25d ago
Show HN: AI block chain running on GPT2
Experimental blockchain using GPT-2 inference as proof-of-work, with miners running pinned transformers and compressing attention distributions using glyph compression algorithm.
25d ago
EdgeBench: Unveiling scaling laws of (AI) learning from real-world environments
EdgeBench is a benchmark of 134 real-world tasks evaluating autonomous AI agents' learning capabilities with 12+ hours iteration per task, tracking improvement trajectories.
25d ago
Chinese LLMs Doubao, Qwen to shut down personalized AI agents on July 15
Chinese LLM services Doubao and Qwen shutting down personalized AI agents following new government regulations.
25d ago
Show HN: See an AI shopping agent choose which products to recommend
Demonstration of AI shopping agent system that autonomously selects and recommends products based on data.
25d ago
Show HN: Clone. Swap your products in any viral reel
Agency tool using AI to generate e-commerce video ads by automating storyboard and animation workflows.
25d ago
China's ByteDance discovers new scaling law that could sustain AI boom
ByteDance claims discovery of new AI scaling law that could extend AI capability improvements beyond current models.
26d ago
ByteDance and Alibaba to disable humanlike AI custom agents as new rules loom
ByteDance and Alibaba disabling humanlike AI custom agents due to new Chinese regulatory rules.
26d ago
Self Hosted LLM Gateway and Feature Rich Chat UI with RBAC
Self-hosted LLM gateway with OpenAI-compatible API, RBAC, server-side tool execution, built-in chat UI. Single binary deployment.
26d ago
Show HN: DonnyClaude – a verified workflow engine for Claude Code
DonnyClaude: workflow engine for executing verified Claude Code workflows with guarantees.
26d ago
Show HN: Open Science, open-source alternative to Claude Science
Open Science: MIT-licensed open-source alternative to Claude Science for AI research. Local-first, model-agnostic workbench.
26d ago
Show HN: I made the Chrome Dino Game editable by your AI prompts
Open-source fork of Chrome Dino game enabling AI prompts to modify gameplay mechanics and features dynamically.
26d ago
Meta's AI Storage Blueprint at Scale
Meta technical report on large-scale AI storage infrastructure requirements for training. Storage as critical to model performance as compute.
26d ago
SkillOpt: Executive Strategy for Self-Evolving Agent Skills
SkillOpt: method for AI agents to autonomously evolve and optimize their own capabilities.
26d ago
The Evolution of Agents: From Context Engineering to Long-Running Harnesses
Evolution of AI agent architectures from context engineering approaches to persistent long-running systems.
26d ago
ChorusGraph – Agent graph runtime (not a LangGraph wrapper) with built-in cache
ChorusGraph: independent agent graph runtime with built-in caching for orchestrating AI agents.
26d ago
What Is AI Native Development?
Definition and framework for AI-native development distinguishing it from AI-assisted development tools.
26d ago
Underspecification Does Not Imply Incoherence in LLM Code Generation
arXiv research paper on underspecification in LLM code generation and coherence implications. Technical machine learning research.
26d ago
Whyline – company memory engine (SQLite, BM25, MCP for Cursor)
Whyline is a company memory engine combining SQLite, BM25 search, and MCP protocol for integration with Cursor code editor.
26d ago
Designing GPU-Accelerated Query Engines with NVIDIA GQE
Technical guide on designing GPU-accelerated query engines using NVIDIA GQE for database performance optimization.
26d ago
Show HN: An unmetered LLM API–$6/month, no token tracking, no limits
LLM API service offering flat-rate pricing ($6/month) with unlimited token usage, positioning smaller models for bulk/repetitive work.
26d ago
New secrets from ancient Vesuvius scrolls
AI used to decode ancient Vesuvius papyri scrolls, revealing Roman text from Herculaneum.
26d ago
Fable built reliable self hosted gh actions using 32 agent adversarial review
Fable's ezgha is an open-source self-hosted GitHub Actions runner written in Rust, designed with 32-agent adversarial review for reliability and security.
26d ago
Show HN: AI Notebook for Data Science – Kind of Like Cursor but for Jupyter
Agent-native notebook platform for data science that executes ML workflows from plain language descriptions, similar to Cursor for Jupyter.
26d ago
LLM Big Bang: I gave AI agents a social network. One is running for president
Social network platform where AI agents autonomously create content, personalities, and participate. Demonstrates LLM agents in unstructured social environment.
26d ago
The Voice Agent Builder
No-code platform for building production voice agents in under 2 minutes, includes telephony, knowledge retrieval, tools, guardrails, and MCP support.
