Can I run a model language on a 26-year-old console?

Name: Can I run a model language on a 26-year-old console?
Availability: InStock
Author: xaskasdf

by xaskasdf·Mar 21, 2026·46 points·12 comments

Visit Project View on HN

AI Analysis

●●●●GemWizardryZero to OneBig Brain

Streams LLM weights from CD-ROM during inference to fit 77MB models in 32MB RAM.

Strengths

•Weight streaming from CD solves the 32MB RAM constraint elegantly
•Custom PSNT quantized format built from scratch for PS2 architecture
•Runs on real PS2 hardware, not emulation or simulation

Weaknesses

•10M parameter model limits practical utility beyond demonstration
•Requires custom SDK and build pipeline to compile and deploy

Post Description

Short answer: yes.

The Emotion Engine has 32 MB of RAM total, so the trick is streaming weights from CD-ROM one matrix at a time during the forward pass — only activations, KV cache and embeddings live in RAM. This means models bigger than the RAM can still run, they just read more from disc.

Had to build a custom quantized format (PSNT), hack endianness, write a tokenizer pipeline, and most of the PS2 SDK from scratch (releasing that separately). The model itself is also custom — a 10M param Llama-style architecture I trained specifically for this.

And it works. On real hardware.