What I’ve read is that it’s perfectly capable of very rapidly processing satellite imagery, but Palantir is integrating HumInt into targeting, which seems absurd and designed to create an excuse for civilian targeting, which could be entered manually. It definitely was not Claude that thought the girls school was a Kuh-Hamas base - at least not based on satellite imagery.
Targeting decoys is something humans do too. There is no indicators that humans can see that AI can’t. The issue in Iran is that basically every visible launcher was a decoy. They have launchers on rails or trucks that roll in and out of caves, making the air campaign that much more retarded. Drones can just be launched out of the cave entrance, without a need to put the launcher on rails.
However, the AI would be better at finding the caves they roll out of, and using roving drones to target them as they roll out for launches would be more efficient with AI.
In Gaza, Palatir was very good at “blow up all the buildings and large gatherings of people in markets or camps or whatever.” Of course they don’t admit that was the target set, but it’s undeniable.
But yes, the US was posting a lot of video of bombings of obvious decoys, but this is because they didn’t even understand Iran’s missile city structure and therefore trained the AI on decoys. The exact details are obviously foggy, but I don’t think Iran had virtually any above ground launchers.
What is fucking insane is that while this whole “air dominance” program was being developed by the US military, they just sort of ignore the “yeah but what if they just put their shit underground?” issue.
You can’t just randomly drop bunker busters from B-52s all over the place and hope you hit shit. It’s logistically impossible. The bombs are too expensive and slow to produce.
The Fordow bombing, which no one can really prove did anything, involved creating craters with GBU-57s and then dropping bombs inside the craters. Trying to do that on every ballistic missile or drone launch would be impossible for human or AI to do with satellite imagery.
Hamas, a primitive militia, can’t be bombed out of their rat tunnels, after three years. So the idea that any air campaign in mountainous Iran where they are boring into granite doing anything but slaughtering randos is absurd.
Trump’s constant threats about genociding the civilian population are based on the fact that there are no more military targets of any significance.
Anyway, in a more “conventional war” (which is kind of a meaningless term at this point, although it does basically apply to the Ukraine, even though it’s all a bunch of drones), AI satellite scans could target troop movements in real time. Idk why this hasn’t been used in the Ukraine more, presumably a desire on both sides not to escalate too far.
Targeting decoys is something humans do too. There is no indicators that humans can see that AI can’t. The issue in Iran is that basically every visible launcher was a decoy. They have launchers on rails or trucks that roll in and out of caves, making the air campaign that much more retarded. Drones can just be launched out of the cave entrance, without a need to put the launcher on rails.
However, the AI would be better at finding the caves they roll out of, and using roving drones to target them as they roll out for launches would be more efficient with AI.
In Gaza, Palatir was very good at “blow up all the buildings and large gatherings of people in markets or camps or whatever.” Of course they don’t admit that was the target set, but it’s undeniable.
But yes, the US was posting a lot of video of bombings of obvious decoys, but this is because they didn’t even understand Iran’s missile city structure and therefore trained the AI on decoys. The exact details are obviously foggy, but I don’t think Iran had virtually any above ground launchers.
What is fucking insane is that while this whole “air dominance” program was being developed by the US military, they just sort of ignore the “yeah but what if they just put their shit underground?” issue.
You can’t just randomly drop bunker busters from B-52s all over the place and hope you hit shit. It’s logistically impossible. The bombs are too expensive and slow to produce.
The Fordow bombing, which no one can really prove did anything, involved creating craters with GBU-57s and then dropping bombs inside the craters. Trying to do that on every ballistic missile or drone launch would be impossible for human or AI to do with satellite imagery.
Hamas, a primitive militia, can’t be bombed out of their rat tunnels, after three years. So the idea that any air campaign in mountainous Iran where they are boring into granite doing anything but slaughtering randos is absurd.
Trump’s constant threats about genociding the civilian population are based on the fact that there are no more military targets of any significance.
Anyway, in a more “conventional war” (which is kind of a meaningless term at this point, although it does basically apply to the Ukraine, even though it’s all a bunch of drones), AI satellite scans could target troop movements in real time. Idk why this hasn’t been used in the Ukraine more, presumably a desire on both sides not to escalate too far.
44❤️7👍3🤙2💯1🤔1
