A Modern Self-Referential Weight Matrix That Learns to Modify Itself · HackerPapers discussions

A Modern Self-Referential Weight Matrix That Learns to Modify Itself — mentioned by logicchains

Wed, 18 Mar 2026 09:07:30 GMT

There's already a model capable of autonomous learning on the small scale, just nobody's tried to scale it up yet: https://arxiv.org/abs/2202.05780

Paper and discussion history. A mention is not an endorsement.

A Modern Self-Referential Weight Matrix That Learns to Modify Itself — mentioned by logicchains

Sat, 07 Feb 2026 08:01:04 GMT

Schmidhuber solved it at a small scale: https://arxiv.org/abs/2202.05780 .

Paper and discussion history. A mention is not an endorsement.

A Modern Self-Referential Weight Matrix That Learns to Modify Itself — mentioned by logicchains

Wed, 15 Jan 2025 07:54:53 GMT

>Contextual real-time weight modification is definitely one of the breakthroughs required for AGI. It's already been invented: https://arxiv.org/abs/2202.05780 . That design is just very inefficient to scale up / use as a transformer backbone.

Paper and discussion history. A mention is not an endorsement.

A Modern Self-Referential Weight Matrix That Learns to Modify Itself — mentioned by logicchains

Thu, 14 Dec 2023 19:20:33 GMT

>atrophying his skills too much to do anything now that the industry is moving so quikcly. His recent papers are still cutting edge. He's already solved self-improving AI: https://arxiv.org/abs/2202.05780 , it just needs to be scaled up.

Paper and discussion history. A mention is not an endorsement.

A Modern Self-Referential Weight Matrix That Learns to Modify Itself — HN submission

Wed, 13 Apr 2022 16:12:30 GMT

A modern self-referential weight matrix that learns to modify itself

166 points; 46 discussion replies on HN. Paper and source history. A submission is not an endorsement.

A Modern Self-Referential Weight Matrix That Learns to Modify Itself — HN submission

Thu, 24 Feb 2022 01:05:59 GMT

A Modern Self-Referential Weight Matrix That Learns To Modify Itself

1 point; 0 discussion replies on HN. Paper and source history. A submission is not an endorsement.