SQL Performance issue

testdata

-- Create table
CREATE TABLE book
(
  id SERIAL NOT NULL,
  released timestamp without time zone NOT NULL,
  author_id character varying(64) NOT NULL,
  CONSTRAINT book_pkey PRIMARY KEY (id),
  CONSTRAINT book_uq1 UNIQUE (author_id, released)
)
WITH (
  OIDS=FALSE
);

-- Create index for author_id and released columns
CREATE INDEX book_idx1
	ON book
	USING btree
	(author_id, released);

-- Add test data (50 authors, each have 100000 Books). Takes ~3 minutes to add.
do $$
begin
for r in 1..50 loop
	for j in 1..100000 loop
		insert into public.book(released, author_id) values(to_timestamp(extract(epoch from now()) + j), r);
	end loop;
end loop;
end;
$$;

-- I need to search for the latest (created column) Book at certain time, for all or some of the author_ids. One Book per author.

-- This one using max() and group by is too slow, around 300ms (but I can search for multiple authors). 
EXPLAIN ANALYZE
SELECT max(released), author_id
     FROM book
	 WHERE author_id in ('1', '2', '3', '4', '5')
	 AND released<=to_timestamp(2e9)
	 AND released>=to_timestamp(0)
	 GROUP BY author_id

-- This query works fast really fast, around 0.2ms (but I cannot search latest books for multiple authors with same query).
EXPLAIN ANALYZE
SELECT released, author_id
     FROM book
     	 WHERE author_id = '1'
	 AND released<=to_timestamp(2e9)
	 AND released>=to_timestamp(0)
	 ORDER BY released desc
         LIMIT 1