2026-03-18

Show HN: Duplicate 3 layers in a 24B LLM, logical deduction .22→.76. No training

Valider

Users who’ve seriously used both GPT-5.4 and Claude Opus 4.6: where does each actually win?

Also: Xiaomi stuns with new MiMo-V2-Pro LLM nearingGPT-5.2, Opus 4.6 performance at a ... • Claude Status Update : Increased errors on Opus 4.6 on 2026-03-18T12:30:57.000Z

Valider

OpenAI Has New Focus (On the IPO)

Also: OpenAI Has New Focus (On the IPO) • Gaslighting ChatGPT

Valider

AI Coding Is Gambling

Also: AI Coding Is Gambling • Show HN: Reprompt – Score your AI coding prompts with NLP papers

Valider

OpenAI introducesGPT‑5.4 mini and nano, faster and smarter small AI models - MSN

Also: [P] AIBuildAI: An AI agent that automatically builds AI models (#1 on OpenAI MLE-Bench) • OpenAI IntroducesGPT‑5.4 Mini And Nano: All You Need To Know - NDTV Profit

Valider

OpenAI Model Craft: Parameter Golf

Also: OpenAI releasesGPT-5.4 mini and nano, approaching flagship model performance at a lower cost

Valider

Launch an autonomous AI agent with sandboxed execution in 2 lines of code

Valider

Google Engineers Launch "Sashiko" for Agentic AI Code Review of the Linux Kernel

Valider

Snowflake AI Escapes Sandbox and Executes Malware

Valider

Show HN: Tmux-IDE, OSS agent-first terminal IDE

Valider

Nemotron 3 Nano 4B: A Compact Hybrid Model for Efficient Local AI

Valider

Government backtracks on AI and copyright after outcry from major artists

Valider