Dev Systems - Building Systems That Scale

Ranking Engineer Agent (REA): The Autonomous AI Agent Accelerating Meta’s Ads Ranking Innovation

Meta’s Ranking Engineer Agent (REA) autonomously executes key steps across the end-to-end machine learning (ML) lifecycle for ads ranking models.This post covers REA’s ML experimentation capabilities: autonomously generating hypotheses, launching training jobs, debugging failures, and iterating on results. Future posts will cover additional REA capabilities.REA reduces the need for manual intervention. It manages asynchronous workflows spanning days to weeks through a hibernate-and-wake mechanis

Resiliency Patterns for Azure Front Door: Field Lessons

AbstractAzure Front Door (AFD) sits at the edge of Microsoft’s global cloud, delivering secure, performant, and highly available applications to users worldwide. As adoption has grown—especially for mission‑critical workloads—the need for resilient application architectures that can tolerate rare but impactful platform incidents has become essential.This article summarizes key lessons from Azure Front Door incidents in October 2025, outlines how Microsoft is hardening the platform, and—most impo

GenAI-based development platform - part 2: How Idea to Code turns an idea into working, tested software

This article is the third in a series about the GenAI-based development platform (aka. harness) that I’ve been developing to make GenAI-based coding agents like Claude Code more productive and less frustrating.The platform is a set of tools, constraints and feedback loops that guide their behavior, catch mistakes and prevent them from generating large amounts of poor-quality code that is often incomplete and untested.The complete list of articles in the series is as follows: Part 0 - My GenAI d

Ask HN: Critique the published validation work for my blackjack simulator

I’m building a blackjack simulator and have published the architecture, methodology, validation reports, references, and known discrepancies here:https://ether-ore.github.io/BJW/The code is *not* open source, so I’m not asking anyone to audit the implementation itself. What I am asking is whether the *public validation work* looks serious and where its blind spots may be.In particular:1. If you were trying to explain the remaining differences with industry-standard tools like

Show HN: Agents shouldn't operate software–they should coordinate commitments

I've been working on Covenant Layer — an open protocol and framework for shifting AI agent systems from tool orchestration to outcome coordination.The core idea: agents shouldn't operate software step by step. They should publish objectives, compare competing provider offers, accept the best one under policy, and let providers fulfill outcomes with evidence and settlement.Why this matters: we're still building agents as "software operators" — better interns that click th

Future After the AI Revolution

Current AI revolution is building larger models, using feedback to fine-tune, building agents around them and such.I was thinking what will be the next revolution. We will have a true leap forward.We will have self-aware beings among us. John von Neumann architecture will be done for good. There will be zero software as a consequence. ( All in learning models ). Even biology is not there actually (DNA is a lot like John von Neumann than we would think), so this is a very tall claim.We may potent

Ask HN: Are there any CS niches safe from AI?

I'll be graduating with a SWE degree soon and the thought of spending my career reviewing AI code just seems both awful and unsustainable.Programming has been fun because it is difficult. It required skill which is continously grown by the act of writing more code.Using code gen is boring, doesn't require much skill, and tends to atrophy your understanding of the subject.Does anyone have any experience with nearby niches to software engineering which might have some inherent property t

Show HN: AntroCode-A zero-dependency,single-file local AI client,159 clone in 4D

Hi HN,I'm the developer of AntroMind. Today I want to share AntroCode with you, an open-source project I've been refining for the past few months.Project IntroductionAntroCode is a zero-dependency, single-file local AI client designed specifically for models like DeepSeek. Its core design philosophy is simplicity—you only need to download an Python file, open it, and use it immediately, without installing any dependencies or configuring any environment.GitHub Address: https:/&#x2F

Ask HN: If everyone is selling, then who is buying?

The cost of building software products is collapsing. A solo dev with LLMs can ship in a weekend what took a team months (Seriously some devs can do it). Has demand scaled with supply?Every other industry that went through this (fashion, food, furniture) followed the same arc which started with craftsmanship → mechanization → overproduction → commoditization → value migrates elsewhere. All of these industry feels and tastes the same. Consider food, everyone wants to open the next luxurious dinin

Show HN: Got tired of AI copilots just autocompleting, and built Glass Arc

Hey HN,Over the last few months, I realized I was paying $20/month for an AI that essentially just acts as a really good autocomplete. It waits for me to type, guesses the next block, and stops. But software engineering isn't just writing syntax, it's managing the file system, running terminal commands, and debugging stack traces.So I pivoted my project and built Glass Arc. It’s an agentic workspace that lives directly inside VS Code.Instead of just generating text, I gave it actu

