible. Moving from candidates to evaluation to selection gives viewers a process they can follow.

These ideas could inform climbing games, NPC path-selection debug views or training simulators. They are design possibilities suggested by the footage, not implemented features established by the original post.

What real slope monitoring needs to demonstrate

Image-based monitoring operates on different evidence and timescales. According to NHAZCA's official IRIS description, the system aligns images, handles lighting variation and compares them against a baseline to identify scene changes and track displacement over time. The vendor describes it as a complement to conventional geotechnical instruments, with human review available for critical alerts.

That is a separate vendor's product description; no verified connection to this video has been established. Assessing a real system requires its measurement method, errors, longitudinal records and validation results. Compelling graphics alone cannot supply that evidence.

The creative value of this clip lies in making observation, comparison and feedback easy to follow. Keeping the creator's description attached to it lets readers appreciate that design without confusing the effect with a field-tested instrument.

Sources

• Poonam Soni's post on X
• Original Instagram reel and creator description by @thebigflatnow
• NHAZCA IRIS official product description

Prepared from the public posts, creator description and video imagery. No slope-monitoring system was field-tested for this article.