frontpage.
newsnewestaskshowjobs

Made with ♥ by @iamnishanth

Open Source @Github

fp.

Mac mini will be produced in the US for the first time later this year

https://twitter.com/tim_cook/status/2026351829928624257
1•tosh•1m ago•0 comments

Off Grid: On-device AI-web browsing, tools, vision, image gen, voice – 3x faster

1•ali_chherawalla•1m ago•0 comments

Hacking an old Kindle to display bus arrival times

https://www.mariannefeng.com/portfolio/kindle/
1•mengchengfeng•1m ago•0 comments

Show HN: Interactive 3D Moon with real NASA data and WebGPU

https://moon.oddurs.com
1•oddurs•2m ago•0 comments

V1.3.0 Spring CRUD Generator- MariaDB Support + Null Exclusion in REST Responses

https://github.com/mzivkovicdev/spring-crud-generator/releases/tag/v1.3.0
1•mzivkovicdev•3m ago•0 comments

Earliest known writing dates back over 40k years

https://www.popsci.com/science/earliest-human-writing-germany/
1•_vaporwave_•4m ago•0 comments

Show HN: I applied Markowitz port. theory to agent teams / proved it in a zkVM

https://www.mnemom.ai/showcase
1•alexgarden•4m ago•0 comments

Hegseth warns Anthropic to let the military use company's AI tech as it sees fit

https://apnews.com/article/anthropic-hegseth-ai-pentagon-military-3d86c9296fe953ec0591fcde6a613aba
1•zzzeek•4m ago•0 comments

When newspapers cut book coverage, communities lose more than reviews

https://www.poynter.org/business-work/2026/why-dont-newspapers-review-books-anymore/
1•speckx•4m ago•0 comments

Meta's Internal Research about harms of social media

https://metasinternalresearch.org/
1•shrubby•5m ago•1 comments

Reddit fined more than £14M over age verification checks

https://news.sky.com/story/reddit-fined-14m-by-information-commisioners-office-over-age-verificat...
2•ivewonyoung•6m ago•1 comments

GPT-5.3-Codex is now available for all developers

https://twitter.com/OpenAIDevs/status/2026379092661289260
1•tosh•7m ago•0 comments

Anthropic's Existential Negotiations with The Pentagon

https://www.theverge.com/ai-artificial-intelligence/883456/anthropic-pentagon-department-of-defen...
1•czottmann•7m ago•1 comments

Show HN: Agently an AI Work OS that turns docs, chats, and tasks into execution

https://www.agently.dev/
1•Bondig•9m ago•1 comments

The sitting president is selling watches

https://gettrumpwatches.com/
1•vjvjvjvjghv•10m ago•0 comments

Writing about Agentic Engineering Patterns

https://simonwillison.net/2026/Feb/23/agentic-engineering-patterns/
1•Anon84•10m ago•0 comments

Show HN: Building to Remember. Using AI to Wrangle My Daily Mess

https://chrisberry-tech.pages.dev/
1•WarcrimeActual•10m ago•0 comments

Show HN: Srclight – Deep code indexing MCP server (FTS5 and Tree-sitter)

1•srclight•11m ago•0 comments

Show HN: Prompt → Schema → CRUD API and Admin UI (New Codehooks Template)

1•knutmartin•12m ago•0 comments

I built a tool that scores how replaceable you are in the AI economy

https://candidate.perfectly.so/roast
2•luogary•13m ago•1 comments

In a replay of 2019, Apple says a single Mac will be manufactured in the US

https://arstechnica.com/apple/2026/02/in-a-replay-of-2019-apple-says-a-single-desktop-mac-will-be...
4•rbanffy•13m ago•0 comments

OpenMedicare – 10 years of Medicare physician data analyzed for fraud patterns

https://www.openmedicare.us
1•kianoconnor•13m ago•0 comments

Minimalism

https://en.wikipedia.org/wiki/Minimalism_(computing)
1•tosh•14m ago•0 comments

GridCalc: An RPN Spreadsheet for iOS

https://tailrecursion.com/~alan/GridCalc.html
1•wooby•14m ago•0 comments

DJI Romo robovac had security so poor, man remotely accessed them

https://www.theverge.com/tech/879088/dji-romo-hack-vulnerability-remote-control-camera-access-mqtt
1•cglong•14m ago•0 comments

Laser irradiation method for additive manufacturing of WC–Co cemented carbide

https://www.sciencedirect.com/science/article/pii/S0263436825005906
2•PaulHoule•16m ago•0 comments

words – my own personal dictionary

https://words.clarkdinnison.com/a
2•kaniksu•16m ago•2 comments

Show HN: Writher – offline voice assistant for Windows (Whisper and Ollama)

https://github.com/benmaster82/writher
1•bcorp•17m ago•1 comments

Show HN: Selfie bodyfat % scan (offline, no server upload)

https://stateofutopia.com/experiments/bodyfat/
2•logicallee•17m ago•0 comments

Can Elon Musk run AI in space?

https://www.economist.com/insider/inside-tech/can-elon-musk-really-run-ai-in-space
1•andsoitis•21m ago•1 comments
Open in hackernews

I built a governance layer for multi-agent AI coding – lessons after 6 months

2•vincentvandeth•2h ago
Six months ago I started coordinating multiple AI coding agents (Claude Code, Codex CLI, Gemini CLI) across parallel terminals for a production project. The agents were productive, but I had no idea what they were actually deciding or why.

The problem wasn't capability — it was accountability. An agent would make a choice buried in a 50-file commit, and I'd only find out weeks later when something broke. No trace of which agent did what, when, or based on what context.

So I built a governance layer on top. The core idea: every agent decision gets recorded in an append-only receipt ledger (NDJSON). Each receipt links a specific agent action to a git commit, a dispatch ID, and a quality verdict. The orchestrator (T0) reviews receipts and decides what happens next — approve, hold, or redispatch.

Some things I learned: 1. Sub-agents are a black box. I never use them. When a bug surfaces, you can't trace which agent's context was polluted. Instead, I run independent agents in separate terminals with their own context windows, reporting back to T0. 2. Quality gates need to be deterministic, not LLM-based. An automated advisory checks every completion against pre-registered rules (file size limits, test coverage, open blockers). The LLM proposes, the gate validates. No vibes. 3. Context rotation is unsolved by the ecosystem. When an agent fills its context window mid-task, most workflows just fail. I built an automated rotation pipeline using Claude Code hooks — detects context usage, writes a structured handover, clears the window, and resumes. Zero human intervention. 4. The receipt ledger is the most valuable artifact. After 1100+ entries, patterns emerge: which types of tasks fail, which agents struggle with what, where context pollution happens. That data feeds back into dispatch planning. 5. Terminal locking prevents chaos. Each terminal can only work on one dispatch at a time. Sounds obvious, but without it you get overlapping work, merge conflicts, and agents overwriting each other's changes.

The system runs across 4 tmux panes (T0 orchestrator + 3 worker tracks), supports multiple AI providers, and everything is filesystem-based — no database, no cloud dependency. Open-sourced it recently.

Happy to answer questions about the architecture or specific failure modes.