Dev Systems - Building Systems That Scale

Building Prometheus: How Backend Aggregation Enables Gigawatt-Scale AI Clusters

We’re sharing details of the role backend aggregation (BAG) plays in building Meta’s gigawatt-scale AI clusters like Prometheus.BAG allows us to seamlessly connect thousands of GPUs across multiple data centers and regions.Our BAG implementation is connecting two different network fabrics – Disaggregated Schedule Fabric (DSF) andNon-Scheduled Fabric (NSF).Once it’s complete our AI cluster,Prometheus, will deliver 1-gigawatt of capacity to enhance and enable new and existing AI experiences across

Show HN: Standardized robot brain with hardware safety – 10 patents in 4 days

I'm a solo inventor in rural Pennsylvania. Over four days last week, I filed 10 provisional patents (70 claims, $650 total) describing a complete hardware specification for a standardized AI compute module for robots -- what I've been calling a "robot brain."The spec defines three standardized sizes (drone to surgical robot), a universal connector system (the Manufacturer Interface Module, or MIM) so any brain works with any robot body, and a hardware safety architecture wher

Show HN: Simulation Studio Inside ArchtSoft

We just shipped Simulation Studio in ArchtSoft.It lets you simulate and compare software architectures before writing code — focusing on performance, scalability, resilience, and cloud cost trade-offs.What it does: Simulates architecture behavior under different load scenarios Compares alternative designs side-by-side Evaluates AWS / Azure / GCP implications early Surfaces bottlenecks and failure risks before production Helps reduce redesign cycles and cloud overspendThe goal is simple

The Null Pointer Crisis: Running God-Mode Software on Legacy Hardware

We spend a lot of time discussing AI Alignment—how to ensure the machine’s values match ours. But I’d like to open a thread on a parallel issue that seems to be causing a "Kernel Panic" in the current generation (specifically Gen Z and Alpha): The Human Alignment Problem.I approach this not as a psychologist, but from a systems engineering perspective. If we look at the rising rates of anxiety, burnout, and suicide not as "illnesses" but as System Errors, a terrifying archite

Show HN: I spent 2 years building a feedback tool with only the features I need

Hey everyoneI’m a solo founder and developer, and I’ve been building software since 2011. Over the years, I watched customer feedback tools go from simple and focused to bloated, enterprise-heavy platforms. Tons of features, endless configuration… and yet I kept using maybe 10% of what was there. The rest mostly got in the way.So I decided to build the feedback tool I actually wanted to use.That became Zigpoll — a fast, focused survey and feedback platform designed for real-world businesses. It

Show HN: Artifact Keeper – Open-Source Artifactory/Nexus Alternative in Rust

I'm a software engineer who keeps getting pulled into DevOps no matter how hard I try to escape it. I recently moved into a Lead DevOps Engineer role writing tooling to automate a lot of the pain away. On my own time outside of work, I built Artifact Keeper — a self-hosted artifact registry that supports 45+ package formats. Security scanning, SSO, replication, WASM plugins — it's all in the MIT-licensed release. No enterprise tier. No feature gates. No surprise invoices.Your package m

Show HN: I'm 75, building an OSS Virtual Protest Protocol for digital activism

Hi HN,I’m a 75-year-old former fishmonger from Japan, currently working on compensation claims for victims of the Fukushima nuclear disaster. Witnessing social divisions and bureaucratic limitations firsthand, I realized we need a new way for people to express their will without being “disposable.”To address this, I designed the Virtual Protest Protocol (VPP) ? an open-source framework for large-scale, 2D avatar-based digital demonstrations.Key Features:Beyond Yes/No: Adds an "Observe&

Why a solid app failed without marketing focus

I’ve built several apps over the years that were technically solid.They worked. They were stable. They solved real problems. Some were even used by paying customers.And yet — they went nowhere.For a long time, I framed this as a distribution problem, or a timing problem, or “I just didn’t push it hard enough.” What I eventually realized is simpler, and harder to accept:Building software and growing a product are fundamentally different disciplines.I enjoy building systems. I enjoy architecture,

GenAI-based software delivery needs a fast flow architecture

GenAI doesn’t change the fundamentals of software architecture.In fact, GenAI-based software development is an amplifier of organizational capabilities, not a replacement for sound architectural principles.It can make high-performing organizations even better and make low-performing organizations worse.For organizations adopting GenAI-based software delivery, fast flow is therefore not optional, but essential.This article explores what GenAI-based software delivery means for architecture and org

