Episode Link
https://share.snipd.com/episode/e915ab63-dc66-454d-b9b6-af4a4739e72c
Episode publish date
August 19, 2025 8:54 PM (UTC)
Last edit date
Aug 21, 2025 1:38 PM
Last snip date
August 21, 2025 2:37 PM (GMT+1)
Last sync date
August 21, 2025 2:38 PM (GMT+1)
Show
Latent Space: The AI Engineer Podcast
Show notes link
https://assets.flightcast.com/track-v2/01K31AMDS3GJWXK9Y9MWE7ER3V.mp3
Snips
13
Warning
⚠️ Any content within the episode information, snip blocks might be updated or overwritten by Snipd in a future sync. Add your edits or additional notes outside these blocks to keep them safe.
‣
Your snips
‣
[05:00] Choose Focus Over Fast Market Hype
‣
[07:10] Hire Slowly To Protect Culture
‣
[09:13] Chroma's Zero-Config Onboarding Story
‣
[12:23] Context Engineering Defined
‣
[13:59] Context Rot Makes Engineering Essential
‣
[20:15] Community Research Fills Labs' Developer Gap
‣
[24:01] Use Two-Stage Retrieval Then LLM Re-Rank
‣
[27:50] Indexing Is A Write/Query Trade-Off
‣
[32:41] Precompute & Benchmark During Ingestion
‣
[35:20] Retrieval Might Stay In Latent Space
‣
[41:38] Use Offline Compaction For Memory Quality
‣
[42:36] Generate QA Pairs To Benchmark Retrieval
‣