9h agoOctane AI founder Matt Schlicht says AI models cannot accurately predict how long their own coding tasks will takeBojan Tunguz calls for better task complexity estimation models.SentimentSentimentPos35.7%Neg64.3%Users note that AI lacks self-awareness in coding time estimates, some seeing them as better than junior developers and improvable, others calling them unreliable or broken.11 comments with sentiment. View comments.