context

81 stories and discussions about context, aggregated from every source we track.

1.

We introduce Context Language Models (CLMs), language models that natively manage their own context. We implement this by treating the context as a file and allowing the model to make unrestricted updates to this file.…

163 points•emersonmacro•about 2 months ago•44 comments•
2.

A collaborative AI workspace, built on your company context. Build and orchestrate agents right alongside your team's projects, meetings, and connected apps.

149 points•erichocean•about 2 months ago•46 comments•
3.

<p>Additional context: <a href="https://www.youtube.com/watch?v=M1si1y5lvkk" rel="ugc">https://www.youtube.com/watch?v=M1si1y5lvkk</a></p> <p>No abstract.</p>

129 points•theelx•7 months ago•64 comments
4.

We’re on a journey to advance and democratize artificial intelligence through open source and open science.

76 points•victormustar•about 2 months ago•7 comments•
6.

<p>Context: I'm the author of the Rust patch</p>

25 points•ChrisDenton•over 2 years ago•21 comments
7.

We keep hearing the same advice: Give the AI more context. Add the README. Add AGENTS.md. Add...

16 points•robertadam987_•about 2 months ago•8 comments
8.

This is a submission for the Sanity Challenge, Path One: Ship an Agent That Queries Real...

15 points•shreyansh_agrahari_2009db•about 2 months ago•2 comments
9.

Most engineering teams working on long-context agents hit the same billing wall around turn twenty. A...

12 points•reidmarlow•about 2 months ago•18 comments
11.

Open Model Context Protocol (omcp) — a community-driven hard fork. Build without permission. - enclawed/omcp

8 points•enclawed•2 months ago•1 comment•
12.

Eliminating the widget builder tax, closure fatigue, and the context.watch trap in Flutter: how 1:1 symmetry between containers and BuildContext unlocks cleaner, faster reactive apps.

8 points•randalschwartz•2 months ago•1 comment
13.

Message your team and AI agents in one place. Praxos lets people and agents talk, share work, and get things done together.

6 points•mogusian•2 months ago•0 comments•
14.

High-assurance in-memory Tree-Sitter AST context firewall and pruning MCP server for coding agents (-72.4% token mass). - heuristicolab/ctxfw

5 points•mikemo88•about 2 months ago•0 comments•
15.

Having to manually copy-paste context between your code environments (like IDEs) and Claude Desktop,...

5 points•techwithsam•about 2 months ago•1 comment
16.

This is a submission for the Sanity Challenge, Path Two: Build a Knowledge Base or Context...

5 points•rajan_mishra_a9f78ad216b4•about 2 months ago•0 comments
17.

Model Context Protocol has changed the way AI applications interact with external tools. Instead of...

5 points•therealmrmumba•2 months ago•0 comments
18.
19.

Autonomous flight simulator for AI agents — co-evolutionary adversarial red-teaming with 3D MAP-Elites quality-diversity. Discovers zero-days in LLM agents before production. MCP-native. - zariffro...

4 points•zariflatif•about 2 months ago•0 comments•
20.

Build an AGENTS.md file for your project from project context, focus areas, technology stack, custom instructions and working preferences.

4 points•ecce_homo•about 2 months ago•1 comment•
21.

Context, limitations, and evaluation criteria for OliverDB performance results.

4 points•pnr-consulting•2 months ago•0 comments•
22.

Technology, startups, programming, technical management and software architecture

4 points•felixc•over 3 years ago•0 comments
23.

Family AI assistant that runs on Claude Code and speaks Matrix - dmitry-markin/silta

3 points•dmitry-markin•about 2 months ago•0 comments•
24.

We introduce Context Language Models (CLMs), language models that natively manage their own context. We implement this by treating the context as a file and allowing the model to make unrestricted updates to this file.…

3 points•tomatomatomato•about 2 months ago•1 comment•
25.

How a tree of business domains makes analytics context easier to debug, maintain, and retrieve without overloading the agent.

3 points•aloys_hn•about 2 months ago•0 comments•
26.

Most developers still treat prompt injection as a leakage problem. Someone types an adversarial...

3 points•reidmarlow•about 2 months ago•1 comment
27.

Model-requested semantic context compaction for Codex, with complete checkpoints and no per-turn cap. - manuelcecchetto/codex-context-gc

3 points•manuelcecchetto•2 months ago•1 comment•
28.

At some point in growing a large-scale software system, you’ll require “out of band” context: data which is not explicitly passed as an argument to a function but rather implicitly attached to the request, job or event…

3 points•ebababi•over 4 years ago•2 comments
29.

The KV Cache Is the New Memory Wall: Three Regimes and Five Domains of Long-Context LLM Inference TL;DR — Long-context LLM decode inference is bound by memory bandwidth,…

2 points•acossta•about 2 months ago•0 comments•
30.

How a context engine works, stage by stage: the 5-stage architecture behind agents that ship mergeable code across 200+ developers and multi-repo orgs.

2 points•ijidak•about 2 months ago•0 comments•
31.

Up to 8.3 times cheaper inference, 6 times more context, and 10 times the speed.

2 points•quantbagel•about 2 months ago•1 comment•
32.

