X

@poteto

Grok Bot last week: asks people what they got done with Grok Bot last week. post (https://x.com/poteto/status/2112246281016119780) · 2,180❤ · 151 RT

You Can Ship Everything: posts the line "You Can Ship Everything", followed in-thread by "but should you?". post (https://x.com/poteto/status/2112356481597317544) · 2,056❤ · 79 RT

Action Button: quotes the Grok Bot announcement that it can be set as the iPhone Action Button and pressed to talk with the bot. post (https://x.com/poteto/status/2112261641505231330) · 1,608❤ · 47 RT

@rauchg

Open-source HTML artifact builder: Vercel open sourced a fast HTML artifact builder (fx0.vercel.dev) powered by libfx, which makes the agent durable against server crashes, flaky LLMs, and function limits, the plumbing v0 needed. post (https://x.com/rauchg/status/2112247046082363821) · 1,879❤ · 114 RT

Mercury 3 speed: quotes Mercury 3 from Inception going live on AI Gateway at a 50% discount (diffusion model, 1M-token context, tool calling) and notes it runs at 1,000 tokens per second. post (https://x.com/rauchg/status/2112256718281551955) · 267❤ · 11 RT

More free AI: quotes MiniMax M3.1 Flash going live on AI Gateway for free through Oct 22 when served by MiniMax. post (https://x.com/rauchg/status/2112194000639967590) · 294❤ · 9 RT

@hwchase17

Vertical vs horizontal agents: from a talk with Harry Stebbings, argues vertical agents differentiate through context (prompts, skills, tools), harness, and model, in increasing difficulty; most teams still tweak context then harness, and should revisit all three every 3 to 6 months as base models change. post (https://x.com/hwchase17/status/2112471992435667163) · 30❤ · 7 RT

LangChain x kapa: endorses the kapa integration that gives LangChain agents a single company-knowledge index (docs, GitHub, Notion, Slack, Zendesk) with search and read-doc tools. post (https://x.com/hwchase17/status/2112179638802006194) · 54❤ · 4 RT

Harness first: agrees with starting from a harness such as deepagents, adding context with skills and tools, then middleware, and only then a LangGraph workflow if needed. post (https://x.com/hwchase17/status/2112489626061738209) · 1❤ · 0 RT

@swyx

End of chat: replies to poteto's question on whether chat is the final form of agent UI with a Gita-style dialogue arguing people keep announcing the end of chat while chasing what is new rather than what is true. post (https://x.com/swyx/status/2112182228977348963) · 60❤ · 1 RT

Silent

@martinfowler, @cursor_ai, @karpathy, @simonw, @HamelHusain: no original posts in window

Gaps

@xAI: API not authorized to read this account's posts (skipped)

GitHub trending

MadsLorentzen/ai-job-search (https://github.com/MadsLorentzen/ai-job-search): local AI job application framework built on Claude Code that evaluates postings, tailors CVs, writes cover letters, and preps interviews (Python, +3,856★ today)
anomalyco/opencode (https://github.com/anomalyco/opencode): the open source coding agent (TypeScript, +1,311★ today)
akitaonrails/ai-memory (https://github.com/akitaonrails/ai-memory): long-term memory for coding-agent CLIs and handoff between different agent vendors (Rust, +666★ today)
Comfy-Org/ComfyUI (https://github.com/Comfy-Org/ComfyUI): modular diffusion model GUI, API, and backend with a graph and nodes interface (Python, +1,011★ today)
ruvnet/ruflo (https://github.com/ruvnet/ruflo): agent harness for multi-agent swarms and autonomous workflows with adaptive memory, vector RAG, and Claude Code and Codex integration (TypeScript, +564★ today)