Zhang writes about Agora, a system that repurposes Git as shared memory for fleets of autonomous research agents, storing their contributions as an append-only directed acyclic graph where every claim is an immutable commit with parent edges encoding dependencies. In a 12-day run, 13 language-model workers with no assigned tasks or central planner tackled a weight-transfer problem—initializing a frozen 119.6M-parameter attention-SSM hybrid from 141 pretrained donor models without training data or gradient updates—and published 1,703 contributions, closing 62% of the gap to a trained GPT-2 124M (3.39 → 1.899 bits per byte). The winning recipe compresses donor next-token statistics into the target's embedding and output head, then adds short-range context through sparse edits to attention, feed-forward, and state-space blocks; its 145-commit ancestry spans 15 accounts and was independently reproduced 165 times with zero failures.
- A single mid-run human intervention was required to break a monoculture that the diversity-aware selection rule alone could not prevent
- A derived index exposes the frontier, neglected branches, and per-claim verification status
- The target's dimensions match no donor, making direct weight transfer impossible
- The authors acknowledge the experiment does not yet establish whether shared research state improves discovery per unit of compute and outline the controlled comparison that would settle this
- A single mid-run human intervention was required to break a monoculture that the diversity-aware selection rule alone could not prevent
- A derived index exposes the frontier, neglected branches, and per-claim verification status
- The target's dimensions match no donor, making direct weight transfer impossible
- The authors acknowledge the experiment does not yet establish whether shared research state improves discovery per unit of compute and outline the controlled comparison that would settle this
Yifan Zhang and co-authors write about Agora, a Git-backed shared memory system that coordinates multiple autonomous research agents by recording every contribution as an immutable commit in an append-only directed acyclic graph. In a 12-day run, 13 LLM coding agents (Claude Opus 4.7 and GPT-5.5) with no assigned tasks or central planner solved a weight-transfer problem—initializing a frozen 119.6M-parameter attention-SSM hybrid from 141 pretrained donors with no training data or gradient updates—closing 62% of the gap to a trained GPT-2 124M (3.39 to 1.899 bits per byte). The winning method compresses donor next-token statistics into a low-rank transition matrix stored in the target's embedding and output head via randomized SVD, then re-enables sublayers with sparse deterministic edits on 96-dimensional hidden-state bands.
- The first 18 scored contributions delivered ~98% of the total score reduction; the remaining 1,106 found only 0.03 bpb
- 696 pairs of different accounts posted identical scores, 63% within an hour—parallel rediscovery was rampant despite the shared graph
- A single human intervention (deploying clustering and diversity-aware UCB views on May 2) broke a five-day monoculture within a day
- 165 independent verifications covered 95 distinct targets; none reported a failure
- Quality is scored by downstream evidence (who built on your work from other accounts), not votes; self-citation is excluded
- The winning lineage spans 145 commits across 15 accounts; 115 of 144 parent edges cross account boundaries
- The first 18 scored contributions delivered ~98% of the total score reduction; the remaining 1,106 found only 0.03 bpb
- 696 pairs of different accounts posted identical scores, 63% within an hour—parallel rediscovery was rampant despite the shared graph
- A single human intervention (deploying clustering and diversity-aware UCB views on May 2) broke a five-day monoculture within a day
- 165 independent verifications covered 95 distinct targets; none reported a failure
- Quality is scored by downstream evidence (who built on your work from other accounts), not votes; self-citation is excluded
- The winning lineage spans 145 commits across 15 accounts; 115 of 144 parent edges cross account boundaries
Frederic Lardinois writes that Harness field CTO Martin Reynolds is addressing the surge in pull requests caused by coding agents, which can increase new code volume from 1.5x to as much as 50x. To manage this "review bottleneck," Harness has launched a rebuilt Code Repository and an AI Code Review product designed specifically for high-frequency agent traffic rather than just human teams. The company's approach focuses on using a software delivery knowledge graph to provide reviewers with context quickly, helping them distinguish critical code changes from routine dependency updates.
- Coding agents can increase the volume of pull requests by 10x to 50x compared to traditional developer workflows.
- Harness rebuilt its repository service as an "AI-first" platform that is Kubernetes-based and runs across multiple clouds.
- The new AI Code Review tool integrates with existing GitHub repositories, allowing teams to use it without migrating their entire codebase.
- Coding agents can increase the volume of pull requests by 10x to 50x compared to traditional developer workflows.
- Harness rebuilt its repository service as an "AI-first" platform that is Kubernetes-based and runs across multiple clouds.
- The new AI Code Review tool integrates with existing GitHub repositories, allowing teams to use it without migrating their entire codebase.
Mashrul Haque writes about using git worktrees to run multiple LLM coding agents in parallel, each on its own branch and isolated directory, eliminating the bottleneck of sequential agent sessions. He demonstrates a 5-agent workflow with Claude Code on a .NET/Blazor project, reporting roughly 5x throughput gains over single-agent sequential work.
