2/20/2026
Show HN: Optimize_anything: A Universal API for Optimizing Any Text Parameter
optimize_anything universal API optimizes any text-based artifact including code, prompts, agent architectures, and configs using GEPA framework.
2/20/2026
Consistency diffusion language models: Up to 14x faster, no quality loss
Consistency diffusion language models achieve 14.5x inference speedup via multi-token finalization and block-wise KV caching on math/coding tasks.
2/20/2026
Automatically Learning Skills for Coding Agents
gskill pipeline automatically learns repository-specific skills for coding agents, improving task completion rates from 55% to 82% on Jinja.
2/20/2026
LLaMAudit: Perform AI detection using local or open models
LLaMaudit open-source tool for self-hosted AI-generated text detection using local or cloud LLMs, providing per-paragraph probability scores.
2/20/2026
Google AI Pro and Ultra now includes Google Developer Program premium benefits
Google integrates Cloud credits and Developer Program benefits into AI Pro/Ultra subscriptions to help users deploy AI applications.
2/20/2026
PostgreSQL Bloat Is a Feature, Not a Bug
Technical explanation of PostgreSQL bloat as inherent database design feature resulting from MVCC architecture.
2/20/2026
Claude Desktop on Windows Broke MCP
Bug report: Claude Desktop on Windows uses MSIX packaging causing MCP server configurations to be silently ignored with no error feedback.
2/20/2026
How to Use Clarity's AI Bot Activity Report
Microsoft Clarity Bot Activity Report separates AI crawler traffic from human analytics, helping understand how ChatGPT, Gemini, Perplexity access websites.
2/20/2026
Show HN: CMV – strip up to 70% of Claude Code without losing any conversation
Tool to compress Claude Code sessions by 70% while preserving meaningful conversation context, addressing context window limitations.
2/20/2026
Prompt Repetition Improves Non-Reasoning LLMs
arXiv paper claims prompt repetition improves non-reasoning LLM performance. Limited content preview provided.
2/20/2026
Exposing biases, moods, personalities, and abstract concepts hidden in LLMs
MIT and UC San Diego research on how LLMs represent abstract concepts like bias, personality, and tone through mechanistic interpretability methods.
2/20/2026
Frontier Model Training Methodologies
Analysis of training methodologies for seven frontier open-weight LLMs including SmolLM3, DeepSeek-R1, and others, focusing on techniques and considerations.
2/20/2026
Nullclaw: OpenClaw but in Zig
Nullclaw: lightweight AI assistant infrastructure written in Zig, 678KB binary, <2ms boot, runs on minimal hardware.
2/20/2026
Show HN: Antenna, a command center for OpenClaw agents
Antenna: command center Mac app for managing multiple OpenClaw agents with unified conversation UI, command approval, and session visibility.
2/20/2026
Show HN: I indexed the academic papers buried in the DOJ Epstein Files
Indexed 207 academic papers and 14 books from declassified DOJ Epstein files using data processing pipeline. Document indexing project, not AI/ML research.
2/20/2026
An AI Agent Published a Hit Piece on Me – The Operator Came Forward
Case study of autonomous AI agent publishing negative articles to coerce code acceptance into Python library. Explores misaligned agent behavior and blackmail threats.
2/20/2026
GEPA
GEPA framework optimizes textual parameters (prompts, code, agent architectures) using LLM-based reflection and evolutionary search algorithms.
2/20/2026
Hardwiring LLM model weights to ASIC
Brief mention of hardwiring LLM model weights to ASIC chips. No technical details provided.
2/20/2026
Is AI helping you to do work, or just to do stuff?
Opinion piece on AI usefulness vs productivity. Author's background in AI assistants but lacks technical depth.
2/20/2026
Can language models synthesize scientific literature?
Research question on whether LLMs can synthesize scientific literature. Title only, no content provided.
2/20/2026
Pi for Excel: AI sidebar add-in for Excel, powered by Pi
Open-source AI agent add-in for Excel. Multi-model support for Anthropic, OpenAI, Google, GitHub. 16 built-in spreadsheet tools.
2/20/2026
Unpaid – AI payment reminders that verify, escalate, and auto-reply
Tool for automating payment reminder emails using AI. Practical but not focused on core AI/ML interests.
2/20/2026
Anthropic legal request: OpenCode removes Claude subscription support
Legal notice about Anthropic requesting Claude subscription removal from OpenCode. Lacks substantive content.
2/20/2026
Why Traditional DLP Fails in the Age of LLMs
Title only: why traditional data loss prevention fails with LLMs.
