17h agoJeremy Howard, fast.ai co-founder, says Opus 4.8 reduces over-agentic coding behavior compared to GPT 5.5The model's high running cost remains a notable constraint.SentimentSentimentPos87.5%Neg12.5%Many users praise Opus 4.8 for its stop-and-ask behavior in coding because it reduces wasted tokens, loops, and fixes compared to 4.7, while a few note higher costs or weaker performance outside code.16 comments with sentiment. View comments.