Agentic test processes, LLM benchmarks, and other notes on agentic coding · HackerLangs