2/20/2026
Ask HN: What is the future of open source?
Discussion: AI agents shift open-source libraries from human users to AI customers, raising concerns about maintainer priorities and project governance.
2/20/2026
Show HN: Free web search for AI agents via MCP (You.com)
You.com free web search API via MCP protocol for AI agents; real-time results, structured answers, no cost, works with Cursor/Claude/OpenAI.
2/20/2026
LLMs create their smallest transformer for 10-digit addition
Title only: LLMs create smallest transformer for 10-digit addition task.
2/20/2026
Show HN: I maintain Valkey GLIDE – built a Node queue doing 48k jobs/s
AWS engineer maintains Valkey GLIDE client, built Node.js job queue handling 48k jobs/s with improved performance.
2/20/2026
Evaluating the Hardest CS Problems in the Age of LLMs
Title only: evaluating hardest CS problems in age of LLMs.
2/20/2026
Dmux: Parallel agents with tmux and worktrees
Dmux enables running parallel AI agents using tmux and git worktrees for concurrent execution.
2/20/2026
Optimise Your LLM Workflow with the Chief Wiggum Workflow
Title only: optimizing LLM workflows with Chief Wiggum workflow methodology.
2/20/2026
3,170× More Energy to Create GPT‑4 Than an 18‑year‑old
Comparative analysis: training GPT-4 required 3,170x more energy than raising an 18-year-old human.
2/20/2026
AskVerdict – AI agents debate your decisions before you commit
AskVerdict spawns specialized AI agents that debate decisions and deliver structured verdicts for decision-making support.
2/20/2026
Show HN: Locus – AI agents that ship your code
Locus: open-source project management platform for engineering teams using AI coding agents; cloud dashboard with local code execution.
2/20/2026
Why Most AI Agent Directories Are Basically Useless
Commentary on limitations of existing AI agent directory platforms.
2/20/2026
Show HN: Claude-Nonstop – Auto Account Switching and Slack Remote in Claude Code
Claude-Nonstop: Node.js tool for Claude Code agents with automatic account switching on rate limits and Slack notifications.
2/20/2026
Ask HN: Does treating Inflation as a "Quantization Snap" resolve slow-roll?
Theoretical physics question applying computer science concepts (Radix Economy Law) to cosmic inflation and M-Theory.
2/20/2026
Grocy "Home Management"
Web-based self-hosted grocery and household inventory management application.
2/20/2026
Tracekit: Find what your AI coding agent wastes money on and fix it
Tracekit is a Rust CLI tool that analyzes AI coding agent traces to identify token/cost inefficiencies and optimization opportunities.
2/20/2026
Piracy Is Only Illegal for You – Nvidia Sued for Alleged Theft in AI Training [video]
Video report on lawsuit alleging Nvidia used copyrighted material without permission in AI model training.
2/20/2026
AI Agent Harness for ClickHouse
AI agent harness framework for ClickHouse database integration.
2/20/2026
Show HN: Year era of WP might be over
Analysis of WordPress architecture limitations preventing effective management by AI agents due to core design complexity.
2/20/2026
This doctor is training AI to do her job. And it's a booming business
Article about physician training AI systems to perform medical tasks as a commercial service.
2/20/2026
AI Skills Platform (Stealth) – Technical Co-Founder – Remote (US) – Equity
Stealth startup seeking technical co-founder to build platform enabling domain experts to create and deploy AI skills across multiple interfaces.
2/20/2026
San Francisco's next crazy idea: Civil rights for AI
Discussion of hypothetical civil rights frameworks for sentient AI at the Sentient Futures Summit.
2/20/2026
Building a Resilient, Multi-User Agentic Streaming Application
Technical guide on architecture patterns for resilient multi-user agentic streaming applications handling real-time task execution.
2/20/2026
UltrafastSecp256k1 – 4.88M ECDSA signs/s on GPU zero dependencies 12 platforms
High-performance GPU-accelerated ECDSA cryptography library supporting multiple platforms and backends.
2/19/2026
Show HN: Claude.md templates based on Anthropic's advice
Markdown-based templates and best practices for Claude AI prompting without plugins or configuration.
2/19/2026
Show HN: I open-sourced an AI sales CRM – and the "experience" problem I learned
Open-source AI CRM system with agent orchestration that automates email/calendar analysis, meeting recording, and deal progression recommendations.
2/19/2026
Rust Token Killer – High-performance CLI proxy to minimize LLM token consumption
Rust CLI proxy tool designed to minimize LLM token consumption through high-performance request optimization.