Show HN: Chitram – Open-source image hosting with automatic AI tagging

I wanted to learn distributed systems by building something real instead of just reading about them. So I built an image hosting service that automatically tags uploads using GPT-4o-mini vision.What it does: Upload an image → Celery worker picks it up → sends to OpenAI Vision API → tags get saved to PostgreSQL. Takes about 10 seconds, costs ~$0.004/image.Why I built it this way:I kept running into the same architecture decisions at work (swapping providers, handling async tasks, testing bac

Show HN: Calfkit – an SDK to build distributed, event-driven AI agents on Kafka

I think agents should work like real teams, with independent, distinct roles, async communication, and the ability to onboard new teammates or tools without restructuring the whole org. I built backend systems at Yahoo and TikTok so event-driven agents felt obvious. But no agent SDKs were using this pattern, so I made Calfkit.Calfkit breaks down agents into independent services (LLM inference, tools, and routing) that communicate asynchronously through Kafka. Agents, tool services, and downstrea

Show HN: 33rpm – A vinyl screensaver for macOS that syncs to your music

I built a macOS screensaver that displays a spinning vinyl record synced to whatever you're listening to. The record spins while music plays and stops when you pause.Works with Spotify, Apple Music, or any app that reports to Now Playing (podcasts, YouTube, etc.).A few technical details HN might find interesting:- Zero permissions required. I use distributed notifications and the MediaRemote framework to read now-playing state without requesting any system access.- Album art comes from Spot

Show HN: LayerClaw – Observability tool for PyTorch training

Hi HN! I built LayerClaw (https://github.com/layerclaw/layerclaw), a local-first observability tool for PyTorch training. The problem: When training neural networks, things go wrong silently. Your loss explodes at step 47,392. Your gradients vanish in layer 12. Your GPU memory spikes randomly. By the time you notice, you've wasted hours or days of compute.I got tired of adding print statements, manually checking TensorBoard files, and tracking down training issues after

Show HN: Hibana – choreography-first protocol safety for Rust

Hi HN — I just open-sourced Hibana and hibana-agent.Hibana is an Affine MultiParty Session Types (MPST) runtime for Rust.It targets protocol drift in distributed systems. Instead of maintaining separate hand-written state machines in each component, you define interaction once as a global choreography and project role-local behavior at compile time. At runtime, only valid protocol transitions are executable, so invalid moves such as skipping, reusing, or taking the wrong branch are rejected by t

Show HN: Vibefs – A file preview server for remote vibe coding

I vibe-coded this tool entirely from my phone with OpenClaw, to solve a problem I encountered in this workflow. It's called vibefs.The use case: when you're remotely controlling an AI agent and want to preview a file it generated — a development plan, a piece of code, or some log it found — you can have the agent run vibefs allow /path/to/file and send you back an accessible URL. Open it in your browser and you're looking at the file with syntax highlighting.For lar

Show HN: Astrolabe – Navigate Your Data Universe in Nextcloud

Astrolabe is a new Nextcloud app that I've developed that enables semantic search across your documents, notes, calendar, recipes, etc.It leverages the nextcloud-mcp-server (same author) as the backend to enable AI workflows via MCP that you can use from any compatible MCP client such as Claude Code, Gemini CLI, Opencode, etc.This is my first foray into NC app development, and I'd really appreciate any productive criticism. I learned a lot about php, authentication within Nextcloud, as

Show HN: ARM64 Android Dev Kit

GUI-first, multi-service gRPC scaffold for an Android Development Kit style workflow on an AArch64 system.

Show HN: Seele AI – Generate playable games from text prompts using MLLM

I'm a co-founder of Seele AI. We are building a generative engine designed specifically for game development.Unlike typical code-generation tools that rely on general-purpose LLMs, we have developed our own Multimodal LLM (MLLM) optimized for gaming. It doesn't just output text/code; it natively handles 3D models, spatial coordinates, and visual information to construct the game world.Key Features:Engine Support: We currently support real-time generation for Three.js and Unity Web

Show HN: I built Clash to avoid conflicts when running AI agents in parallel

Hey HN,I made Clash, an open-source cli tool that helps you/your coding agent manage merge conflicts that occur during parallel development on the same repository.Example: Two agents working on different features both modify src/main.rs. Clash shows you this conflict immediately and not after lots of time and tokens spent during the merge time.It is a simple tool that your agent can use as part of the development process to check if the commits agents make are conflicting with any othe

Show HN: aTerm – a terminal workspace built for AI coding workflows

I’ve been doing more AI-assisted development lately using tools like Claude Code and Aider, and I kept running into the same friction: too many terminal tabs and split windows to manage at once — agent, shell, dev server, tests, git, etc.So I built a small macOS app called aTerm — a terminal workspace designed around multi-pane, agent-driven coding workflows rather than single shells.Core ideas:- project-based terminal workspaces- predefined layouts for common AI coding setups (AI + Shell, AI +