CLI Command to mount an SMB share on Apple Silicon that isn't abysmally slow

I've finally got tired of how bad the latency and transfer speeds are when mounting my TrueNas SMB shares on my macbook. I looked online for some s…

Pedro Cuenca (@pcuenq)

Transformers에 추가되는 모든 모델이 즉시 Apple Silicon에서 사용 가능해야 한다는 목표를 위해, mlx-lm용 Skill과 테스트 하네스를 구축했다는 내용입니다. 기여자와 리뷰어 모두를 돕기 위한 개발 인프라 개선으로, 모델 지원과 검증 과정을 표준화하는 데 의미가 큽니다.

https://x.com/pcuenq/status/2044793328978497590

#transformers #mlxlm #applesilicon #opensource #modelsupport

Pedro Cuenca (@pcuenq) on X

🔈 Every model added to transformers has to be available on Apple Silicon 🍎 at once. We built a Skill and test harness for mlx-lm to get us closer 🔥 It's designed to help contributors AND support reviewers. Read on to see what we did and why it matters.

X (formerly Twitter)

Awni Hannun (@awnihannun)

Transformers에서 새로운 모델을 mlx-lm으로 자동 포팅하기 위한 새로운 스킬과 테스트 하네스를 만드는 블로그 글이 소개되었습니다. 모델 이식 자동화와 검증을 개선해, Apple Silicon에서의 모델 지원을 더 빠르고 안정적으로 확장하려는 개발 도구 업데이트입니다.

https://x.com/awnihannun/status/2044804263621312993

#transformers #mlxlm #applesilicon #tooling #automation

Awni Hannun (@awnihannun) on X

Great blogpost from @pcuenq on making a new skill + test harness to automate porting new models from Transformers to mlx-lm

X (formerly Twitter)

N8 Programs (@N8Programs)

Qwen3.6 4bit DWQ가 MLX에 추가되었습니다. 4bit MLP와 8bit 나머지 레이어를 쓰는 커스텀 양자화 방식과 DWQ를 결합해, 기본 모델 대비 KL 손실은 0.0225로 낮고 PPL도 거의 동일하며, 단순 4bit 양자화보다 성능이 크게 개선됐습니다. 비트당 비용도 0.25BPW만 추가됩니다.

https://x.com/N8Programs/status/2044951328560967888

#qwen #mlx #quantization #llm #applesilicon

N8 Programs (@N8Programs) on X

Qwen3.6 4bit DWQ now up on MLX, uses custom quantization scheme (4bit MLP 8bit everything else) + DWQ for additional gains. It gets 0.0225 KL w/ the base model, and matches it on PPL - versus 0.0819 for a naive 4bit quant. Adds only 0.25BPW! https://t.co/CU8UzX27lU

X (formerly Twitter)
M5 Pro : l’insolente supériorité d’Apple Silicon face au futur d’Intel http://dlvr.it/TS4n3h #AppleSilicon #M5Pro
M5 Pro : l’insolente supériorité d’Apple Silicon face au futur d’Intel

Souvent imitée, jamais égalée. Alors que l’architecture Apple Silicon compte déjà quelques années d’exploitation, elle continue de faire la course en tête face aux architectures concurrentes. Si la formule de Cupertino demeure dans ses grandes lignes...

MacGeneration

The best emulation device for now may be, surprisingly, the #MacBookNeo

https://www.youtube.com/watch?v=15Ma2hnQIpg

Notice, it's be thest cheap laptop you can buy, it runs #macOS and there is also capable of light gaming.

#Videogames #Gaming #Games #Emulation #MacBook #Apple #AppleSilicon #Laptop #Laptops

Emulation on the MacBook Neo

YouTube
@speisekarte
Ja, #applesilicon bzw. der Umstieg von #intel auf #arm im #mac war mit Abstand das Beste, was sie machen konnte. Auch mir tut es weh, dass ich keinen geilen arm based Laptop habe. Aber so ist das halt. Der alte Intel hat 11 Jahre gehalten und 5 davon seinen Dienst mit #macos getan.
Auch, dass sie gute Qualität liefern, muss man #apple lassen.

Python Trending (@pythontrending)

Apple Silicon에서 동작하는 LLM 추론 서버 omlx가 공개되었습니다. 연속 배칭(continuous batching)과 SSD 캐싱을 지원하며, macOS 메뉴바에서 관리할 수 있어 로컬 AI 추론 운영을 더 편리하게 만듭니다.

https://x.com/pythontrending/status/2044728935124959549

#llm #inference #applesilicon #macos #aiinfrastructure

Python Trending 🇺🇦 (@pythontrending) on X

omlx - LLM inference server with continuous batching & SSD caching for Apple Silicon — managed from the macOS menu bar https://t.co/oqacZn91Rt

X (formerly Twitter)
DFU mode

How Intel Macs with T2 chips, and Apple silicon Macs can be rescued from the apparently dead by connecting them to another Mac in DFU mode.

The Eclectic Light Company