Updates
CiteClaw: Reference Infrastructure for Agentic Research
Blog entry introducing CiteClaw as reference infrastructure for agentic research, with emphasis on citation tools, MCP access, Zotero state, and the bibliography curation skill.
On the Scaling of PEFT: Towards Million Personal Models of Trillion Parameters
Publication entry added for a study reframing PEFT as a substrate for persistent personal models scaled up, down, and out.
Macaron-V1-Preview
Original English repost of Mind Lab's Macaron-V1-Preview release for the 749B Mixture-of-LoRA Agent Model post-trained from GLM5.1.
Agent-Own Loop Harness
Blog entry on turning OpenSpec from a human-attended workflow into an agent-owned SDD loop.
delta-mem: Efficient Online Memory for Large Language Models
Publication entry added for delta-mem, an efficient online memory approach for large language models.
MinT: Managed Infrastructure for Training and Serving Millions of LLMs
Publication entry added for MinT, a managed infrastructure system for training and serving large numbers of LLMs.
Harness Engineering
Blog entry framing harnesses as the operating shell around agentic systems.
Triquetra for LoRAs
Blog entry on how LoRA rank, alpha, and learning rate move together when update scale changes.