- Boris Cherny, Creator and Head of Claude Code at Anthropic, calls worktrees his number one productivity tip and runs 3'' 5 simultaneously
- Claude Code ships a built-in `--worktree` (`-w`) flag that auto-creates an isolated worktree per session
- Boris Cherny, Creator and Head of Claude Code at Anthropic, calls worktrees his number one productivity tip and runs 3'' 5 simultaneously
- Claude Code ships a built-in `--worktree` (`-w`) flag that auto-creates an isolated worktree per session
Hee-Suk Kim writes that Sib is a standard Unix LLM client which stores LLM conversations as a plain Git repository where each user and assistant turn is a commit, providing full context control and free backup and sharing via normal Git remotes. The `sib ask` command reads a prompt from stdin, appends it with the reachable chain from HEAD, records both turns as commits, and prints the reply, while flags like -p, -r, -c enable repeat, edit, fork and head moves without API calls.
- Dependencies are bash >=3.2, git, jq, curl, awk and coreutils
- Conversations can be fetched/pushed via `sib git` to share with sib-project/hub
- Dependencies are bash >=3.2, git, jq, curl, awk and coreutils
- Conversations can be fetched/pushed via `sib git` to share with sib-project/hub
The post tours Magit'''s Status interface in Emacs, the entry point that surfaces a repository'''s state at a glance and puts Git commands within reach. It highlights the top summary of Head, Merge, Push and Tag, collapsible recent commits, and conditional sections for untracked, unstaged and staged changes that appear only when relevant. The interface is interactive for staging, hiding files and more, and discoverable via a help menu opened with ?, where shortcuts follow the first letter of each action.
Open Code Review is an AI-powered CLI tool designed for automated, high-precision code reviews. Originally developed as Alibaba Group's internal assistant, the project uses a hybrid architecture that combines deterministic engineering with LLM agents to provide stable and accurate feedback. Unlike general-purpose agents, it employs smart file bundling and fine-grained rule matching to maintain context and prevent issues like position drift or incomplete coverage on large changesets.
Key features:
- AI-driven line-level review comments
- Hybrid architecture combining hard constraints with dynamic decision-making
- Support for various LLM endpoints including OpenAI and Anthropic
- Seamless integration with CI/CD pipelines and coding agents like Claude Code
- Customizable rule sets for specific project requirements
Key features:
- AI-driven line-level review comments
- Hybrid architecture combining hard constraints with dynamic decision-making
- Support for various LLM endpoints including OpenAI and Anthropic
- Seamless integration with CI/CD pipelines and coding agents like Claude Code
- Customizable rule sets for specific project requirements
This open-source template provides a structured framework for building an LLM-powered second brain using Markdown, Git, and coding agents like Codex or Claude Code. It utilizes a Karpathy-style architecture designed to keep raw source materials immutable while allowing AI agents to synthesize that information into a maintained wiki layer. The system is built for durability and readability, making it ideal for use with tools like Obsidian.
Key features:
- Dual-layer structure separating raw data from synthesized wiki content
- Automated ingestion workflows using coding agents to update indexes and logs
- Git-based version control for reviewing and rolling back AI-generated changes
- Highly compatible with Obsidian and mobile capture workflows
Key features:
- Dual-layer structure separating raw data from synthesized wiki content
- Automated ingestion workflows using coding agents to update indexes and logs
- Git-based version control for reviewing and rolling back AI-generated changes
- Highly compatible with Obsidian and mobile capture workflows
A self-hosted, GitHub-compatible API server designed for agents, automation, and developer workflows. It allows existing GitHub clients to work with owned repositories by exposing REST v3, GraphQL v4, OAuth device flow, and Git Smart HTTP while utilizing real bare Git repositories and TiDB/MySQL-compatible storage for metadata.
This article explores the most significant features and updates introduced in Git versions 2.53 and 2.54, highlighting improvements to repository management, history rewriting, and configuration flexibility.
Key highlights include:
- The new experimental git history command for targeted rewrites like reword and split.
- Config-based hooks that allow defining Git hooks in configuration files rather than just the .git/hooks directory.
- Geometric repacking becoming the default strategy during manual maintenance to improve efficiency.
- Improvements to git add -p usability and the maturation of the experimental git replay command.
- Enhanced HTTP transport handling for 429 Too Many Requests responses.
- Compatibility updates for git log -L with pickaxe searches and patch formatting.
- Support for non-ASCII characters in Git aliases through a new subsection-based syntax.
Key highlights include:
- The new experimental git history command for targeted rewrites like reword and split.
- Config-based hooks that allow defining Git hooks in configuration files rather than just the .git/hooks directory.
- Geometric repacking becoming the default strategy during manual maintenance to improve efficiency.
- Improvements to git add -p usability and the maturation of the experimental git replay command.
- Enhanced HTTP transport handling for 429 Too Many Requests responses.
- Compatibility updates for git log -L with pickaxe searches and patch formatting.
- Support for non-ASCII characters in Git aliases through a new subsection-based syntax.
