---
format: "aidr-story-markdown/v1"
id: "acf865c168dd2f965b28f9000f0a87e271cd88ea48c664c762ea23cdd499a5fa"
canonical_url: "https://aidr.today/acf865c1?lang=en"
title: "Sakana AI Trains 1,000 Layer Neural Nets Without Backpropagation Using PC-ALM"
lang: "en"
requested_lang: "en"
available_langs: ["en","vi"]
translation_fallback: null
fallback_fields: []
published_at: "2026-09-14T15:20:35.000Z"
category: "Research"
topics: ["sakana-ai","llm","research","fine-tuning","inference","neural-network","backpropagation","training"]
source_urls: ["https://huggingnews.com/ai/sakana-ai-trains-1000-layer-neural-nets-without-backpropagation-using-pc-1e1de59d","https://marketbrief.now/ai/sakana-ai-trains-1000-layer-neural-nets-without-backpropagation-using-pc-1e1de59d"]
summary: "Local dynamics and distributed optimization can now be used to manage credit assignment in artificial neural networks. Sakana AI"
---

# Sakana AI Trains 1,000 Layer Neural Nets Without Backpropagation Using PC\-ALM

> [Open the canonical story](<https://aidr.today/acf865c1?lang=en>)

**Published:** 2026-09-14T15:20:35.000Z
**Category:** Research
**Topics:** sakana\-ai, llm, research, fine\-tuning, inference, neural\-network, backpropagation, training

## Summary

Local dynamics and distributed optimization can now be used to manage credit assignment in artificial neural networks\. Sakana AI

## Sources

- [Story source](<https://huggingnews.com/ai/sakana-ai-trains-1000-layer-neural-nets-without-backpropagation-using-pc-1e1de59d>)
- [Story source](<https://marketbrief.now/ai/sakana-ai-trains-1000-layer-neural-nets-without-backpropagation-using-pc-1e1de59d>)