Where do you think my confusion came from? All it says is that ai assists in resolving the gyroscope lock path, not why they decided to model the gyroscope lock path to begin with.
replyPlease, keep your offensive comments to yourself when a clarifying comment might have sufficed.
Even worse, the other child comments are speculating (and didn't RTFA either) when the answer is clear in the article.
reply> We found this defect by distilling a behavioural specification of the IMU subsystem using Allium, an AI-native behavioural specification language.
That's the opposite of clear to me.
replyHas the article been updated?
reply2nd paragraph starts with: "We used Claude and Allium"
And later on: "With that obligation written down, Claude traced every path that runs after gyros_busy is set to true"
> distilling
replyA.k.a. as fabricating. No wonder they chose to use "AI".