Using AI to Solve AI's Problem
Better compression, not bigger data centres
The AI industry is betting trillions on infrastructure. We think there's a smarter path — using AI as a research tool to explore next-generation lossless compression.
The Infrastructure Bubble
Everyone's building bigger. We're thinking smaller.
OpenAI's committed infrastructure spend over ten years. Their annual revenue? About $13 billion.
Wafers per month required for OpenAI's Stargate project by 2029, triggering a global memory chip crisis.
— Reuters, December 2025
The AI industry's solution to every problem is scale: more GPUs, more power, more cooling, more data centres. It's economically fragile and environmentally unsustainable.
We're asking a different question: What if the data itself could be smaller?
Our Research
Using AI to reduce AI's footprint
We use large language models as research tools — not products to scale infinitely. Testing compression theories. Validating formulas at scale. Exploring whether data can be encoded as continuous waveforms rather than discrete bits.
AI as Tool
Testing thousands of compression formula variations automatically
Waveform Encoding
Exploring data as continuous signals rather than discrete bits
Lossless Target
Compression that preserves every bit of the original data
The goal isn't to consume more compute. It's to eventually need less. If data can be encoded more efficiently, the infrastructure problem becomes smaller — not something to outspend.
How We Get There
Products fund research
While we work on the hard problems, we're building software that embodies our values: efficient, private, local-first. No cloud dependency. No wasted resources. No subscription traps.
Every product we ship funds continued research into next-generation data encoding.
Local-First Architecture
Your data stays on your machine. No cloud servers burning energy to store what could live locally.
No Subscription Lock-in
Buy once, own forever. MIT-licensed binaries. We don't hold your workflow hostage.
Built With
Rust · Tauri · Svelte
Native performance. Cross-platform. No Electron bloat. Our desktop and mobile applications are built with the same efficiency principles we're researching.
The future of computing isn't bigger data centres.
It's smarter data.