Claude Opus 4.5, and why evaluating new LLMs is increasingly difficult | Hacker News

Hacker Newsnew | past | comments | ask | show | jobs | submit

		Claude Opus 4.5, and why evaluating new LLMs is increasingly difficult (simonwillison.net)
		6 points by jonesn11 27 days ago \| hide \| past \| favorite \| 1 comment

musha68k 27 days ago [–]

Prompt injections + context window engineering are the combined Archilles heel of the "agentic revolution".

Guidelines | FAQ | Lists | API | Security | Legal | Apply to YC | Contact