Building the visualization layer for AI
The Hook
“These two kids grew up in two of the most remote parts of Nepal. They had never heard of MIT, Y Combinator, or Venture Capital.”
The Hook
“But they flew across the world, got into MIT, went through YC, and raised $4.8 million. Those two kids are me and my co-founder Kartikesh.”
Problem Agitation
“We met at MIT, but the idea behind Lamina started much earlier. We both knew what it felt like to struggle through difficult ideas without having the right explanation, teacher, or visuals in front of us. Then LLMs arrived and made knowledge abundant, but knowledge is not the same as understanding.”
Problem Agitation
“The world's most intelligent models still answer every question in the same format: paragraphs on a screen, because text was the easiest thing for computers to generate. We believe that is about to change. When a student asks why, the answer should become a lesson. When a workflow changes, the documentation should become a walkthrough. When an AI agent reaches a conclusion, it should be able to show you exactly what it means.”
Product Reveal
“So we built Simi. Give it a prompt, document, or workflow, and it generates a complete narrated explainer video in seconds. But Simi is only our first product.”
Feature Teaser
“Lamina Labs is an applied AI lab building the fastest video infrastructure for LLMs. Infrastructure that allows any model, agent, or application to generate precise, controllable video at the moment it is needed.”
Call to Action
“Today, we're announcing $4.8 million in funding, led by Foundation Capital with participation from Link Ventures and Y Combinator, to build Lamina Labs. We're not building another AI video app. We're building the layer that turns machine intelligence into something people can see, follow, and understand.”
Synthesizable in Remotion · key components of the founder-origin-infrastructure-debut template.
Simulates hand-placed archival paper snippets and photos snapping onto surface with subtle drop shadow and rotation jitter.
Use Remotion spring on translation and rotation with discrete frame intervals or step functions to replicate 12fps tactile stop-motion.
Frosted glass container overlay positioned in 3D perspective over live-action video with typing text feed.
Absolute positioned div with backdrop-filter blur, white border stroke at 0.15 opacity, and inner text revealed via string slice keyed to frame count.
Words scale and shift color as founder speaks, transitioning from low-contrast charcoal to accented emerald teal.
Map transcript array with interpolateColor and spring-driven translateY offset triggered at exact audio timestamps.
Multiple progress bars race horizontally; competitor bars move slowly while product bar accelerates instantly to completion.
Animate flexbox width using spring configs with high damping for competitors and low damping with high velocity for Simi.
Purpose. Establish personal grit and dramatic transformation through tangible paper artifacts.
Execution. Overhead stop-motion hands placing photos and university logos onto cardstock.
Purpose. Contrast text-heavy LLM outputs with true visual understanding.
Execution. Frontal live-action dialogue overlaid with floating paragraphs and kinetic text.
Purpose. Demonstrate the capability of Simi and benchmark its raw infrastructure speed.
Execution. Input bar screencast, diagram animation, and comparative speed bar metrics.
Purpose. Solidify venture credibility and deliver concluding thesis statement.
Execution. Bento card reveals of VC logos followed by a clean typography and brand lockup.