Wafer now powers Brilliant's AI tutor Koji, ending speculative prefetching for latency
ycombinator · x · 2026-09-25
Wafer's founder shares that Brilliant's AI tutor Koji needs to respond near-instantly as students work through math and coding problems. Before Wafer, Brilliant had to speculatively prefetch AI-generated responses to hide latency; Wafer's inference stack now serves the tutor. A personal twist: the founder grew up in Mexico relying on Brilliant's logic courses to get into a top US university.
More from Infra
- Deep Inference-Query Engine Integration: Custom Scheduler and Workload-Aware KV Cache for Prefill-Only AI Filters — charles_irl · 2026-09-25
- Diffusion LLM goes production: Augment Code's Mercury 2.5 switch cuts latency 82%, cost 90% — cen6wkf · 2026-09-25
- Oracle says force majeure notice doesn't signal data center delays, rent still due — AIFlow_ML · 2026-09-25
- SemiAnalysis turns more bullish on China's WFE localization after CSEAC 2026 — zephyr_z9 · 2026-09-25
- Meta Muse runs on just 2 cores of an AMD EPYC server CPU — firstadopter · 2026-09-25
- M5 Ultra 80-core tested with GLM-5.3-Flash: RAM is great, GPU is the bottleneck — dreamingwell · 2026-09-25