context

The ASRN Adaptive Sparse Recurrence Network introduces a copy layer for language models that looks up earlier occurrences of the current context via learned hash tables and reproduces what followed, achieving memory usage linear in sequence length as a cheaper alternative to full attention.

A developer (ex-Postman) built Sannr, a headless API client designed for coding agents like Claude Code and Codex that quietly builds and persists API context locally, so agents stop relearning the same edge cases in every session.

A rap song humorously portrays 'Grok Bot' as an AI agent that runs on context, handles tasks, and uses tools for automation, blending tech themes with creative expression.

Click is a product that provides live research context for ChatGPT and Claude, enhancing AI responses with real-time information.

A thought piece arguing that as AI agents take over operational work, companies will shift from managing people to managing context — the shared data, SOPs, and decision logic that forms the company's real competitive advantage.

Gergely Orosz notes a growing trend of apps and companies positioning themselves to gather work context in a single space, addressing the problem of scattered tools and documents.

Hexis is a product that provides Git-backed skills, tools, and context for AI agents, enabling teams to manage agent capabilities as versioned code.

A Twitter thread argues that the key to powerful AI agents is not better prompts but accumulated personal context and memory systems, highlighting Obsidian as a tool for compounding knowledge. The author predicts a widening gap between those using AI alone and those combining it with personal context.

This paper reveals that while large language models appear robust to task-irrelevant context at the aggregate level, their predictions can flip on individual examples, with performance degrading on some and improving on others, highlighting tail risks that aggregate accuracy conceals.

A discussion on how AI agents learn in three layers (model, harness, context) and why capturing user corrections as labeled examples is key to building self-learning agents that improve over time.

A developer describes building a multi-agent voice social-deduction game, solving turn-taking with a central conductor but struggling with shared memory and preserving social subtext when compressing conversation history into structured state.

Explores whether analytics agents should incorporate contextual data from tools like Linear, Sentry, and Notion, or remain purely metrics-driven.

An opinion piece arguing that AI agents in enterprises need a structured 'Company Brain' memory layer to reliably access context, policies, and permissions, rather than relying solely on RAG and tool access.

Zaro allows users to build agents and apps on top of their context with a single prompt.

The author shares lessons from designing a testing framework in Guile, focusing on how adding context to test definitions makes tests more reusable and improves developer experience.

A discussion about the focus of AI evaluations, questioning whether practitioners are optimizing prompts, context, or the entire harness, and noting a shift toward holistic optimization.

This is the sixth article in the series, explaining in detail the concept of subagent, its working principles, and its role in coding agents, including tool call and runtime mechanisms, as well as the applicable scenarios of different subagent types (fresh child, forked child, partial fork).

The article argues that coding agents need continuity—preserving execution history and project state in the repository—rather than simply larger memory or context windows, to avoid losing the operational thread between sessions.

Google announced the Open Knowledge Format, an open standard based on Karpathy's LLM wiki concept, designed to provide context for AI agents using simple markdown files.

Google's Open Knowledge Format (OKF) proposes a portable standard for organizational knowledge to help AI agents retrieve correct context, addressing fragmentation across data catalogs, wikis, and code.