Mastodawn

https://blog.gslin.org/archives/2026/04/04/12973/bonsai%ef%bc%8c%e5%ae%a3%e7%a8%b1-1-bit-%e7%9a%84-llm-model/

Bonsai，宣稱 1-bit 的 LLM model

#1Bit #bonsai #language #large #llm #model

Bonsai，宣稱 1-bit 的 LLM model

看到的是「Show HN: 1-Bit Bonsai, the First Commercially Viable 1-Bit LLMs (prismml.com)」這則，官網是 https://prismml.com/。其中 1-bit Bonsai 8B 這個 model 的介紹是這樣： The first commercially viable model with 1-bit weights. Requiring only 1.

Gea-Suan Lin's BLOG

Gea-Suan Lin 22h ago

https://blog.gslin.org/archives/2026/04/04/12972/macos-tohoe-26-%e5%85%a7%e5%bb%ba%e7%9a%84-llm/

macOS Tohoe (26) 內建的 LLM

#26 #3b+ #apple #foundation #language #large #llm #macos #model #silicon #tohoe

macOS Tohoe (26) 內建的 LLM

看到「Arthur-Ficial/apfel (via)」這個專案，macOS Tohoe 內都有附一個 model，但內建的工具中只有 Siri 會用到他，所以作者就跟 Claude Code 一起刻了一個介面出來： Every Mac with Apple Silicon has a built-in LLM - Apple's on-device foundation model, shipped as...

Gea-Suan Lin's BLOG

Gayniples😋1d ago

#large #bigtech #penis #gay

Hacker News 2d ago

Trinity Large Thinking

https://openrouter.ai/arcee-ai/trinity-large-thinking

#HackerNews #Trinity #Large #Thinking #AI #Innovation #OpenRouter #Technology

Trinity Large Thinking - API Pricing & Providers

Trinity Large Thinking is a powerful open source reasoning model from the team at Arcee AI. $0.25 per million input tokens, $0.90 per million output tokens. 262,144 token context window, maximum output of 80,000 tokens.

Gea-Suan Lin 3d ago

https://blog.gslin.org/archives/2026/04/01/12967/%e8%a8%98%e5%be%97%e5%8e%bb%e9%97%9c-github-copilot-%e8%a3%a1%e9%9d%a2%e7%9a%84-privacy-%e8%a8%ad%e5%ae%9a/

記得去關 GitHub Copilot 裡面的 Privacy 設定

#ai #copilot #cory #doctorow #enshittification #github #language #large #llm #model #privacy #training

記得去關 GitHub Copilot 裡面的 Privacy 設定

從 4/24 開始，GitHub 將會存取你的 private data 當作 AI 訓練用：「Updates to GitHub Copilot interaction data usage policy」。 private 的重點在這： This program does not use: * Content from your issues, discussions, or private repositories at rest.

Gea-Suan Lin's BLOG

Radio_Azureus 6d ago

LLM insider view

Insightful video. Regardless of your stand on LLMs you will learn a lot from analyzing this vid.
The truth about LLMs

https://www.youtube.com/watch?v=Cn8HBj8QAbk

#LLM #AI #slop #enshittification #programming #large #language #model #technology #hidden #whistleblower #insightful

CBC Top Stories Mar 28

Scientists captured rare video of a sperm whale birth. What they learned from that remarkable moment
Rare footage of a sperm whale giving birth has offered scientists a window into the behaviour of these large, elusive mammals.
https://www.cbc.ca/news/science/sperm-whale-birth-video-9.7144808?cmp=rss

CBC Technology Science Mar 28

Я⊛ḶḶ model ☭Mar 28

Trump “is getting a little bored with Iran,” a senior White House official told Jake Traylor of MS NOW. “Not that he regrets it or something—he’s just bored and wants to move on.”

https://open.substack.com/pub/heathercoxrichardson/p/march-27-2026

#buffoonary #writ #large

March 27, 2026

The ongoing battle over funding Transportation Security Administration (TSA) agents at U.S.

Letters from an American

Gea-Suan Lin Mar 28

https://blog.gslin.org/archives/2026/03/28/12959/intel-%e8%a6%81%e6%8e%a8%e5%87%ba%e5%a4%a7-vram-%e7%89%88%e6%9c%ac%e7%9a%84%e9%a1%af%e5%8d%a1-arc-pro-b65-%e8%88%87-arc-pro-b70/

Intel 要推出大 VRAM 版本的顯卡 Arc Pro B65 與 Arc Pro B70

#arc #b65 #b70 #card #gpu #intel #language #large #llm #memory #model #pro #size #video #vram #xe2

Intel 要推出大 VRAM 版本的顯卡 Arc Pro B65 與 Arc Pro B70

Intel 推出大 VRAM 版本的顯卡：「Intel Announces Arc Pro B70 and Arc Pro B65 GPUs, Maxes Out Xe2 Battlemage Architecture (via)」。 367 TOPS (INT8) 與記憶體的速度只有 608GB/sec，這兩個差不多是隔壁 Nvidia 入門級顯卡的規格，但賣點是 32GB VRAM 與目前傳言...

Gea-Suan Lin's BLOG