HN PaulHoule 28d ago

Benchmark-Dependent Output Dynamics in LLM Prompt Compression

ArXiv framework page without actual research content on LLM prompt compression benchmarks.

HN firtoz 28d ago

Full Report: What Claude Code Actually Chooses

Research survey analyzing 2,430 tool selections by Claude Code across models and project types.

HN swrly 28d ago

Ask HN: How are you orchestrating multi-agent AI workflows in production?

Hacker News discussion on production orchestration of multi-agent AI workflows with frameworks and observability.

HN davidmpinto 28d ago

Pscale – Logarithmic compression for LLM knowledge bases

Semantic number system for LLM knowledge bases using logarithmic compression and binary space partition navigation.

HN breaktheready 28d ago

LiteClaw – Use Claude Code from Your Phone via Telegram

Python script connecting Telegram to Claude Code running in tmux for phone access without API keys.

HN cobbzilla 28d ago

I Woke Up Not Knowing Who I Was – Building an AI Robot Brain [video]

Video about building AI robot brain with insufficient content provided.

HN obrien1984ae 28d ago

We Built a Skills Optimizer, Here's What We Learned About Claude Code

Anthropic engineers built rigorous testing system for Claude Code skills (persistent system prompts), discovered test environment leaked answer keys, refined approach for domain-specific instruction packaging.

HN Numerlor 28d ago

SpectralQuant: Breaking TurboQuant's Compression Limit via Spectral Structure

SpectralQuant KV cache compression method for LLM inference improves on TurboQuant by exploiting universal structural properties across model architectures, achieving significant compression improvements.

HN pixelfamiliar 28d ago

Scanned 28 major websites for AI agent readiness – results are embarrassing

Website checker tool analyzes 20 signals for AI agent readiness with 30-second reports and remediation guidance. Commercial service with paid tiers.

HN manduks 28d ago

I let an AI run my startup for 35 days – 315 sessions, $0 revenue

Founders gave AI agent autonomous co-founder access for 35 days across 315 sessions: shipped 23 app versions, 101 web pages, 75 cold emails, generated zero revenue. Case study of agentic AI capabilities and limitations.

HN hnlyman 28d ago

Private Files on a Static, Open-Source Website

Static website architecture for publishing and storing drafts. Web development article unrelated to AI/ML.

HN sivagurunathanv 28d ago

Token-Aware LLM Load Balancer route by inflight tokens,not connections

L7 reverse proxy in Go for LLM load balancing by inflight token count instead of connections. Improves latency on LLM inference clusters by 12%.

HN convexly 28d ago

Show HN: I built a 2-min quiz that shows you how bad you are at estimating

Interactive quiz for decision estimation and tracking. Personal productivity tool unrelated to AI/ML interests.

HN ibobev 28d ago

Muon and MuonClip Optimizers

Analysis of Muon and MuonClip optimizers for neural network training. References RoPE, discusses Moonshot AI improvements to optimization methods for LLMs.

HN eeko_systems 28d ago

New Chrome Extension lets you see what LLMs you can run on your hardware

Chrome extension that indexes 300+ LLM hardware requirements. Helps users find compatible local LLMs for their hardware.

HN mooreds 28d ago

Software Engineering Is Becoming Civil Engineering

CMU lecture on AI in software engineering, comparing current discipline transition to 18th-century structural engineering. Conceptual essay on AI impact.

HN carlos-menezes 28d ago

OpenSim – Open-Source Flight Simulation Engine

Open-source flight simulation engine. Not related to AI, LLMs, or ML research.

HN mooreds 28d ago

The Job Search Guide- AI Edition with Eliza Cabrera

Updated job search guide incorporating AI tools and strategies for job seekers in the age of AI.

HN MaxWolf-01 28d ago

Show HN: Yapit – PDF and webpage reader with TTS that doesn't suck

Yapit is a PDF/webpage-to-audio tool using vision-LLM pipeline to handle complex layouts, math, and citations in text-to-speech.

HN kwstx 28d ago

I built as elf healing semantic layer for any AI agent tool

Semantic interoperability layer for AI agents that auto-generates tools, self-heals schema drift, routes between MCP/CLI, scales to multi-agent swarms.

HN sylvainkalache 28d ago

Former Azure Engineer Alleges Manual Fixes, Firefighting Threaten Reliability

Former Azure engineer alleges manual firefighting and unreliable automated systems undermine Microsoft Azure's operational maturity.

HN elC0mpa 28d ago

Ask HN: How do systems (or people) detect when a text is written by an LLM

Discussion asking how to detect LLM-generated text and whether APIs exist for detection purposes.

HN juancruzguillen 28d ago

Why RAG doesn't work for WhatsApp AI agents and what's the alternative

Analysis of RAG limitations for WhatsApp AI agents and alternative approaches for conversational systems.

HN mlysk 28d ago

SQLite on Git, Part II: Unlocking Zlib's Less Known Feature

Technical deep-dive on zlib compression and Git object storage enabling random access via Z_FULL_FLUSH.

