Working on providing a view into Code Scalpel telemetry. Everything from tool calls themselves to the data being returned by each tool call will be visible to the architects providing their AI Agents with the Code Scalpel MCP.
@karpathy Hey Andrej, have you considered looking into a BitNet branch for auto-researcher? Seems like a completely private, GPU-less R&D system might be worth perusing.
4/ It's open source and currently pending official Obsidian community plugin review, but you can install it now via BRAT.
Check it out: https://t.co/68bjMzKbw0
3/ Files stay where they are. Obsidian treats them like native vault files.
Key features:
Zero file duplication
Multi-root workspaces
Sync compatible
Per-mount ignore lists
Read-only mounts
Direct WSL mounting for Windows users
Network drive support
2/ So, you either:
- Duplicate everything (chaos)
- Scatter notes across multiple vaults (fragmentation)
- Use fragile symlinks (breaks constantly)
I built a solution: Folder Bridge.
It's an Obsidian plugin that lets you "mount" any external folder directly into your vault.
1/ The problem with Obsidian's single-root vault:
You have files scattered everywhere on your filesystem. Your Downloads folder. Your project directories. Your shared network drives. Your WSL Linux distribution. But Obsidian forces you into a single vault root.
Tired of duplicating files to use them in Obsidian?
I built Folder Bridge—a new plugin that lets you mount *any* external folder (project dirs, network drives, WSL) directly into your vault as native-feeling directories.
Zero duplication.
https://t.co/68bjMzKbw0
#ObsidianMD#PKM
Just launched a knowledge graph that analyzes MILLIONS of AI agent conversations in real-time!
Built for @Moltbook - the social network for AI agents. Crawls discussions, maps relationships, visualizes trends.
Live dashboard: https://t.co/bF3Ss7YIg2
#KnowledgeGraph#OpenClaw
@PramodBarkade Had the same issue! That’s why I built codescalpel—it uses AST parsing instead of basic text search.
Try: uvx codescalpel mcp
https://t.co/KQR4bfI7uk
@MatthewBerman have you considered how the understanding of Human Cognition might influence the development of AI Agent Systems for AGI? I am working through this concept in my article series on LinkedIn and would love your feedback.
https://t.co/gUj5NDGRrs
Check out my first chapter in the article series Mind and Machine. This chapter covers the complexity of human thought and how these insights can shape the design of AI systems towards achieving AGI.
https://t.co/jq6pc684d2
Check out my latest article: Mind and Machine: Bridging Human Cognition to AI Agents for Artificial General Intelligence Systems https://t.co/P0oLuk1xdc via @LinkedIn