Show HN: Scryer – Visual architecture modeling for AI agents

I've been working on this desktop tool (FSL license, free for commercial use) for the past month because I now spend more time in a terminal prompting Claude Code instead of using a code editor. It generally works quite well if I ask the right questions, but I still often find a lot of dead code, stubs, or poor architectural choices when I finish a session, and understanding the codebase itself can be jarring after making major changes through vibecoding.The idea for Scryer is to provide a

Show HN: Open-Source Workflow Builder SDK

Hi HN,I'm Maciej, founder of Workflow Builder. Over the last few years our team has been building diagramming and workflow tools for complex systems (industrial automation, AV system design, financial workflows, etc.).One thing we repeatedly noticed while working with clients is that many companies did not want to adopt full workflow automation platforms. Tools like Zapier, n8n or Camunda are great when you want an entire automation platform. But many teams we worked with wanted something d

Zirco.ai – AI employee for dental front desk operations

Hi HN,I'm building Zirco.ai — an AI administrative employee for dental practices.The problem: dental front desks spend 2–3 hours every day manually verifying insurance benefits through carrier portals. On top of that, they're handling inbound scheduling calls, sending reminders, coordinating referrals, and managing new patient intake — all manually, all repetitive, all expensive. A single front desk employee costs $40–50K/year and turns over at 40% annually.What I built: an AI tha

Show HN: SiMM – Distributed KV Cache for the Long-Context and Agent Era

We built SiMM because LLM context lengths are growing much faster than GPU memory.With long Chain-of-Thought reasoning and multi-turn agents, prompts are getting much longer. According to OpenRouter’s State of AI 2025, average context length has grown about 4× in the past year.This creates two problems in inference systems:• Slow TTFT — long contexts make prefill expensive • High GPU memory cost — KV cache quickly exhausts HBMInstead of recomputing long prompts or keeping all KV cache in GPU mem

Tell HN: AI tools are making me lose interest in CS fundamentals

With powerful AI coding assistants, I sometimes feel less motivated to study deep computer science topics like distributed systems and algorithms. AI can generate solutions quickly, which makes the effort of learning the fundamentals feel less urgent.<p>For those who have been in the industry longer, why do you think it’s still important to stay strong in CS fundamentals?

Show HN: AgentLog – a lightweight event bus for AI agents using JSONL logs

I’ve been experimenting with infrastructure for multi-agent systems.I built a small project called AgentLog.The core idea is very simple, topics are just append-only JSONL files.Agents publish events over HTTP and subscribe to streams using SSE.The system is intentionally single-node and minimal for now.Future ideas I’m exploring: - replayable agent workflows - tracing reasoning across agents - visualizing event timelines - distributed/federated agent logsCurious if others building agent sy

A lock-free state management engine for DualHeart Financial Association

We just finished an internal overhaul of the data synchronization layer at DualHeart Financial Association. The goal was to replace our traditional mutex-heavy architecture with a lock-free, event-driven state machine.Key technical wins for DualHeart Financial Association:Throughput: Achieved a 3x increase in message processing by eliminating lock contention.Architecture: Implemented a minimalist, high-end schema that maps directly to memory-mapped files for near-instant persistence.Low Latency:

Ask HN: Why do the majority of vibecoded projects fail?

This was recently asked on reddit in the context of a hypothetical "slack killer" Ws chat app that is proofed in 20 minutes and works on localhost, with the "failure" being because AI can't build scalable distributed systems, but discussion not so deep, a lotta jokes.Here's what I think:I think AI can build scalable, distributed systems, too, you just have to know what you're doing.to be slightly more substantial: there's a lot of "implicit" deci

Becoming a Forest Civilisation

A forest is not one tree. It is many different things growing in the same place, competing, cooperating, dying, regrowing. No single species runs it. It holds together because the diversity itself is the structure.Human civilisations tend toward monocultures. One way of thinking crowds out the others - not because anyone chooses it, but because successful ideas spread. This has always happened. It is happening faster now.A forest civilisation resists this deliberately. Not by fighting successful

Show HN: Replacing $50k manual forensic audits with a deterministic .py engine

I’m a software architect, and I recently built Exit Protocol (https://exitprotocols.com), an automated forensic accounting engine for high-conflict litigation.Problem: If you get divorced and need to prove that a specific $250k in a heavily commingled joint bank account is your "separate property" (e.g., from a pre-marital startup exit), the burden of proof is strictly mathematical. Historically, this meant paying a forensic CPA $500/hour to dump years of blurry bank PDF