Creează un cont pentru a te abona la acest blog și a-i urmări articolele pe pagina Lectură.
-
A paper and reproducible study asking whether a sixteen-persona LLM panel preserves treatment responses in repeated games, and showing that passing coarse marginal checks is cheap.
-
A mobile relay for coding agents that keeps Claude Code and Codex running on your computer while you check in, answer questions, and approve actions from your phone.
-
An article on putting the repository at the center of the agent stack.
-
An open-source graph-based presentation, where each node carries the slides it belongs to and the graph pulls nodes in and pushes them out as the talk advances.
-
An article on the hippocampus and cortex split as a model for long-running agents: a frozen model is a cortex, and the hard part is the organ that decides what it learns.
-
A mathematical explanation of how information sharing without action coordination can worsen group outcomes.
-
An auditable, held-out-gated self-improvement loop demonstrated on LongMemEval.
-
An interactive site for learning ActiveGraph concepts through hands-on examples.
-
An open-source reference coding agent built on ActiveGraph.
-
Experiments comparing deterministic, semantic, and hybrid architectures for auditable agent memory.
-
An arXiv paper introducing event-sourced reactive graphs for auditable and forkable agentic systems.
-
An event-sourced reactive graph runtime for auditable, forkable, long-running agents.
-
A technical history comparing nine iterations of BabyAGI across three years.
-
A live AI pitch agent that interviews founders and VCs and extracts structured startup data.
-
An AI-generated sea shanty for Claude Code lovers.
-
A minimal autonomous assistant.
-
A daily prompting game.