The industry's defence against data poisoning was arithmetic: an attacker needs a percentage of the corpus, and a percentage of 260 billion tokens is unobtainable. In October 2025 the largest poisoning study ever run showed the r…
Model registries were built to version a trained artefact you own. An LLM application's behaviour comes from a hosted snapshot that retires on someone else's calendar, plus prompts, an index and tools you change weekly. The relea…
An LLM agent with a million-token window still forgets what you told it last week. Memory is not a bigger context; it is a separate system that decides what to write down, how to summarize it, and what to pull back. Here is how t…
In 2021 a benchmark of 18 retrieval datasets found that a scoring function from 1994 still beat most neural retrievers outside their training domain. BM25 is not a heuristic that got lucky. It is a deliberate approximation of a p…
A million-token window promises perfect recall of everything you feed it. Controlled tests on 18 frontier models show recall degrading steadily, unevenly, and well before the window fills, a pattern researchers now call context rot.
An LLM-written pseudo-document lifts BM25 by 15 nDCG@10 points on TREC DL 2019, and turns a 16 ms search into one that waits over two seconds for the model. Query understanding decides what the engine searches for before ranking …
Teams tune the embedding model and the vector database, then wonder why answers are still wrong. Both are the least important parts of the stack. Retrieval is a two-stage ranking problem, and information retrieval solved the shap…
Most teams reach for fine-tuning when they should reach for retrieval. The cost is not just dollars - it is months of confused engineers trying to figure out why the model started lying again.
In 2019 a masked language model recalled facts almost as well as a relation extractor with an oracle entity linker, and people asked whether knowledge graphs were finished. They were not, but the reason is narrower than their adv…