How to Survive the Vortex of LLM Change
summary
Progress Semantic Rack is a production intelligent search system that adapts by running RAG and a small LLM as an internal evaluator that scores every user query; a captioning model mistake (poetic output for an unrelated image) and a hallucinated base64 image reveal why leaderboard wins fail in production. Providers impose dynamic rate limits and silent deprecations, so the engineering stack includes a deprecation calendar, mandatory model migration plans, strict testing and continuous evaluation before rolling new models.
AI-written summary · about
chapters (9)
- 0:00Introduction and team background
- 1:10LLMs and emotional prompting
- 2:57Rapid model and API changes
- 5:06Evaluation beyond leaderboards
- 7:12Research committee and trend filtering
- 9:17Evolving client expectations
- 13:09Key survival strategies
- 15:38Human connection and burnout
- 17:25Q&A session
related talks
22mPractical Strategies for Implementing Domain-Specific LLMs
59mThe Evolution of RAG
31mBeyond LLMs and Maturing RAG use cases
20mReal-time RAG: How to augment LLMs with Redis and Amazon Bedrock
28mNavigating Challenges and Enhancing Performance of LLM based Applications
27mOptimizing RAG Systems in Practice: Challenges and Proven Solutions
49mLLM—Search Engine Killer?
19mLLM codegen fails and how to stop 'em