frontpage.
newsnewestaskshowjobs

Made with ♥ by @iamnishanth

Open Source @Github

fp.

Open in hackernews

I built an AI agent in Zig that runs on Windows XP with 64 MB RAM

2•bcorp•1h ago

Comments

bcorp•1h ago
I wanted to see how far back I could push an AI agent — not the LLM itself, but the client that talks to it, parses tool calls, and acts on the results.

retro-agent is a terminal-based AI agent written in Zig 0.15. It connects to Ollama (or any OpenAI-compatible API) over local HTTP, supports function calling, and provides built-in tools for system diagnostics: processes, network, disk, services, memory, and arbitrary command execution.

The target is Windows XP SP3 x86 on hardware as old as a Pentium III with 64 MB RAM. The binary is ~750 KB, single-threaded, no dependencies, no UCRT/MSVC runtime. It also cross-compiles to Linux x86/x64/ARM.

Some things I had to deal with:

- Win32 Console API for the TUI (CP437 box-drawing, colored output) - Automatic CP850 → UTF-8 conversion for localized Windows command output - UTF-8 → ASCII sanitization for console display - A compatibility shim for RtlGetSystemTimePrecise (doesn't exist on XP) - Conversation history with a sliding window to stay within memory limits - Command whitelist and approval mode for security

The LLM runs on a separate machine on the network — Ollama can't run on XP. The agent is just the thin client: parse, call tools, feed results back, loop until you get a text response (max 10 iterations).

Tested with llama3, qwen2, mistral, and command-r. Any model with function calling support works.

MIT licensed. Would love feedback, especially from anyone still managing legacy Windows systems or working with Zig's cross-compilation story.

https://github.com/benmaster82/retro-agent

Why Do They Want to Get Rid of Software Engineers?

https://jrswab.com/blog/why-do-they-want-to-get-rid-of-software-engineers/
1•abnercoimbre•59s ago•0 comments

What Entertainment Might Look Like in 20 Years

https://www.wsj.com/arts-culture/entertainment-in-20-years-d6d5ac9b
1•fortran77•1m ago•1 comments

BookGraph: Moving beyond naive RAG with graph-native AI reasoning

https://github.com/sumant1122/bookgraph
1•paperplaneflyr•3m ago•1 comments

The Obvious Is Taking Its Revenge on Trump

https://www.theatlantic.com/international/2026/03/iran-war-trump/686314/
1•Jtsummers•4m ago•1 comments

Tenstorrent TT-QuietBox 2

https://tenstorrent.com/waitlist/tt-quietbox
1•LorenDB•4m ago•0 comments

Show HN: Launching Our XR-Native Operating System

https://explore.getxeneva.com/
1•ayush_xeneva•5m ago•0 comments

Nvidia's Huang argues AI creates jobs, not destroys them

https://www.coindesk.com/tech/2026/03/10/nvidia-s-huang-argues-ai-creates-jobs-not-destroys-them-...
2•Vaslo•5m ago•0 comments

Ask HN: What are your favorite alternate history scenarios?

1•StartupsWala•7m ago•0 comments

Simple, fast and open-source PHP framework frozen in C-extension

https://www.iceframework.org/
1•mooreds•9m ago•0 comments

Show HN: Fast-Axolotl – Rust extensions that make Axolotl fine-tuning 77x faster

https://github.com/neul-labs/fast-axolotl
1•ticktockten•10m ago•0 comments

How to Design Antibodies

https://press.asimov.com/articles/antibody-design
1•mailyk•11m ago•0 comments

Extend Protobuf/FlatBuffers Schema IDL with Shared/Circular Reference Support

https://fory.apache.org/blog/extend_protobuf_flatbuffers_with_shared_circular_refs/
1•chaokunyang•12m ago•3 comments

Code Canary: Realtime Reporting of Coding Agent Performance

https://fredbenenson.com/blog/2026/03/11/introducing-code-canary/
1•mecredis•12m ago•0 comments

Designing AI Chip Hardware and Software

https://docs.google.com/document/d/1dZ3vF8GE8_gx6tl52sOaUVEPq0ybmai1xvu3uk89_is/view
1•matt_d•14m ago•0 comments

Pro-Worker AI

https://www.brookings.edu/articles/building-pro-worker-ai/
1•Cerchie•14m ago•0 comments

Client-Side AI React Hooks Powered by Transformers.js and Web Workers

https://github.com/baskvava/react-zero-ai
1•baskvava•15m ago•1 comments

Brazilian Age-Verification Law: I Posit It Does Not Apply to Open-Source OSes

https://www.planalto.gov.br/ccivil_03/_ato2023-2026/2025/Lei/L15211.htm
1•replooda•16m ago•1 comments

Programmable Property-Based Testing

https://arxiv.org/abs/2602.18545
1•PaulHoule•17m ago•0 comments

Yahoo Introduces MyScout, the First Personalized Homepage for AI Answers

https://www.yahooinc.com/press/yahoo-introduces-myscout-the-first-personalized-homepage-for-ai-an...
1•drtz•18m ago•0 comments

I paired NotebookLM with Claude Code, and it feels like a dream team

https://www.xda-developers.com/paired-notebooklm-with-claude-code/
1•speckx•18m ago•0 comments

Replit raises $400M at $9B valuation

https://techfundingnews.com/replit-raises-400m-9b-valuation-ai-app-building/
3•exizt88•18m ago•0 comments

Tcl's Nxtpaper 4.0 screen: A review

https://manualdousuario.net/en/tcl-nxtpaper-4/
1•rpgbr•19m ago•0 comments

Sam Altman says OpenAI will tweak its Pentagon deal after surveillance backlash

https://www.businessinsider.com/openai-amending-contract-with-pentagon-amid-backlash-mass-surveil...
1•doener•21m ago•3 comments

YouTube just approved 30-second unskippable ads for TV

https://www.androidcentral.com/apps-software/youtube/youtube-on-tv-30-seconds-unskippable-ads
1•LorenDB•21m ago•0 comments

Goldman executive says private markets clients glad about Iran war 'distraction'

https://www.ft.com/content/9232dbce-0255-4949-8c4c-ea58d86a4166
2•alephnerd•21m ago•0 comments

Most AI chatbots will help users plan violent attacks, study finds

https://www.engadget.com/ai/most-ai-chatbots-will-help-users-plan-violent-attacks-study-finds-163...
1•mikece•22m ago•0 comments

ChatGPT Took The Pentagon's Killer Robot Deal: Boycott Now

https://quitgpt.org/pentagon?link_id=2&can_id=3b2cebf422aaa35898d6d8ce17355809&source=email-week-...
3•doener•22m ago•0 comments

The Web Is a Guitar Amp Now (Literally)

https://www.silverorange.com/blog/the-web-is-guitar-amp
2•speckx•22m ago•0 comments

The Bay Area Considers the Unthinkable: Life Without BART

https://www.nytimes.com/2026/03/10/us/bart-bay-area-san-francisco-transit.html
1•radley•23m ago•0 comments

ChatGPT Uninstalls Skyrocket

https://twitter.com/SensorTower/status/2029250034772963513
1•doener•23m ago•0 comments