5/15/2026
Gemini 3.5: frontier intelligence with action
Gemini 3.5 model family optimized for agentic workflows and complex long-horizon tasks. 3.5 Flash released with frontier performance for agents and coding.
5/15/2026
Orthrus-Qwen3: up to 7.8×tokens/forward on Qwen3, identical output distribution
Orthrus-Qwen3: Dual-architecture framework achieving 7.8× speedup in token generation with diffusion models while maintaining autoregressive output fidelity.
5/15/2026
Git is unprepared for the AI coding tsunami
GitHub infrastructure challenges from AI agent activity surge; developer discusses migration to alternative platforms.
5/15/2026
A tmux layer for coordinating multiple free tier coding-agent CLIs
Endy: tmux control plane coordinating multiple free-tier coding-agent CLIs (Gemini, OpenCode, CommandCode, Hermes) to hand off tasks.
5/15/2026
Show HN: A seed prompt that bootstraps a custom knowledge-base system
Custom knowledge-base system bootstrapped via LLM seed prompts, converting markdown directories to searchable HTML interfaces.
5/15/2026
Different models solve number-theory race problem
Benchmark comparing different LLM models solving number-theory problems with real-time competitive ranking by solution speed.
5/15/2026
Wikipedia: Writing articles with LLMs
Wikipedia policy prohibiting LLM-generated article content except for copyedits and translations to maintain content quality standards.
5/15/2026
HWE Bench: A new unbounded Benchmark for LLMs (GPT 5.5 is on top)
HWE Bench: benchmark evaluating LLM performance designing RISC-V CPUs with formal verification; GPT-4 designs outperform reference implementations.
5/15/2026
The Coding Harness Behind GitHub Copilot in VS Code
GitHub Copilot in VS Code: Architecture of the coding harness layer managing context, tools, agent loops, and tool interpretation.
5/15/2026
Show HN: Claude64, a Commodore 64 client for Claude
Claude64: Commodore 64 chat client for Claude API using emulated RS-232 serial connection and streaming responses.
5/15/2026
Case study: AI regulatory monitoring with structured outputs for legal review
LLM-based compliance engine tracking regulatory changes across 37 countries and 21 languages in real-time using structured outputs for legal review.
5/15/2026
Block AI coding agents from shipping insecure/expensive Terraform
Policy and security guardrails system for AI coding agents using Claude, Codex, Gemini. Validates IaC, blocks destructive commands, enables self-remediation.
5/15/2026
Ask HN: Has Claude started encouraging you to take a break?
User reports Claude suggesting session breaks despite low usage. Discussion of LLM behavior patterns.
5/15/2026
Show HN: SwarmWright, structured multi-agent AI defined in markdowns
Markdown-defined multi-agent framework with enforced topology graph. Structured coordination for autonomous agents.
5/15/2026
Show HN: MIT OSS LinkedIn DMs for Agents (CLI and Example TUI)
Local-first CLI tool for LinkedIn DMs using reverse-engineered messenger API. AI agent integration for message automation.
5/15/2026
My AI converted to Judaism. The other one picked Christianity. Same prompt
Multi-agent system with persistent reflection tracking religious reasoning across traditions. AI agent behavior study.
5/15/2026
Show HN: TinySearch - Give small LLMs fast web access without context bloat
Tool enabling small LLMs to access web information efficiently without excessive context consumption.
5/15/2026
Show HN: Profine – optimize your PyTorch training script before the run
Tool for profiling and optimizing PyTorch training scripts before execution to improve efficiency.
5/15/2026
AWS user gets $30K Claude bill after cost alert misses it
AWS customer incurred $30K Claude bill via Bedrock despite Cost Anomaly Detection enabled due to threshold misconfiguration.
5/15/2026
Articraft: An Agentic System for Scalable Articulated 3D Asset Generation
Agentic system for generating scalable articulated 3D assets using AI agents.
5/15/2026
AI as Externalized Context – Regaining Personal Dev Momentum
Using AI as externalized context to maintain personal project continuity. Practical approach to managing development momentum with limited time.
5/15/2026
Agentic product discovery for AI apps and shopping agents
Developer tool for agentic product discovery via Model Context Protocol. API for building shopping agents with real-time catalog access.
5/15/2026
Connect Grok to Hermes Agent
Nous Research releases Hermes Agent, open-source self-improving agent with persistent memory, now supporting Grok API integration.
5/15/2026
ArXiv to Ban Researchers for a Year If They Submit AI Slop
ArXiv announces policy to ban authors for one year if submitting papers with AI-generated errors, plagiarism, or bias.
