Loading TechWhoKnows…
Reporting, analysis and primary-source coverage on Tencent.

Models & Research
Tencent's new Hy4 flash model uses hybrid attention, LatentMoE and Engram architecture to deliver 150+ tokens per second at a fraction of predecessor costs, now free on WorkBuddy through November 30.
TechWhoKnows

Agents & Tools
Compare vLLM, SGLang, and TensorRT-LLM in 2026 across PagedAttention, RadixAttention prefix caching, TensorRT compilation, and inference server security.
Jin

Asia & SEA Tech
Why independent home tutors in Singapore still juggle WhatsApp chats, PayNow screenshots, and spreadsheets for scheduling and invoicing, and how dedicated…
Jin