#recommendation-systems
-
2026-03-12
WTF is the Recovering Difference Softmax Algorithm?Duolingo's notification algorithm isn't just A/B testing with extra steps. It's a bandit that knows when to shut up — and that's the hard part.
-
2026-03-12
Signal Fusion: How Semantic, Relational, and Direct Signals Combine to Make Recommendations That Don't SuckEvery recommendation system that works well is fusing multiple signal types. The ones that don't understand this ship vibes-based retrieval and wonder why users leave. A taxonomy of signals, how they combine, and what the SOTA ecosystem gets right and wrong.
-
2026-03-11
WTF is Two-Tower Recommendation?The architecture behind every recommendation system that actually works at scale — why splitting the model in half is the key to serving billions of candidates in milliseconds.
-
2026-03-10
Through-Line Detection: What LLMs See That Rule Systems Can'tCross-modal pattern detection across behavioral data, free-text, and metadata — the capability that actually justifies using an LLM in a recommendation system.
-
2026-03-10
Signal Stability Classification: Inference Cost-Benefit in Hybrid Recommendation SystemsNot all behavioral signals deserve the same compute budget. Genre affinity changes over weeks; session mood changes in seconds. Classify by stability, infer by tier, and stop pretending daily batch is the answer to everything.
-
2026-03-10
Negative Signals as First-Class Citizens in RecommendationWhat users don't do matters more than what they do. Most recommendation systems are built entirely on applause. Here's why the silence is louder.
-
2026-03-10
Filter Bubble Mitigation in Personalized SystemsHow to avoid turning personalization into an algorithmic echo chamber — and why the fix isn't as simple as randomly throwing garbage at your users.