I find it odd that the US is touting the use of AI in Iran AND blaming a "targeting error" for bombing the shit out of a girls' school, and no one in the media seems to be asking the obvious question.

March 12, 2026 - 02:20 UTC

8
4
250
The simplest solution doesn't involve AI. A CENTCOM planning cell used DIA data (see NYT rpt) and did not review its currency or request update. They will plead "ops tempo" (operations pace rqmts), but there is no excuse for verification and review rqmts. Humans, not AI, failed.
0
0
4
It was the first question I had this morning. bsky.app/profile/ryanhealy.bsky.social/post/3mgs6npc6h22k
1
0
3
🫶
0
0
1
bsky.app/profile/10periods.bsky.social/post/3mgs6oqsxkc2e
0
0
3
I don't know enough about military weapons to know how to start digging into this, but I'm willing to try if someone points me in the right direction.
1
0
14
I believe the best answer is the simplest: AI flagged the target using obsolete intelligence. No human validated the target selection before approving it.
2
0
4
Kinda missed the defense scoop
0
0
0
What does AI mean in this context?

Like, I can totally imagine a model (random forests based?) that takes ship locations, target destinations, and air defense coverage and builds Tomahawk flight paths that are least likely to be shot down

And that is NOTHING like asking a chatbot for targets
1
0
1
We should just stop saying AI and demand everyone else does too

Did you use a chatbot? A regression? A clustering algorithm?

Be more specific
0
0
1
HA HA HA HA
No.

They're not using Anthropic anymore because Anthropic told them to go fuck themselves in the face because basically using Claude for War violates their ToS.

This administration RUNS on spin.
0
0
0
The targeting error was that they didn't have a human review the targeting recommendations before passing them along to be acted on.
1
0
2
💯
0
0
0