“Good enough” as in we don’t see any point in 1-shotting everything we want to build in lightning speed. If Opus 5.5 can 1 shot it then that product is essentially commodified, no point in any one building it except as an internal tool.
If Opus can’t 1-shot it, then it must rely on our human intelligence which can be complemented well enough with a dumb model as a frontier model.
If Opus can’t 1-shot it, then it must rely on our human intelligence which can be complemented well enough with a dumb model as a frontier model.