Hacker News Reader

Modukeon Qwen3.8 27B scores 52 on Artificial Analysis

Very exciting indeed. It is in the works. Their current dense offering, Gemma 4 31B, sits at ~1800t/s

https://news.ycombinator.com/item?id=49308715

Modukeon Qwen3.8 27B Coming to Cerebras

https://xcancel.com/cerebras/status/2088419572416147582

Modukeon Moonshot AI suspends new subscriptions due to Kimi K3 demand

> we're temporarily pausing new subscriptions and prioritizing compute for current members

emphasis on _new_ subscriptions but otherwise yes that’s exactly it.

Modukeon Tinybox – Offline AI device 120B parameters

https://xcancel.com/jgarzik/status/2031312666036146460?s=20

Modukeon Kalman Filter Explained Simply

I enjoyed the simplicity of this explanation as well:

https://praveshkoirala.com/2023/06/13/a-non-mathematical-int...

https://news.ycombinator.com/item?id=36971975

Modukeon Apple unveils new iPad mini with breakthrough performance in new design

Do you have a source for the claim that they use the A15 in the iPad mini? It stood out to me that they left the details of the SoC out of the announcement.