Ben Golub@ben_golub
@shashj Okay but when pangram says 100%, it is basically always fully or almost fully AI, right?
Any known counterexamples in the wild?
8:21 AM · Jun 21, 2026 · 3.1K Views
Pangram updates its AI detector to resolve false positives triggered by human text mimicking Claude
The tool previously flagged Claude-like human prose as 100% AI
736223.4K
Sentiment
Sentiment building, check back later.
Cluster Engagement
Digg Deeper
No Digg Deeper questions have been answered for this story yet.
Posts from X
Most Activity
Most Activity
VIEWS71REPLIES1
thebes@voooooogel
@gleech @ben_golub @shashj the new pangram model distinguishes this one nowadays!
gavin leech (Non-Reasoning)@gleech
@ben_golub @shashj lots of adversarial counterexamples but that's not what OP is on about
2hViews 71Likes 0Bookmarks 0
LIKES2
gavin leech (Non-Reasoning)@gleech
@voooooogel @ben_golub @shashj you love to see
2hViews 18Likes 2
gavin leech (Non-Reasoning)@gleech
@ben_golub @shashj lots of adversarial counterexamples but that's not what OP is on about
2hViews 55Likes 2
Ben Golub@ben_golub
@gleech @voooooogel @shashj I wonder if it knows the one from the daughter of @boazbaraktcs
2hViews 8