A Blog post by David Corvoysier on Hugging Face

2 points•dacorvo•about 2 months ago•0 comments•
33.
2 points•simonpure•about 2 months ago•0 comments•
34.

This author has chosen to make their posts visible only to people who are signed in.

2 points•LeoPanthera•about 2 months ago•0 comments•
35.

For the folks who found this blog from Anthropic&#8217;s site, welcome! For everyone else, I&#8217;d better give some context. Last month, I had a blog post titled &#8220;It Only Counts When AI Get…

2 points•Hbruz0•about 2 months ago•0 comments•
36.

Compare GPT-6 Sol and Luna pricing, context limits, reasoning, and use cases to choose a model for coding or high-volume applications.

2 points•flashbrew•about 2 months ago•0 comments•
37.

🛠️ toolhub - Turn ANY console script into an LLM tool in 2 seconds. Lightweight, self-hosted, tree-structured alternative to MCP & LangChain. - Talos-popcorn/toolhub

2 points•soidkeo•2 months ago•0 comments•
38.

Here's a chat someone thought you'd want to see.

2 points•GodelNumbering•2 months ago•2 comments•
39.

A semantic layer is a governed catalog of definitions that sits on top of your existing data and tells agents and people what to trust. Here's how we built one into PostHog's context warehouse.

2 points•eustoria•2 months ago•0 comments•
40.
1 points•simianwords•about 2 months ago•0 comments•
41.

Gobstopper makes long Claude Code and Codex sessions smaller. A local proxy compacts live requests, and saved sessions get a smaller copy beside the original.

1 points•benzguo•about 2 months ago•2 comments•
42.

A free Mac app that turns meetings, dictation, and writing into local files your AI can use. It transcribes on your Mac, no bot joins your calls, and it's open source.

1 points•justinbetker•about 2 months ago•0 comments•
43.

Try out Artifacts created by Claude users

1 points•Anon84•about 2 months ago•0 comments•
44.

AI writes code better than any of us now. What still has value: context, product feel, initiative and clear communication.

1 points•montenegrohugo•about 2 months ago•0 comments•
45.

A transparent agent runtime in Rust: context, tools, permissions and requests as explicit state. Plus an MCP bridge and a terminal agent. - ljedrz/nachalnik

1 points•ljedrz•about 2 months ago•0 comments•
46.

Podcast Episode · Latent Space: The AI Engineer Podcast · September 29 · 1h 32m

1 points•adamdaniel•about 2 months ago•1 comment•
47.

An enterprise Model Context Protocol (MCP) gateway adds shadow MCP discovery, IdP-driven access, tool-call security, and audit reporting for large companies.

1 points•ritzaco•about 2 months ago•0 comments•
48.

AI context engineering and agentic commerce for LLMs & AI applications — RAG Chunk Inspector, Memory Architecture Visualizer, Agent Trace Inspector, and Agent Commerce Gate for x402 + ERC-8004 pre-payment checks.

1 points•contextiq•about 2 months ago•0 comments•
49.

What's new in Nola 0.1.14: context statements, colon types for asks, function calls without the dots, a prompt you can read and a new agent-loop starter.

1 points•emykhailenko•about 2 months ago•0 comments•
50.

Heimr 570M is a 2.7B-parameter, 570M-active hybrid of Mamba-3 and Heimr attention with mixture-of-experts layers, built for phones and laptops. Its decode throughput and memory stay nearly flat to 64K tokens of…

1 points•ilreb•about 2 months ago•0 comments•
51.

Jev reads the message, Gradium Voice Design gives the agent a voice that fits. Every step on the critical path is traced.

1 points•segu•about 2 months ago•0 comments•
52.

A lot of prior work addressed key-value (KV) cache selection and compression by sparse attention to enable long-context inference for transformer language models without excessive hardware budgets. We provide a new…

1 points•theanonymousone•about 2 months ago•0 comments•
53.

Read-only Postgres MCP server: a drop-in replacement for the archived server-postgres, with schema context for correct answers. - contextflo/postgres-mcp

1 points•depthfirst•about 2 months ago•1 comment•
54.

The first frontier LLM with ten million tokens of context.

1 points•ClintEhrlich•about 2 months ago•0 comments•
55.

LLM coding agents operate by constructing trajectories that accumulate reasoning, tool calls, and results to enable multi-step decision-making. However, the conventional append-only trajectory architecture found in…

1 points•matt_d•about 2 months ago•0 comments•
56.

Make decisions from the command line, in scripts, and in agent skills. No code required. - vsekhar/decide

1 points•vsekhar•about 2 months ago•0 comments•
57.
1 points•marowa-labs•about 2 months ago•0 comments•
58.

Why LLMs changed what we expect from monitoring, and why centralized, queryable context makes every alert investigable.

1 points•khazit•about 2 months ago•0 comments•
59.
1 points•jlnppk02•about 2 months ago•1 comment•
60.

Large language model (LLM) agents often handle streams of related tasks, yet standard harnesses repeatedly ask the model to reconstruct the same control decisions inside each task's context. We study whether task…

1 points•acossta•about 2 months ago•0 comments•

Related topics