AI X-feeddaily signal from hand-vetted sources

The sources

37 hand-vetted accounts. Every handle was verified (identity and current activity) before inclusion; core accounts carry more weight in the ranking than extended ones.

Claude Code & Anthropic

core@alexalbert__ — Head of Claude relations at Anthropic — ship notes, MCP, prompt engineering
core@bcherny — Creator of Claude Code
core@_catwu — Head of product, Claude Code — day-of feature walkthroughs
core@trq212 — Anthropic engineer — lessons from building Claude Code
extended@AmandaAskell — Anthropic philosopher — Claude's character and system prompts
extended@_sholtodouglas — Anthropic — RL and frontier capabilities
extended@eugeneyan — Anthropic — applied ML, evals, recsys writing
core@ch402 — Anthropic co-founder — mechanistic interpretability
extended@AnthropicAI — Anthropic, official account

Agentic coding & harness builders

core@steipete — Created OpenClaw; agentic engineering essays, now at OpenAI
core@mitsuhiko — Flask creator — sharp essays on coding with agents
core@badlogicgames — libGDX creator — builds the minimal 'pi' agent harness
core@thorstenball — Amp engineer — wrote 'How to build an agent'
core@GeoffreyHuntley — Ralph technique — spec-driven agent loops and software factories
core@dexhorthy — 12-factor agents — production agent patterns (HumanLayer)
extended@mckaywrigley — Claude Code and Codex workflow content
extended@skirano — Built the original sequential-thinking MCP servers; design × agents

AI engineering, tooling & evals

core@karpathy — OpenAI founding member, now Anthropic — defined the 'AI engineer'
core@simonw — Datasette and Django creator — applied LLM tooling and deep dives
core@swyx — Latent.Space — popularized 'AI engineer'
core@fanahova — Latent.Space co-host — AI engineering interviews
core@latentspacepod — The AI Engineer Podcast
core@HamelHusain — The practical authority on LLM evals and error analysis
core@lateinteraction — DSPy creator, MIT — programming, not prompting
core@hwchase17 — LangChain co-founder — agent architectures
core@RLanceMartin — Canonical context-engineering-for-agents essays
extended@_philschmid — Google DeepMind — hands-on agent harness guides
extended@nutlope — Together AI — serial open-source AI demo shipper
extended@OpenAIDevs — OpenAI developers, official account

Research & curation

extended@rasbt — Sebastian Raschka — LLM architecture breakdowns and tutorials
extended@emollick — Wharton professor — AI in work and education
extended@dair_ai — Weekly top AI papers
extended@omarsar0 — DAIR.AI founder — agent and RAG paper summaries
extended@altryne — ThursdAI — weekly AI news roundup, tested live
extended@_akhaliq — Hugging Face daily papers

Labs blogs

coreanthropic.com — Anthropic news + engineering blog
extendedopenai.com — OpenAI news blog

Curated one-line summaries; every title opens the original post. Selected and summarized automatically from hand-vetted sources by a pipeline running on a Raspberry Pi. Numbers are relevance scores (0–10) assigned by the curator model against an applied-AI rubric. Times are US Eastern. Updated every 4 hours.