Federico Viticci

viticci.macstories.net

25 September 2026 09:36

This M5 Ultra Mac Studio feels made for Qwen3.8-Flash-Next. - 100+ TPS at 4 or 5-bit - Token prefill is up 150% (!) compared to M3 Ultra - Can use subagents with room to spare Flash-Next feels GREAT for agentic coding, especially at these speeds. www.macstories.net/stories/m5-u...

Federico Viticci

viticci.macstories.net

24 September 2026 08:30

Apple also sent me an M5 Pro Mac mini, 64GB of RAM. I spent more time with the M5 Ultra, but for the price and size, this thing also rips! Will do more testing on this soon.

Lorenzo 'kelset' Sciandra

kelset.dev

11 September 2026 06:11

the cat is out of the bag: we have rebuilt from the ground up Shop in SwiftUI and Jetpack Compose. it still feels weird to think that it actually worked out lol. needless to say, it was a crazy few months. Can you believe it? full rewrite, and now it's already out in prod 🤯

Antonio Pitasi

anto.pt

16 July 2026 07:41

beach time in viareggio, #streetphotography 📍 Viareggio, Tuscany, IT #GrainSocial https://grain.social/profile/did:plc:ocqnlmhmdmcnmnbqgnnjrcfu/gallery/3mqmns52orq24

Zhihang Xie

zhihangxie.bsky.social

08 July 2026 14:07

🚀 New paper: Speech-XL for long-form SpeechLLMs 📄 arxiv.org/abs/2602.05373 🧩 Uses Speech Summarization Tokens to compress local speech intervals into compact KV states efficiently. ✨ Improves long-form speech understanding while reducing memory and FLOPs on 10-minute audio.