ambiguity

This blog post explores parsing ambiguities in C23 involving `auto` as a type inference specifier or storage-class specifier, showing how GCC and Clang disagree on parsing declarations like `auto x = 67;` when `x` is a typedef, and how attributes complicate the situation.

A discussion on the varying definitions of 'AI agents' and whether everyone means the same thing when using the term.

A discussion about tasks people think AI agents are ready for but aren't, highlighting the challenge of interpreting ambiguous human input like annoyed but unclear messages.

A developer describes the challenge of building multi-agent AI assistants that fail to handle unexpected situations gracefully, relying on explicit rules that lead to a whack-a-mole problem instead of enabling autonomous reasoning about ambiguity.

Introduces PRIG, a gradient attribution method that localizes prompt ambiguity in large language models by training a linear probe to distinguish clear from ambiguous prompts and attributing the probe score to token representations in the residual stream, achieving strong performance on synthetic and human-written benchmarks.