in experimental sciences, you don't need confidence intervals if you replicate your experiments on several datasets and conditions.
http://andrewgelman.com/2017/06/07/no-conf-intervals-no-problem-got-replication/