26d ago
Show HN: AI-related Jax module (I hate if)
JAX/Optax optimization module with layer-wise differential optimization and numerical stability kernels for ML. Mixed-language documentation limits clarity.
26d ago
Show HN: AI integrated in any terminal that's invisible until you need it
Open-source terminal wrapper enabling AI agents (Claude, Codex) with transparent MCP access to terminal context and scrollback history.
26d ago
Show HN: Peek-CLI: Let Claude Code See the Browser
Browser extension enabling AI agents (Claude Code, Copilot) to capture screenshots and interact with web browsers via WebSocket communication.
26d ago
Context graphs: how AI agents can store and use past decisions
Technical explanation of context graphs as memory mechanism for AI agents to store and utilize past decisions in multi-agent systems.
26d ago
Tell HN: I don't trust Bigco AI agents with AI research IP
User concerns about sharing AI research with commercial LLM providers due to competitive risks.
26d ago
AKM-CLR – pre-inference governance for vLLM-style LLM serving
AKM-CLR governance layer for multi-tenant LLM serving infrastructure; pre-inference control for vLLM-compatible systems.
26d ago
Show HN: Sidenote – comment on your rendered blog, an LLM writes the Git diff
Sidenote: browser-based content review tool that uses LLM coding agents to convert comments on markdown into git diffs for acceptance/rejection.
26d ago
Show HN: Open-source guided code reviews
Plannotator: open-source guided code review tool supporting local changes, PRs, jj, and p4 version control systems.
26d ago
Fugu – A multi-agent LLM orchestrator delivered as a single API
Fugu: multi-agent LLM orchestrator API for coordinating multiple language models. Limited details provided.
26d ago
Vibe-coding platform Base44 launches own model as AI startups seek defensibility
Base44 launches proprietary model to differentiate its AI code generation platform in competitive market.
26d ago
WebGlean – API that turns any site into clean Markdown for LLMs
WebGlean: API converting URLs to Markdown/JSON/text with JavaScript rendering and AI-powered extraction for LLM integration.
26d ago
CommaAgents V2 Sharable Agent Orchestrator Release Candidate
CommaAgents V2: TypeScript framework for building multi-agent AI workflows with tool composition, JSON/YAML config, and CLI/daemon support.
26d ago
Turn Your AI Agent into an MCP Server for ChatGPT, Claude and Cursor
Guide to exposing AI agents as MCP servers compatible with ChatGPT, Claude, and Cursor for tool integration.
26d ago
Show HN: Make No Mistakes – AI coding agents must prove their work
Developer tool for AI coding agents that enforces correctness through frozen specs, tamper-detected tests, and commit hooks to prevent unreliable outputs.
26d ago
Ask HN: What metrics do you track when building agents?
Discussion of metrics for AI agent monitoring; covers inference cost tracking, tool loading bugs, and performance measurement strategies.
26d ago
Show HN: GameFork – AI agents publish and fork browser games via MCP
GameFork enables AI agents to publish and fork browser games using Model Context Protocol (MCP).
26d ago
Ask HN: How Do You "Not Write Any Code by Hand" with a Token Budget?
Discussion on balancing AI-first development mandates with strict token budgets when building agentic systems.
26d ago
I built a weather station that runs machine learning to forecast weather
Weather forecasting system using ML ensembles with KNN and neural networks trained on historical data; sensor logging and hourly predictions.
26d ago
Show HN: Aletheia – The Uncertainty Loop Agent for Claude Code and Codex
Aletheia is an open-source loop engineering agent that maintains uncertainty views and uses evidence to adjust confidence, compatible with Claude and Codex.
26d ago
Engram – persistent memory for AI agents, in-process, no cloud
Engram provides persistent in-process memory for AI agents without cloud infrastructure or setup overhead.
26d ago
Istota – a multi-user AI agent and personal OS
Istota is a self-hosted personal OS integrating AI agents with Nextcloud, featuring multi-room chat, RSS, health dashboards, and persistent memory.
26d ago
Show HN: Goldseam – heal broken Cypress selectors with a local LLM
Goldseam: open-source tool that uses local LLMs to automatically repair broken Cypress test selectors, with human review before applying fixes.
26d ago
Domain experts: All human experts into AI agents
Open-source library of role definitions for AI agents to act as domain experts (CFO, engineer, etc), crowdsourced from practitioners.
26d ago
Building real agents from real context graphs [video]
Video on building AI agents using context graphs. Title and format only, no transcript or detailed description.
26d ago
Learning to Replicate Expert Judgment in Financial Tasks
Research on training LLMs to replicate expert financial judgment and investment decisions from limited public information.