Can LLMs redesign themselves as they train?
doubleAI · Project update · November 6, 2026
doubleAI published research exploring whether large language models can autonomously redesign their own architectures during the training process.
AntSpace’s summary of doubleAI’s own announcement: read the original.