Skip to content

DEV Community

Jax Sinclair

Semiconductor analyst. I write about AI silicon bottlenecks: advanced packaging, HBM, test, and the yield math underneath. Numbers first.

Joined on Oct 29, 2026 https://jax-silicon-notes.surge.sh

Work

Semiconductor analyst

Nov 2

A 671B model reads 36GB per token and still needs 671GB of HBM

#llm #hardware #performance #gpu

7 min read

Want to connect with Jax Sinclair?

Create an account to connect with Jax Sinclair. You can also sign in below to proceed if you already have an account.

Already have an account? Sign in

Nov 1

Halving the KV Cache Doubles Capacity and Buys 1% of Speed

#ai #llm #performance #machinelearning

6 min read

Oct 31

Decode Is 300x Memory-Bound. Batching Fixes That Until the Context Gets Long.

#ai #llm #performance #hardware

6 min read

Oct 30

The KV Cache Is Your Inference Bill: Bytes Per Token, From Three Real Configs

#ai #llm #performance #machinelearning

4 min read

Oct 30

AI Packages Are Bending Under Their Own Size. The Money Moved to Prediction.

#ai #hardware #architecture #performance

3 min read

Oct 30

Advanced Packaging's Yield Killer Isn't Assembly. It's Test.

#discuss #ai #hardware #performance

3 min read

loading...