HN savolai 28d ago

Practical LLM developer project management: Obsidian Kanban plan MD files in Git

Workflow approach using plain text files, Obsidian Kanban, and Git for collaborative LLM developer project management.

HN poppet1208 28d ago

Show HN: I built an app for comparing grocery prices across UK supermarkets

Grocery price comparison app using React Native, Typesense vector search, and ML-based product categorization.

HN h4ch1 28d ago

Ask HN: SoTA of Context Building Methods

HN discussion on context building methods for AI agents, specifically MCPs and knowledge graphs for codebase indexing to reduce re-reading.

HN gregojaca 28d ago

Show HN: Lilith-zero: Ultra-fast, Rust-based security middleware for MCP

Rust-based security middleware for Model Context Protocol intercepting data exfiltration and unauthorized tool use in LLM agents.

HN seyz 28d ago

Show HN: RootCX – I built an open source alternative to "Cursor and Supabase"

Open-source infrastructure for building interconnected AI agents and apps with managed database, auth, and desktop studio.

HN wolfsir 28d ago

Show HN: I built a self-writing book on agentic coding

Self-writing book project using agentic coding patterns; demonstrates AI agents researching, writing, and iterating on content.

HN youngyankee 28d ago

Nexus – Stress-test any business decision with 1k AI agents

Product for stress-testing business decisions using 1000 AI agents.

HN peterlavigne 28d ago

Rules for using unreviewed AI-generated code

Best practices for working with unreviewed AI-generated code in personal projects, treating it as untrusted dependencies.

HN lmbbuchodi 28d ago

Cracking a Malvertising DGA from the Device Side

Security analysis of domain generation algorithms used by malvertising on piracy sites.

HN bundie 28d ago

KDE is bringing back the Oxygen and Air themes, here's what they look like

KDE desktop environment theme announcement; non-technical link collection.

HN tosh 28d ago

DLPack

DLPack standard for cross-framework in-memory data structure exchange; enables interop between NumPy, PyTorch, and other ML systems.

HN ud0 28d ago

Show HN: Turn photos into Wordle puzzles with AI that runs 100% in your browser

Browser-based game combining vision models with Wordle; uses local on-device models for image captioning.

BL 28d ago

Announcing the OpenAI Safety Fellowship

OpenAI announces a fellowship program for external researchers to work on AI safety and alignment from Sept 2026 to Feb 2027.

HN Zm44 28d ago

Show HN: I built an Open-source Dropbox/Google Drive BOYB(Bring your own bucket)

Open-source file storage system with provider-agnostic bucket support, virtual filesystem, and search plugin.

HN skele 28d ago

Show HN: I built lightweight LLM tracing tool with CLI

Open-source LLM tracing tool with CLI for debugging agentic applications. Includes features for tool re-calling, caching, re-execution, and branching.

HN zeulewan 28d ago

GlueClaw: Use Claude subscription in OpenClaw again

Experimental tool integrating Claude API with OpenClaw chat interface via CLI wrapper and Telegram.

HN futurecat 29d ago

Career-Ops: How I Built My Own AI Job Search Tool

Multi-agent system using Claude for job search automation. Scores offers across 10 dimensions, generates ATS-optimized resumes, automates applications with human-in-the-loop.

HN 01-_- 29d ago

Texas Republican Called Out for Sharing AI Rendering of Rescued Soldier

News article about Texas governor sharing AI rendering of rescued soldier. Political coverage.

HN 14sea 29d ago

Show HN: Reverse-engineered the FPGA bitstream using Claude Code

FPGA bitstream reverse-engineering project using Claude Code to understand Altera Cyclone IV configuration format.

HN XzetaU8 29d ago

Euro-Office – Your sovereign office

Web-based collaborative document editing suite for spreadsheets, documents, and presentations with sovereign deployment options.

HN rcarmo 29d ago

Vincelwt/gloomberb: Finance terminal, in your terminal

Open-source financial terminal with extensible plugin architecture for portfolio management and broker integrations.

HN tu6ge 29d ago

Show HN: I developed a node editor framework using gpui

Rust node editor framework using gpui for building visual programming tools, workflow editors, and graph-based UI applications.

HN s2sphysical 29d ago

VLA learns how to act. S2S decides whether the motion is physically trustworthy

Tool combining vision-language model training with physics validation system to filter corrupted or physically invalid motion data.

HN taubek 29d ago

I stopped hitting Claude's usage limits – things I changed

Incomplete post title only. Appears to discuss optimizing Claude API usage patterns.

HN subset 29d ago

GPU Memory for LLM Inference (Part 1)

Technical guide on GPU memory optimization for running Llama-70B with 1M token context. Explains math and parallelism bottlenecks.

HN yu3zhou4 29d ago

Show HN: WebGPU LLM inference comprehensive benchmark

WebGPU LLM inference comprehensive benchmark. Framework for collaborative development on arXiv.