Code4Func @Code4Func

Docker logging and monitoring architecture

this is f*cking gold THE AI guy dropped a free masterclass on harness engineering if I had this a year ago, I would've worked 5x faster in the right hands, this changes everything:

Rahul @sairahul1

a day ago

x.com/i/article/2063…

37 132 766 817K 3K

49 Product Manager Skills for AI Agents github.com/deanpeters/Pro…

Most AI/ML books are only useful if they change how you build. Designing AI Systems is useful because it maps the topic to engineering work you actually have to operate. The book covers: • Prepare reliable data and catch quality issues early • Add memory and context beyond one prompt • Use retrieval and company knowledge safely • Let AI systems use tools within clear rules • Route model requests and control cost • Evaluate output quality before production The production angle is the part I would pay attention to. The demo is the easy part. The useful engineering work is making the system reliable once it touches real workflows. Good fit for engineers building real AI systems and wanting a stronger mental model than another clean demo. Link in the first comment.

A harnessed LLM agent, clearly explained! Most people picture this as a model with tools bolted on. The real architecture inverts that relationship. The model itself is deliberately thin. Intelligence gets pushed outward, and the harness composes it at runtime. Three dimensions orbit the harness core: - 𝗠𝗲𝗺𝗼𝗿𝘆 holds the state a model shouldn't carry in weights or context. Working context, semantic knowledge, episodic experience, and personalized memory each have their own lifecycle. - 𝗦𝗸𝗶𝗹𝗹𝘀 hold procedural knowledge. This can cover operational procedures, decision heuristics, and normative constraints that specialize the general model per task. - 𝗣𝗿𝗼𝘁𝗼𝗰𝗼𝗹𝘀 hold the interaction contracts. Agent-to-user, agent-to-agent, and agent-to-tools are three distinct surfaces with their own failure modes. Between the core and these modules sit the mediators, like sandboxing, observability, compression, evaluation, approval loops, and sub-agent orchestration. They govern how the harness reaches out and how state flows back in. The useful question this framing unlocks is: for any new capability, where should it live? - Stable knowledge goes to memory - Learned playbooks go to skills - Communication contracts go to protocols - Loop governance goes to the mediators Harness design becomes a question of what to externalize, and how to mediate it. I'm building a minimal agent harness from scratch and will open-source it soon. In the meantime, my co-founder wrote an article about the anatomy of Agent Harness, covering the orchestration loop, tools, memory, context management, and everything else that transforms a stateless LLM into a capable agent. Read it below.

Akshay 🚀 @akshay_pachaar

2 months ago

x.com/i/article/2040…

68 489 3K 1.7M 7K

Anthropic CEO Dario Amodei: "The cheapest way to use Claude is also the smartest. Most devs do the exact opposite" this is one of the best interviews I've seen in a long time in this interview he breaks down exactly how a system changes everything: - the memory and context features that turn Claude into a second brain - the knowledge architecture most users don't know Claude can build - the integration layer that connects Claude to your actual workflow - why typing one question at a time is the most expensive way to use Claude if you've been using Claude for months and still start every conversation from scratch with zero context, you don't have a Claude problem. you have a system problem instead of another show tonight, watch this make sure to bookmark it before it gets lost in your feed full guide in the article below

Khairallah AL-Awady @eng_khairallah1

a week ago

x.com/i/article/2060…

33 53 351 622K 1K

How to master investing

@e_opore Helpful share

x.com/i/article/2059…

That's so true!

Stanford just proved Claude agrees with you 49% more than a human would The fix is the LLM Council from Andrei Karpathy. Five different advisers debate your decision inside one Claude chat: > contrarian (looks only for what will fail > first-principles thinker > expansionist (finds the upside) > outsider (knows nothing about your industry) > executor (Monday-morning next step) a chairman reads it all and gives you the final call.

SCOTTY BEAM @ScottyBeamIO

2 weeks ago

x.com/i/article/2059…

16 19 148 348K 449

Claude subscribers get a progress bar. This gives you the full picture. Claude Code usage dashboard (token counts, costs, sessions, projects) directly inside VS Code. Reads local JSONL transcripts, no API calls. Extension: Claude Code Usage by Paweł Huryn Install from VS Code marketplace: marketplace.visualstudio.com/items?itemName… See on GitHub: github.com/phuryn/claude-… Hope that helps.

Anthropic just officially released the blueprint for creating a company with Claude Code and it's mind-blowing😭 CEO: 1 human (who sleeps) Employees: several AIs Activities: the AIs divide up the tasks and move forward on their own Work is literally dying... I've summarized the full guide below, read it when you've got 5 min ⤵️ If you want the AI to work while you sleep → save this as a bookmark 🔖

Rahul @sairahul1

2 weeks ago

x.com/i/article/2059…

12 45 353 827K 1K

Almost everyone is building agent harness systems the wrong way. The default move: pick LangChain or LangGraph or the OpenAI Agents SDK, accept the loop, the tools, the memory, the orchestration, the policy engine, the credential store, the budget tracker, all of it, as one decision. Mike, wrote a long piece today on why this shape is wrong, and why every long-running agent team eventually ends up rewriting its harness from scratch. His argument: a harness isn't one thing. It's fifteen separate concerns bundled together because the surrounding ecosystem didn't give you a way to compose them. Turn state machines, provider routing, credential vaults, policy engines, approval gates, budget trackers, hook fanout, context compaction, session trees, OpenTelemetry tracing. Frameworks ship them as one block because that was the only shape available a year ago. It isn't anymore. When every layer is a worker on a shared bus with a typed function contract, "build your own harness" stops meaning "fork a framework." It means swap a worker. Don't like the model catalogue? Write one that hits a live API. Don't like file-backed credentials? Plug in your secrets manager. Want approvals routed through Slack instead of a console? Add a worker that calls approval::resolve. The rest of the stack does not change. The framework era picked a position for you and locked you in. The worker model leaves the choice in your hand. Worth reading in full.

@

57 years ago

0 0 0 0 0

Excited to dive into this - an open source agent designed with memory/continual learning in mind

Quarq @quarqlabs

2 weeks ago

x.com/i/article/2059…

17 35 243 142K 570

Generates 23 Mermaid diagram types in Claude Code github.com/WH-2099/mermai…

Life when nobody knows wtf you're up to

API Design Roadmap. Grab the API Design ebook: codewithdhanian.gumroad.com/l/nbfkk

Here's a version of this for Claude Code. Rebuilt the evidence stack for Claude's memory architecture: Session Memory, MEMORY. md, CLAUDE. md, git history, .claude/skills/, and hooks. Removed the Chronicle reference (no Claude Code equivalent yet). Full prompt in the image. Copy and run it at the start of any session.

Vaibhav (VB) Srivastav @reach_vb

2 weeks ago

UPDATE: Came up with an even better version of this prompt after the feedback Ask Codex to look across your sessions, Memories, and Chronicle, identify patterns, reuse what already exists, and only create the smallest useful skill, subagent, or automation. "Look back over my

97 368 4K 868K 8K

12 must-use features in Claude Code: