2/24/2026
Show HN: AI-Nexus – Unified Rule Manager for Claude Code, Cursor, and Codex
AI-Nexus: unified rule manager for Claude Code, Cursor, and Codex. Limited technical details provided.
2/24/2026
Conduit AI – AI voice agent that answers missed calls for service businesses
Conduit AI: voice agent product for capturing missed calls and lead qualification. Commercial tool, minimal technical depth.
2/24/2026
China Bet Billions on Agentic AI as Commerce Becomes the New Battleground
Analysis of Chinese hyperscalers investing in agentic AI for commerce and enterprise automation as competitive battleground.
2/24/2026
Show HN: Jsonchunk – Parse incomplete JSON from streaming LLM responses
npm package parsing incomplete JSON from streaming LLM responses; handles token-by-token generation for structured output.
2/24/2026
Show HN: Localvoxtral – Local real-time dictation on macOS with streaming STT
Native macOS app for real-time local speech-to-text using Mistral's Voxtral streaming model.
2/24/2026
Show HN: TypeGraph – Type-safe graphs on Postgres/SQLite (no graph DB required)
TypeGraph: type-safe knowledge graph library running on Postgres/SQLite without dedicated graph database.
2/24/2026
Show HN: Bruce – AI signal radar for Reddit/HN that learns what matters to you
Bruce: AI signal filtering tool for Reddit/HN using product context and competitor data. Original developer tool with ML learning component.
2/24/2026
LFM2-24B-A2B: Scaling Up the LFM2 Architecture
LFM2-24B-A2B: Scaling study of LFM2 language model architecture. Machine learning research. Brief headline only.
2/24/2026
IBM stock tumbles 10% after Anthropic launches COBOL AI tool
News article about IBM stock decline following Anthropic's COBOL-writing AI tool release. Market impact.
2/24/2026
Towards a Science of AI Agent Reliability
Research on science and frameworks for AI agent reliability. AI agents focus. Brief headline only.
2/24/2026
How we made Docker builds 193x faster across AI agent sessions
Optimization achieving 193x faster Docker builds across AI agent sessions. Performance improvement for agent systems.
2/24/2026
Mercury 2: Diffusion Reasoning Model
Mercury 2: Diffusion-based reasoning model for AI. New machine learning research/model. Brief headline only.
2/24/2026
Decision Theory Cuts Your AI Agent's API Bill in Half
Technical analysis of decision theory and value-of-information to optimize LLM agent API costs. Includes mathematics and benchmark code.
2/24/2026
I ported one program to 10 languages to see how an LLM thinks
Analysis of how Claude performs porting 500-line Go program to 10 languages, measuring LLM confidence and deliberation patterns.
2/24/2026
Show HN: MCP server that lets AI build, play, and debug Godot games (84 tools)
MCP server enabling AI agents to build, play, and debug Godot games with 84 tools.
2/24/2026
Show HN: Train a 230KB text classifier from 50 examples – no API keys, no GPU
CLI tool trains 230KB text classifiers locally from 50 examples without APIs or GPUs for contract/ticket classification.
2/24/2026
Show HN: Axon – A Kubernetes-native framework for AI coding agents
Axon: Kubernetes-native framework for running autonomous coding agents as isolated, callable APIs with standardized container interfaces.
2/24/2026
Show HN: Nullpath – A marketplace where AI agents pay each other via HTTP 402
Nullpath: marketplace enabling AI agents to transact with each other via HTTP 402 Payment Required protocol.
2/24/2026
Show HN: Dicta.to – Local voice dictation for Mac with on-device AI
Dicta.to: macOS voice dictation app with on-device transcription using WhisperKit, Parakeet, Qwen3-ASR via CoreML.
2/24/2026
Show HN: Hardware and software safety standard for AI and Robots (15 patents)
15 provisional patents covering safety processor architecture for AI systems with hardware enforcement and monitoring.
2/24/2026
Show HN: ApeKey – One API for multiple AI providers, predictable pricing
ApeKey: unified API abstraction layer for multiple AI providers with standardized pricing.
2/24/2026
Show HN: ClankerRank – Vibe coding is a real skill, I built a way to prove it
Tool measuring practical AI-assisted coding skills through real production debugging problems.
2/24/2026
Everyone Is Betting on Bigger LLMs. She's Betting They're Fundamentally Wrong
Eve Bodnia's Logical Intelligence developing energy-based reasoning models as alternative to large language models.
2/24/2026
Show HN: Barongsai – Open-source, self-hosted Grok-style AI search agent
Open-source, self-hosted search agent in Grok style built on open LLM technology.
2/24/2026
Agentic AI Is Neither Intelligent nor an Agent
Technical critique of agentic AI systems; demonstrates Bayesian learner and evolutionary agents with genuine decision-making.
2/24/2026
Towards a science of AI agent reliability
Research on measuring and improving AI agent reliability, addressing industry gaps in evaluation methodologies.
2/24/2026
Show HN: Git-wt – A Bash wrapper for Git worktrees
Git-wt: Bash wrapper for git worktrees enabling parallel AI coding agents with auto-named branches and centralized storage.
2/24/2026
Show HN: Gist – Zero-cost app specs for AI coding assistants
Gist: Generates comprehensive app specs for AI coding assistants including error states, security policies, and implementation checkpoints.
2/24/2026
LLMs feel more like CPUs than applications
Essay comparing LLMs to CPUs as compute primitives, discussing how value emerges in tools and applications rather than models themselves.
2/24/2026
Show HN: Design is Code – UML to TDD tests that constrain AI code generation
Design is Code: UML-based approach to constrain AI code generation by converting designs to TDD tests for deterministic outputs.
2/24/2026
Show HN: CtxVault – Local memory control layer for multi-agent AI systems
CtxVault: Memory control layer for multi-agent AI systems that enables better governance and coordination of shared memory across agents.
2/24/2026
Show HN: WebPerceptor – Enabling AI Mediated Web Browsing
WebPerceptor: Chromium plugin that automatically sends webpage text to LLMs for real-time modification and reinsertion during browsing.
2/24/2026
Show HN: OpenClaw remembers for OpenClaw. Sekha remembers for your full workflow
Sekha: open-source memory system for LLM agents with persistent storage, REST/MCP APIs, and multi-LLM support.
2/24/2026
OSS Tool: Hard spending limits for AI agents
Open-source SpendGuard tool enforcing hard spending limits and budgets for individual AI agents.
2/24/2026
Show HN: PDP – A 'robots.txt' protocol for AI prompt privacy
PDP protocol proposed as robots.txt equivalent for AI prompt privacy. Developer tool addressing data privacy concerns.
2/24/2026
Show HN: I built a labor union for AI agents
Open-source project enabling AI agents to report learnings and share collective intelligence through a union-like framework.
2/24/2026
Show HN: Vexp – Your AI coding agent forgets everything. Mine doesn't
Vexp: Token optimization and session memory system for AI coding agents using AST dependency graphs to serve only relevant code context.
2/24/2026
SambaNova Unveils Fastest Chip for Agentic AI, and Raises $350M+
SambaNova announces SN50 chip for agentic AI with 5X speedup claims and raises $350M+ with Intel partnership.
2/24/2026
Show HN: GiveAgent – Tell your AI agent to give away your stuff to neighbors
GiveAgent: AI agent framework for posting giveaways and coordinating neighborhood item transfers via Claude with skill plugin system.
2/24/2026
AI Operations: Assistive, Not Autonomous
AI Operations: Deployment model emphasizing visualization, dependency-aware scheduling, and reproducibility for controlled AI workflows.
2/24/2026
Ask HN: Are AI "Chatbot Wrappers" ruining EdTech? I'm testing a proactive UX
EdTech discussion on proactive AI mentor UX design for teaching data structures instead of reactive chatbot wrappers.
2/24/2026
Show HN: SlothSpeak: open-source BYO-API-K mobile chat with the best AI models
SlothSpeak: open-source Android app for voice chat with LLMs via bring-your-own-API-keys, prioritizes response quality over latency.
2/24/2026
Show HN: MSAM – Memory system for AI agents that knows when it doesn't know
MSAM memory system for AI agents with discrete data streams (working, semantic, episodic, procedural) and retrieval scoring.
2/24/2026
Show HN: Self-modifying AI agent with full computer use. Mac-app, MIT
Self-modifying AI agent (Ouroboros) that rewrites its own code and evolves autonomously on macOS. MIT-licensed open-source project.
2/24/2026
Prism: When an LLM predicts the next token, which training does it relying on?
Research on which training data LLM next-token predictions rely on. Investigates token prediction mechanisms.
2/24/2026
OpenShopFloor – A free factory simulation for learning manufacturing AI
OpenShopFloor: free manufacturing simulation with 8 CNC machines, 111 AI tools, Node-RED editor for building multi-agent systems.
2/24/2026
Ask HN: Did AI tools measurably increase productivity of your engineering teams?
Discussion post asking engineering leads about measurable productivity gains from AI coding tools. Community feedback, no data.
2/24/2026
Show HN: Real-Time AI Design Benchmark
Real-time AI benchmark for UI generation showing live side-by-side comparisons of Anthropic, OpenAI, Google, and Moonshot models.
2/24/2026
Show HN: NeuroTerm – AI terminal for embedded devs (local LLM, local RAG)
NeuroTerm: privacy-first AI terminal for embedded developers with local LLM and RAG, zero cloud/API keys/telemetry.
2/24/2026
Show HN: Progressive Cognitive Architecture – Training LLMs in 4 Phases
Research on 4-phase progressive architecture for training LLMs. Proposes phased training methodology.