2026-03-05

Introducing GPT 5.4 (OpenAI)

Also: OpenAI's Altman takes jabs at Anthropic, says government should be more powerful than companies • OpenAI introduces GPT-5.4 with more knowledge-work capability

Valider

Pentagon Formally Labels Anthropic Supply-Chain Risk, Escalating Conflict

Also: A U.S. official confirmed that the Pentagon has designated Anthropic a “supply-chain risk.” • Anthropic is back to the table having conversation with Pentagon

Valider

Qwen3 vs Qwen3.5 performance

Also: Did we figure out a system prompt to Jailbreak Qwen3.5? • exllamav3 QWEN3.5 support (and more updates)

Valider

Real-time generative projection mapping with Claude Code

Also: Real-time generative projection mapping with Claude Code • I used Claude Code to build a better Claude Code — 4 agents, 12 skills, self-improvement loop. Open for feedback

Valider

Phi-4-reasoning-vision and the lessons of training a multimodal reasoning model

Valider

Relicensing with AI-Assisted Rewrite

Valider

The L in "LLM" Stands for Lying

Valider

Launch HN: Vela (YC W26) – AI for complex scheduling

Valider

Show HN: PageAgent, A GUI agent that lives inside your web app

Valider

Show HN: Jido 2.0, Elixir Agent Framework

Valider

AI and the Ship of Theseus

Valider

Labor market impacts of AI: A new measure and early evidence

Valider