We benchmarked Lakebase Search on 1M to 4M passages to understand how vector search performs as the index grows. Here's what we learned 👇 Cache size relative to the vector index mattered more than row count. When the cache roughly matched the index, 95–98% of page reads avoided storage and warm vector search p50 latency stayed between 11 and 23 ms. As datasets grow, sizing the cache alongside the vector index helps maintain low-latency search.

The Louisiana Department of Health is warning residents about the risks of using tap water to rinse nasal passages and clean contact lenses after several rare amoeba infections.

The more I learn about the Vegas Loop, the more I think the tunnels are actually the least interesting part of it. The Boring Company has already built 14 stations, 12 emergency egresses, six cross-passages and nine knockout shafts around the system. By the end of the year, they expect 28 stations. And some of the engineering happens in remarkably small increments. Connecting two existing tunnels takes about ten days. Stabilize the earth, cut through one tunnel wall, excavate roughly ten feet a day toward the other, then waterproof and reinforce the new passage into both structures. It made me realize that building an underground transportation network isn’t really one giant tunneling problem. It’s hundreds of small civil-engineering problems that eventually become a city beneath the city. Writer: Oliver

On BuildBetter's own August benchmark, across one workspace with 6,018 calls and 8,533 support conversations, AOT returned 99% of the verified customer evidence, compared with about 11% for query-time search at a hundred passages per question. Book a demo 👀