Method Drift
Which methods have been superseded
A living systematic review per builder-problem. For each problem below, we track which methods have become superseded baselines, what sits on the live frontier, and the grounded receipts — verbatim critique quotes and benchmark numbers, each traced to the paper it came from — behind every “X superseded by Y” claim.
10 builder-problems · 3,164 papers · 6,756 critique receipts · 26,480 benchmark results
Parameter-efficient fine-tuning (LoRA family)
Adapting big models without touching every weight — LoRA and the variants chasing it.
Retrieval-augmented generation
Grounding generation in retrieved context — the architectures replacing naive RAG.
LLM reasoning / chain-of-thought
Teaching models to think before answering — chain-of-thought and what came after.
Mixture-of-experts routing
Scaling capacity without the compute — sparse experts and how they are routed.
KV-cache compression
Shrinking the attention KV cache for long context — eviction, quantization, and reuse.
Speculative decoding
Drafting tokens ahead to speed up inference — the EAGLE/Medusa lineage.
LLM quantization
Running large models in low precision — post-training quantization methods.
Agent / long-term memory
What an agent remembers across a session — long-term memory systems.
Long-context / context-window extension
Reading past the training length — positional tricks that extend context.
Tool use / function calling
Teaching models to call APIs and tools — function-calling and agent toolkits.