Add multi-view SfM benchmark to motivation cards (home + README)

Ground-truth camera intrinsics and metric depth across 10 synced
viewpoints make the dataset a clean reference for SfM/MVS/pose
estimation. Switch the motivation grid from 5x1 to 3x2 to fit
6 cards.

README.md

@@ -37,6 +37,7 @@ This dataset turns those demos into rendered, frame-accurate, fully-annotated tr
 - **Forward dynamics / world models.** `(frame, action) → frame_{t+1}` with all 10 player POVs of the same world state available as supervision.
 - **Spatial-audio conditioning.** Per-player stereo recorded relative to each agent's position and orientation, so models can learn to localize footsteps, gunfire, and callouts.
 - **Multi-agent training.** All 10 perspectives of the same round are kept aligned tick-for-tick — useful for collaborative policies, opponent modelling, and multi-player world models.
+- **Multi-view SfM / depth benchmark.** Ground-truth camera intrinsics and metric depth maps from 10 synchronized viewpoints make this a clean reference for structure-from-motion, multi-view stereo, and camera pose estimation.
 
 ## What's recorded
 

src/routes/+page.svelte

@@ -13,6 +13,7 @@
 	import FilmReelIcon from 'phosphor-svelte/lib/FilmReelIcon';
 	import SpeakerHighIcon from 'phosphor-svelte/lib/SpeakerHighIcon';
 	import UsersThreeIcon from 'phosphor-svelte/lib/UsersThreeIcon';
+	import CubeIcon from 'phosphor-svelte/lib/CubeIcon';
 	import { mode, toggleMode } from 'mode-watcher';
 	import SunIcon from 'phosphor-svelte/lib/SunIcon';
 	import MoonIcon from 'phosphor-svelte/lib/MoonIcon';
@@ -131,6 +132,11 @@
 			icon: UsersThreeIcon,
 			title: 'Multi-agent training',
 			body: 'All 10 perspectives stay tick-aligned for collaborative or competitive policies.'
+		},
+		{
+			icon: CubeIcon,
+			title: 'Multi-view SfM benchmark',
+			body: 'Ground-truth camera intrinsics and metric depth maps from 10 synced viewpoints — free supervision for SfM, MVS, and pose estimation.'
 		}
 	] as const;
 </script>
@@ -236,7 +242,7 @@
 			observability, dense visual signal, and 5v5 multi-agent dynamics. This dataset turns them into
 			trainable rendered video.
 		</p>
-		<div class="mt-8 grid grid-cols-1 gap-4 sm:grid-cols-2 lg:grid-cols-5">
+		<div class="mt-8 grid grid-cols-1 gap-4 sm:grid-cols-2 lg:grid-cols-3">
 			{#each motivations as m (m.title)}
 				<div class="rounded-md border bg-card p-3">
 					<m.icon size={18} weight="duotone" class="mb-2 text-foreground/80" />