Claude Opus 4.5, and why evaluating new LLMs is increasingly difficult (opens in new tab)

(simonwillison.net)

6 pointsjonesn113mo ago1 comments

1 comments

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

j / k navigate · click thread line to collapse

(simonwillison.net)

6 pointsjonesn113mo ago1 comments

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

j / k navigate · click thread line to collapse