
Founder @GolpoAI (YC S25) Prev: CS @ stanford
San Francisco, CA
Joined July 2024
- Tweets382
- Following375
- Followers1K
- Likes3.4K
Pinned Tweet
Today, we are launching Golpo, the world’s most powerful AI video generation model to help you create AI-generated explainer and animation videos.
With just a prompt, Golpo turns your ideas into stunning animated and whiteboard-style videos
Enjoy
P.S. If you want some free credits, please drop your email in the comments below.
Here’s a Golpo… explaining Golpo
its crazy how much ai has improved in spatial reasoning and understanding of time. it can now draw complex out and understand geometric relationships between objects.
this video below explains gradients and directional derivatives in multivariable calculus, using concentric level curves for f(x,y) = x² + y², coordinate axes, gradient vectors, tangent directions and mathematical equations. the geometry here is pretty complex. all the level curves need to share the same center, the point P = (1,1) needs to lie on the correct contour, and the gradient arrow needs to point in the direction of steepest increase while remaining perpendicular to the tangent. even a small error in the positioning of these elements would make the mathematical explanation incorrect. its rrly interesting how ai is getting better at handling these kinds of geometric constraints.
golpo progressively drew the coordinate axes, traced the circular level curves, positioned the arrows and labels, and then worked through the partial derivatives and gradient calculations. the equations eventually show how the gradient at P becomes (2,2), why the maximum directional derivative is 2√2, and why moving along the tangent gives a directional derivative of zero. maintaining the spatial relationships between all these elements while coordinating their appearance over time requires some pretty sophisticated planning. even the handwritten equations have to appear in the correct sequence, w/ fractions, symbols and mathematical notation staying consistent throughout.
what i find rrly cool is how ai is starting to handle all these different technical problems together. understanding the geometry of a multivariable function, representing its level curves visually, reasoning about perpendicular vectors, planning the drawing order, and coordinating everything on a frame-by-frame timeline. the amount of spatial reasoning, mathematical understanding, temporal planning and software execution involved is rrly crazy.
The spatial awareness and reasoning capabilities of video models are improving much faster than I anticipated. The video shows just how far they’ve come. We are now able to render realistic and accurate anatomical structures to accurately depict their spatial relationships and how they work together.
So how good really are current SOTA AI models at explaining concepts visually? I asked one to draw an explanation of mean, median, and mode. After few minutes, it created this. You can obviously it exactly how detailed or simple you want the explanation to be.
1
17
489
This video below is about Wabi Sabi: the concept that finds beauty in imperfection.
It was fully made by AI in 1 prompt and in under 10 mins
Excited to announce that Golpo AI is now available inside Jira and Confluence through our new Atlassian Marketplace apps.
For enterprise teams, this means employees no longer have to leave their existing workflow to create explainer videos through the Golpo website, Claude Code, Codex, or separate API tools. They can generate videos directly from Jira tickets, epics, sprint summaries, and Confluence pages where their work already lives.
This makes it much easier for teams to turn technical context, runbooks, SOPs, onboarding docs, customer support articles, product updates, and engineering decisions into clear, watchable videos. Instead of asking people to read long tickets or pages, teams can now create AI-generated explainer videos in a few clicks inside Atlassian.
The video below was made fully with Golpo
Gemini Omni is a huge step forward but it still is not practical for actually explaining things that are technical. Like the videos generated by Omni still cannot explain you multivariable calculus or a deep codebase.
That is why I think the future of AI video is specialized models like Golpo that prioritize explanations. The below video from Golpo AI shows this in practice .
Thanks to OpenAI for recognizing the work we are doing.
Growing up, whenever I was stuck on a concept and later while tutoring students from very different backgrounds and languages, the thing that consistently helped wasn’t a textbook or lecture. It was a clear visual explanation with intuitive examples, at the right pace.
The problem was that creating these kinds of personalized learning experiences traditionally required enormous manual effort and production time, making them impossible to scale for most educators and institutions.
But with AI, we saw a real opportunity to reimagine education and make learning more engaging, personalized, and accessible for everyone, regardless of geography or language. That’s what led me and @KarShreyas to build @GolpoAI
, where anyone can turn PDFs, lesson materials, scripts, prompts, presentations, or onboarding content into engaging visual explainer videos in any language within minutes.
Today, Golpo (video.golpoai.com) supports educational institutions, teachers, edtech platforms, L&D teams, and enterprises across 55+ countries and in over 50 languages, helping more than 100,000 users create engaging visual explanations and educational videos tailored to their specific curriculum, audience, and learning goals. That includes BreakFree Education transforming curriculum into video lessons for students in juvenile justice programs, the Garnet Valley School District recreating science lessons for a student learning remotely for medical reasons, and organizations, presentations, onboarding materials, and lesson plans into multilingual explainer videos.
We’re very excited about what’s ahead.
Introducing the ChatGPT Futures Class of 2026—26 honorees from the first graduating class to have had ChatGPT throughout all four years of university, who used AI to:
- Map 1.5M previously unknown objects in space
- Detect disaster survivors through walls and debris
- Make 100M+ galaxy images searchable
- Preserve endangered languages
- Build infrastructure to reroute 5M+ pounds of unsold inventory from landfills
Introducing the ChatGPT Futures Class of 2026—26 honorees from the first graduating class to have had ChatGPT throughout all four years of university, who used AI to:
- Map 1.5M previously unknown objects in space
- Detect disaster survivors through walls and debris
- Make 100M+ galaxy images searchable
- Preserve endangered languages
- Build infrastructure to reroute 5M+ pounds of unsold inventory from landfills
Education shouldn’t be limited by geography or language.
Today, we are really proud to announce that Golpo has now supported education institutions, teachers, and students across 55 countries and in over 50 languages.
Golpo delivers powerful, AI-driven video lessons across every domain.
Below is an example of a physics practice problem walkthrough created by Golpo in under 2 minutes with just 1 simple prompt.
Big News: @GolpoAI in @ChatGPTapp !
You can create an explainer video on literally any topic and start the process directly in a ChatGPT conversation.
This is fully AI btw
How did the Bronze Age end? One reason was a group of invaders called the Sea Peoples.
Explainer video on the Sea Peoples created in one prompt with @GolpoAI
I think education is finally democratized. You can make high quality, accurate and engaging educational lessons like these in 50+ languages in just 1 prompt and in under 2 mins





