Jo Kristian Bergum
@jobergum
Search is the natural abstraction for augmenting AI with evolving reality
Search is the natural abstraction for augmenting AI with moving context The initial excitement around RAG centered on vector databases. Since then, we've seen various variations—graphRAG or lazyRAG—each offering a spin tied to a specific technology or vendor. A slight…

Appreciate the self awareness “user friendly” consistently in quotes

Nice cookbooks from the friendly people at @QuotientAI. Tracking relevance 👀
Just dropped: three new cookbooks for building AI research agents with @ExaAILabs, @LangChainAI, @OpenAI, and @AnthropicAI — now with built-in monitoring from @QuotientAI. Track search relevance. Catch hallucinations. Debug real-world agents as they run.
This is extremely well articulated
This is why going against a platform requires you to niche down and build a massive moat in terms of applied expertise in your specific industry so that it is not feasible, possible, or desirable to reach parity from the platform's point of view.
Amazing progress for Omni retrieval, this is really something that all the emerging notetaker apps needs.
Introducing ColQwen-Omni, a 3B omnimodal retriever that extends the ColPali concept of multimodal retrieval with late interaction to audio chunks and short videos, with no performance degradation on visual document retrieval wrt our best models! (1/N)
What’s the accuracy lift by giving coding agents the “Jeff Dean” persona? You are Jeffrey Dean, you are working on a mission-critical project to save humanity by implementing an important CRUD application. IMPORTANT: Mocking unit tests is a fireable offense.
Quality retrieval has never come from general-purpose storage or SQL databases. Decades of search history (web, enterprise, recommendations) show that good retrieval systems are specialized, with purpose-built ranking. Storing a vector doesn’t cut it. The category ‘vector…
Pretty sure the future of vector stores are: AWS S3 vector store and pgvector on Postgres That's it Don't see how dedicated vector DBs can survive unless it serves specific niches
The AI trading cards
BREAKING: Claude Code PMs Boris Cherny and Cat Wu have returned to Anthropic after a brief stint at Cursor.
Maybe just me but I feel empowered by LLMs and coding assistants
Talk to Charlie! Legend.
..from mid-September onwards to supercharge your search & AI do check out my background at thesearchjuggler.com and get in touch, I'm looking forward to my next challenge! [3/3]