What does Kafka give up to handle F failures with F+1 nodes while still having strong consistency while Raft needs 2F+1 nodes?

I am reading Kafka design of replication and fault-tolerant,
and I see:

What Kafka gives up to handle F failures with F+1 nodes while still having strong consistency while Raft needs 2F+1 node?

I am reading Kafka design of replication and fault-tolerant,
and I see: