Dev
-
Your AI Writes Tests That Pass. That Is the Problem.
AI-generated tests often create false confidence by checking that code runs, not that it is robust, correct, or secure.
-
The ROM Listing Spirit Lives On in Apple Silicon
From TRS-80 ROM listings to reverse-engineered Neural Engines: Apple silicon is making home computers weird, intimate, and exciting again.
-
COBOL Was Never the Product. The Toll Booth Was.
Anthropic’s COBOL pitch is not about old code. It threatens the costly consulting bottleneck around banks, insurers, and government legacy systems.
-
PageIndex.ai: A persuasive “vectorless RAG” idea—especially for real PDFs
PageIndex.ai uses tree-structured, reasoning-based retrieval for long PDFs – great for traceability, but test latency/cost on messy docs.
-
Gemini’s Fix: The OpenClaw Security Mess and Apple’s Google Bet Paying Off
OpenClaw’s rapid rise turned into a security nightmare – until Gemini stepped in to patch it fast. Apple’s AI partnership suddenly looks like a very smart move.
