Its still there - storage is still mainly HDFS, but computation layer morphed into Spark on Yarn (Mesos as a scheduler is barely used). Barely anyone runs classic Mapreduce jobs, Spark all around, written in Scala(or java or python).
The managed data science services from AWS/Google is Hadoop/Spark.
True. Some one in the cloud is managing these Hadoop stacks for you. You just pay usage of these services.