Vector search has change into a foundational know-how for AI purposes, enabling all the pieces from semantic code search to contextual retrieval for big language fashions. Nevertheless, a significant problem with vector databases has been the fee as knowledge storage scales.
Turbopuffer is a vector database that focuses on velocity, value and scalability. It was created by Simon Hørup Eskildsen and Justin Li in 2023 and has seen adoption from high-profile firms corresponding to Cursor and Notion.
Simon joins the podcast with Gregor Vand to debate the origin of turbopuffer, its distinctive technical design, the economics of vector storage, and extra.
Gregor Vand is a security-focused technologist, and is the founder and CTO of Mailpass. Beforehand, Gregor was a CTO throughout cybersecurity, cyber insurance coverage and basic software program engineering firms. He has been based mostly in Asia Pacific for nearly a decade and might be discovered by way of his profile at vand.hk.Capital One’s tech group isn’t simply speaking about multi-agentic AI.
They already deployed one.
It’s known as Chat Concierge, and it’s simplifying automobile buying.
Utilizing self-reflection and layered reasoning with dwell API checks,
It doesn’t simply assist patrons discover a automobile they love.
It helps schedule a check drive, get pre-approved for financing, and estimate trade-in worth.
Superior, intuitive, and deployed: that’s how they stack.
That’s know-how at Capital One.
Constructing agentic AI apps isn’t nearly selecting one of the best LLM.
Brokers want brief‑time period reminiscence, lengthy‑time period recall, and lightning‑quick retrieval. With out it, you’re left with clunky prototypes that by no means scale.
You understand, Redis? The world’s quickest caching answer?
It seems quick knowledge is the important thing to good context. And good context is important for quick, correct reminiscence. It’s what makes AI brokers really work together with your knowledge.
Redis for AI. The correct infrastructure. The correct instruments. The one option to scale.
Be taught extra at redis.io/genai
Constructing an app typically appears like a balancing act. You wish to ship options quick – chat, exercise feeds, moderation, video. However constructing them from scratch is sluggish and complicated.
That’s the place Stream is available in. Stream offers developer-friendly APIs that allow you to add real-time communication with out reinventing the wheel.
Why Stream?
First, developer expertise. Stream has clear, open supply SDKs, and nice docs, and you will get a proof of idea working in hours.
Second, velocity to construct. Experiment, prototype, or launch options rapidly – no bank card required to start out.
Stream additionally scales. Over 2,000 international apps, together with Strava, Patreon, Nextdoor, Robinhood, and Peloton, depend on Stream to energy in-app communication for greater than a billion customers.
Whether or not you’re a startup or an enterprise, Stream handles the laborious elements so you possibly can deal with what makes your app distinctive. Get began right this moment at getstream.io/podcast.