diff --git a/README.md b/README.md
index adb450f..4cdad48 100644
--- a/README.md
+++ b/README.md
@@ -28,10 +28,11 @@ At this point your bot should run daily and publish a static website. You can te
 
 **Optional but highly recommended**: 
 
-6. [Set up a slack bot](https://api.slack.com/start/quickstart), get the OAuth key, set it to `SLACK_KEY` as a github secret
-7. Make a channel for the bot (and invite it to the channel), get its [Slack channel id](https://stackoverflow.com/questions/40940327/what-is-the-simplest-way-to-find-a-slack-team-id-and-a-channel-id), set it as `SLACK_CHANNEL_ID` in a github secret.
-8. Take a look at `configs/config.ini` to tweak how things are filtered.
-9. Set the github repo private to avoid github actions being [set to inactive after 60 days](https://docs.github.com/en/actions/using-workflows/disabling-and-enabling-a-workflow)
+7. Get and set up a semantic scholar API key (`S2_KEY`) as a github secret. Otherwise the author search step will be very slow
+8. [Set up a slack bot](https://api.slack.com/start/quickstart), get the OAuth key, set it to `SLACK_KEY` as a github secret
+9. Make a channel for the bot (and invite it to the channel), get its [Slack channel id](https://stackoverflow.com/questions/40940327/what-is-the-simplest-way-to-find-a-slack-team-id-and-a-channel-id), set it as `SLACK_CHANNEL_ID` in a github secret.
+10. Take a look at `configs/config.ini` to tweak how things are filtered.
+11. Set the github repo private to avoid github actions being [set to inactive after 60 days](https://docs.github.com/en/actions/using-workflows/disabling-and-enabling-a-workflow)
 
 Each day at 1pm UTC, the bot will run and post to slack and publish a github pages website (see the publish_md and cron_runs actions for details).
 
diff --git a/main.py b/main.py
index 45012df..849fd62 100644
--- a/main.py
+++ b/main.py
@@ -7,6 +7,8 @@
 from requests import Session
 from typing import TypeVar, Generator
 import io
+
+from retry import retry
 from tqdm import tqdm
 
 from arxiv_scraper import get_papers_from_arxiv_rss_api
@@ -94,6 +96,7 @@ def get_author_batch(
         return response.json()
 
 
+@retry(tries=3, delay=2.0)
 def get_one_author(session, author: str, S2_API_KEY: str) -> str:
     # query the right endpoint https://api.semanticscholar.org/graph/v1/author/search?query=adam+smith
     params = {"query": author, "fields": "authorId,name,hIndex", "limit": "10"}
@@ -142,8 +145,12 @@ def get_authors(
             auth_map = get_one_author(session, author, S2_API_KEY)
             if auth_map is not None:
                 author_metadata_dict[author] = auth_map
-            # add a 10ms wait time to avoid rate limiting
-            time.sleep(0.01)
+            # add a 20ms wait time to avoid rate limiting
+            # otherwise, semantic scholar aggressively rate limits, so do 1s
+            if S2_API_KEY is not None:
+                time.sleep(0.02)
+            else:
+                time.sleep(1.0)
     return author_metadata_dict