88
likes
5
comments
daily view 0
monthly view 0

Share

Live Analytics

Comments 0

Why AI Stopped Lying — TurboPuffer CTO Nikhil Benesch Explains l Dev in the Details Analytics Table

Income Estimates for Why AI Stopped Lying — TurboPuffer CTO Nikhil Benesch Explains l Dev in the Details

Based on this YouTube video's total view count of 79.2K views and industry-standard rates, the estimated total earning is $55 - $158 through ad revenue. Historical data is not yet available to calculate daily, weekly, or monthly averages.

About Why AI Stopped Lying — TurboPuffer CTO Nikhil Benesch Explains l Dev in the Details

Explore Why AI Stopped Lying — TurboPuffer CTO Nikhil Benesch Explains l Dev in the Details with 79,217 views, 88 likes, and 5 comments. Experience the impact of this video content that has captured audience attention.

The Vector Database Powering Notion & Cursor's AI Search | TurboPuffer CTO AI chatbots used to hallucinate constantly. Now they barely do — and it's not because the models got smarter, it's because of search. Nikhil Benesch, CTO of turbopuffer (the vector database powering search for Notion, Cursor, Legora, and Anthropic), joins Dev Rishi to break down why grounding models in real-time search killed hallucinations, and why the next wave of AI infrastructure runs on object storage, not GPUs. Nikhil's path is wild: he dropped out of Harvard for a year to join Whoop, cut his teeth at CockroachDB, spent 6+ years at Materialize (eventually as CTO), and is now CTO of turbopuffer — one of the fastest-growing names in AI infra. In this episode: Why vector search needs a completely different database than Postgres The S3-based insight that makes turbopuffer 10x cheaper than alternatives Why AI stopped hallucinating (hint: it's search, not model size) What a "search agent" is — and why it's the next big AI infra trend Why 0.5% better search relevance was worth $100M to Walmart The line that changes how founders build: "first-time founders focus on product, second-time founders focus on distribution" BYOC (bring-your-own-cloud) and why it's winning over pure SaaS CHAPTERS 0:00 – Intro: Meet Nikhil Benesch, CTO of turbopuffer 1:27 – How he picks which startups to join 6:39 – Why systems programming won him over 7:12 – The Dropbox bug that sold him on CockroachDB 11:24 – Why database revenue is so sticky 12:52 – OLTP vs. OLAP, explained 15:31 – Why he left for Materialize 17:40 – "Napkin math": how turbopuffer plans for scale 18:34 – The one-way door vs. two-way door framework 20:27 – The real story behind the name "turbopuffer" 22:33 – What turbopuffer actually does 23:31 – Vector embeddings, explained simply 25:45 – Why vector search needs specialized infrastructure 26:02 – The scale threshold: when you actually need turbopuffer 27:54 – The big idea: building a vector database on S3 30:08 – turbopuffer's "frenemy" relationship with AWS 31:53 – "Late interaction" search and the next leap in relevance 34:05 – The Walmart stat: 0.5% better search = $100M 34:52 – Finding product-market fit: Simon's cabin & the $30K bill 38:48 – Why RAG lit the fuse on vector search adoption 43:51 – Why AI stopped hallucinating 45:04 – What is a "search agent"? 54:34 – First-time founders chase product, second-time founders chase distribution 55:49 – Why swapping databases terrifies companies (Wiz & Snowflake) 1:00:45 – BYOC: bring your own cloud, explained 1:03:07 – Nikhil's non-consensus take: AI means more software jobs, not fewer 1:09:20 – Closing thoughts GUEST Nikhil Benesch, CTO @ turbopuffer X: https://x.com/nikhilbenesch LinkedIn: https://www.linkedin.com/in/nikhilbenesch/ turbopuffer: https://turbopuffer.com HOST Dev Rishi, GM of AI at Rubrik, former CEO of Predibase Subscribe to Dev in the Details for more conversations at the edge of AI infrastructure. #vectordatabase #aiinfrastructure #TurboPuffer #vectorsearch #RAG #aiagents #startupadvice #techpodcast #softwareengineering #machinelearning #databasesecurity #searchengine #Devinthedetails #buildinpublic #ai

About YouTube Real-Time View Count

With SocialCounts.org’s view counter, track your YouTube video’s live view count and YouTube likes count in real time with fast, reliable updates.

Watch every YouTube video live view count rise with our real-time YouTube views tracker—built for accuracy and minimal delay.

Follow YouTube real time views as they happen, using our dedicated view counter for YouTube videos.

Get up-to-date live view count on YouTube and see real-time growth with SocialCounts.org’s smart tracking tools.

Embed Widget

Parameters:

  • fullscreen=true - Fullscreen counter
  • graph=true - Live graph chart
  • counter=0/1/2 - Select counter (0=likes, 1=views, 2=comments)
URL

Click to copy the embed URL to your clipboard