Why is disabling thinking for coding models a good idea?
I've seen several people recommend disabling thinking for models when used in agent encoding, but I haven't been able to find any reasoning behind it. Could you please share details on this topic? submitted by /u/ThingRexCom […