# When a user posts a tweet:
# 1. Look up all followers (could be millions)
# 2. Insert the tweet into each follower's feed cache
# 3. Reading feed = just read from pre-computed cache
# Pro: Fast reads (O(1) cache lookup)
# Con: Celebrity with 10M followers = 10M writes per tweet

# When a user opens their feed:
# 1. Look up all accounts they follow
# 2. Fetch recent tweets from each (500 queries)
# 3. Merge-sort by timestamp, return top N
# Pro: No write amplification
# Con: Slow reads — 500 queries at read time

# For normal users (< 10K followers): fan-out on write
#   → tweet is pushed to all follower feed caches
# For celebrities (> 10K followers): fan-out on read
#   → their tweets are fetched at read time and merged
# Reading: merge pre-computed cache + celebrity tweets
# Best of both worlds

# Client periodically polls each followed user's tweet endpoint
# Client merges and sorts locally
# Pro: Simple server
# Con: Massive client bandwidth, battery drain, stale data