> I don’t think hallucinations are a major issue with agentic coding a year later? At least, not for me.

Are you still looking at code and agent traces to know if this is true or not?

At work, we still see hallucinations across all model families as we share bot claims for each others code