Skip to content

AI & Automation1 min read

Andrej Karpathy builds two video platforms in one weekend

Andrej Karpathy used Claude models to create two platforms in 48 hours, lowering the barrier for businesses to build internal tools.

by Fini MironFounder @ Balkaris
An abstract composition printed flat in pale cool grey, steel blue, deep navy and bright mustard yellow: one enormous triangle filling the lower right and cropped by two edges, and a single small square stranded alone in the empty upper left.
Drawn for this article on our own machine. It is not a photograph.

Andrej Karpathy built two video platforms over a single weekend in November 2026. According to The-decoder, Karpathy used Claude's Opus 5.5 model and Claude Code to develop HackerTube and TristanTube. HackerTube distilled 43 million comments from 20 years of Hacker News into 6,454 videos. Karpathy stated that Claude Code handled the data analysis and server setup. His manual tasks were limited to securing a Google Cloud API key, a rented server, and a domain.

Why it matters

The speed of development changes the cost of experimentation. For a Swiss business, this means the barrier to creating niche internal tools is lower. You can move from a concept to a functional prototype in hours rather than weeks. This shift rewards those who can define clear requirements quickly.

  • Lower costs for bespoke data processing tools.
  • Faster testing of new product ideas.
  • Reduced reliance on large engineering teams for simple deployments.

What to watch

Businesses should evaluate their current internal tool roadmap. If a project requires weeks of manual coding, it is now a candidate for rapid prototyping. You should focus on refining your data inputs and specific logic requirements. The value shifts from writing code to defining the correct parameters for the model to execute.

In a year, the primary bottleneck for new products will shift from engineering capacity to the clarity of your requirements. This happens if you stop coding and start defining.

Key takeaways

  • The cost of creating functional prototypes has decreased.
  • Businesses can move from concept to prototype in hours.
  • Value is shifting from manual coding to defining specific logic parameters.

Frequently asked

Does this mean I can replace my entire engineering team with AI?

The article suggests reduced reliance on large teams for simple deployments, but it does not claim full replacement of engineering roles.

What specific data do I need to provide to make this work?

The text does not specify the data types, only that you must refine your data inputs and specific logic requirements.