New Runtime

An atlas of agent-shaped software.

We track the systems, signals, and shifts as agents become first-class participants in how software gets work done.

New Runtime observatory map connecting projects, signals, sources, agents, patterns, and field notes.
Raw Signals639live
Patterns15active
Topics79tracked
Field Notes154notes
Sources1331connected
Agent Routes903exposed

Start here today

All Field Notes
  1. 01
    Read firstA Balanced MoE Router Can Still Be Functionally Dead

    Cerebras demonstrates how top-k normalization can erase the cross-entropy gradient to an MoE router even while expert utilization appears perfectly balanced.

    Aug 3, 2026 · 1 sources
  2. 02
    Fresh signalVercel Passport is generally available

    This X-discovered record adds fresh evidence to the model answers become task interfaces, ai native orgs move to review and orchestration lens and lets the site trend graph move as public product and research signals arrive.

    Jul 31, 2026 · notable
  3. 03
    Track patternHarness architecture outlives model choice

    Rules, tools, permissions, memory, evaluation, execution, and recovery are becoming a stable runtime layer while models remain replaceable components.

    6 signals · high
  4. 04
    Build pathGitHub Projects for AI Engineers

    Turn the Telegram export and two public ecosystem references into a ranked repository radar and a navigable map of the wider production AI stack.

    B2 · One local content pass plus a GitHub metric refresh
Post-app observatory

Software is becoming less like a destination and more like a temporary workspace.

Each record ties a capability, an old ritual, and a build path back to public evidence.

A Telegram-corpus-derived GitHub radar plus a source-backed map of the models, orchestration, retrieval, memory, security, automation, and infrastructure layers around production AI systems.

githubai-engineeringagent-runtimecontext-engineering
Status
candidate
Buildability
B2
Signals
238
Patterns
5
Last update
Aug 15, 2026
Recent activity

Outcome: Turn the Telegram export and two public ecosystem references into a ranked repository radar and a navigable map of the wider production AI stack.

Bottleneck: GitHub metrics drift quickly; every refresh needs a current GitHub read and a relevance pass against the corpus.

Open project

Machine routes

Programmatic access to the atlas.

View all routes
GET/openapi.jsonTyped JSON and MCP API contractJSONContractLive
GET/.well-known/mcpResources-only MCP server cardJSONDiscoveryLive
GET/llms-full.txtExpanded reading and citation guideTextGuideLive
GET/machine-consumption.jsonSanitized machine readership snapshotJSONTelemetryLive
GET/signals.jsonCompact raw signal indexJSONStaticLive
GET/patterns.jsonEvidence-linked pattern indexJSONStaticLive
GET/topics.jsonTopic hub indexJSONStaticLive

all work and no play make Jack a dull boy

Who read this page?Machine requests, hidden until opened

Loading the privacy-safe route aggregate…

Open the JSON contract