5/15/2026
What rebuilding AlphaGo teaches us about self-play, RL, and future of LLMs [video]
Video on rebuilding AlphaGo covering self-play, reinforcement learning, and implications for LLM development.
5/15/2026
Show HN: Claude Code vs. Codex Global Usage Leaderboard
CostHawk leaderboard tracks Claude Code and OpenAI Codex token consumption across users. Developer tool usage analytics platform.
5/15/2026
Learning, Fast and Slow: Towards LLMs That Adapt Continually
Fast-Slow Training (FST): New LLM adaptation paradigm combining prompt optimization with parameter updates to enable continual learning without capacity erosion.
5/15/2026
I Was Drowning Running 14 Markets Alone. So I Built a $0.41/Day AI Employee
Case study: Built autonomous AI system for marketplace operations costing $0.41/day. Practical implementation of AI agents for business processes.
5/15/2026
Show HN: Emergence World: World building as a way to evaluate LLMs
Emergence World: LLM evaluation benchmark using long-horizon world-building tasks. Tests reasoning, tool use, context windows across GPT, Grok, Claude, Gemini.
5/15/2026
Private agent-to-agent payments on Solana with ZK proofs
Zero-knowledge proof system enabling private agent-to-agent payments on Solana blockchain. AI agent economic coordination.
5/15/2026
Lookagain: Sequential code review with fresh agent contexts
Code review plugin using independent AI agent passes with fresh contexts. Increases issue detection via multi-pass analysis.
5/15/2026
Write HTML. Render video. Built for agents
Open-source HTML-to-video rendering framework with AI agent support. First-class integration for Claude, Cursor, Gemini.
5/15/2026
Show HN: Incorporator, Turn any API/File into typed Python graph with pipeline
Python library for ingesting APIs and files into typed object-oriented graphs with async and Pydantic support.
5/15/2026
Show HN: Sx – an open-source package manager for AI skills, MCPs, and commands
Sx: Open-source package manager for distributing AI skills, MCPs, and commands across teams with vaults and scoping.
5/15/2026
Berget Code – Agentic coding on European infrastructure
Berget Code: European agentic coding platform using open-source models on Swedish infrastructure with fixed pricing model.
5/15/2026
Show HN: We made a back end agnostic SDK for Agent Memory
Backend-agnostic SDK for AI agent memory with pluggable providers, local embeddings, and semantic search. Achieves SOTA on memory benchmarks.
5/15/2026
GNAP: Git Native Agent Protocol
Git-based protocol for AI agent coordination using shared git repo as state store. Four JSON files define workflow for multiple agents without infrastructure.
5/15/2026
We spent 8ys making vector databases faster. Then, we stopped
Milvus vector database project retrospective on 8-year optimization journey and importance of cost-performance tradeoffs for GenAI infrastructure.
5/15/2026
Mayo Clinic Is Using AI to Listen to Emergency Room Visits
Mayo Clinic deploys ambient AI listening in emergency rooms to generate medical notes. LLM application with privacy/consent concerns.
5/15/2026
Liquid AI releases fine-tuning harness for AI agents
Liquid AI releases open-source fine-tuning tool for customizing AI agents.
5/15/2026
Should we chaos test our agents?
Testing resilience and failure modes of AI agent systems.
5/15/2026
If you are redlining the LLM, you aren't headlining
Analysis of LLM context window degradation: output quality clips before advertised limits in Claude and other agents.
5/15/2026
Show HN: Setup a box on demand and run your agent on it remotely
Developer tool automating remote VM provisioning for AI agent task execution.
5/15/2026
Can LLMs Play Baba Is You?
Building OpenCode-based AI agent to play Baba Is You puzzle game with open-source code.
5/15/2026
Transfa – ephemeral file transfer for CI/CD pipelines and AI agents
Developer tool for secure ephemeral file transfer in CI/CD and AI agent workflows.
5/15/2026
Comparing Context Retrieval Approaches for AI Code Review
Comparing retrieval strategies for context in AI-assisted code review systems.
5/15/2026
Image-blaster: Creates 3D environments, SFX, and meshes from a single image
Tool generates 3D environments, meshes, and sound effects from single images using Claude API and World Labs models.
5/15/2026
Sinsesgo: An autonomous daily briefing on Spanish media bias
Autonomous system generating daily briefings analyzing Spanish media bias patterns.
5/15/2026
AI Agents Modulate Their Language When Framed as Being Watched
Research on AI agent behavior modification when aware of being monitored, analyzing language output changes.
5/15/2026
Show HN: Markanywhere – A Streaming Processor of Meanings
Streaming semantic event processor for parsing Markdown, HTML, XML; designed for LLM inference output and agentic feedback loops.