前几天和 Cursor 里的 AI 聊了聊思维是否都以语言形式存在、「语言即世界」到底在说什么、以及维特根斯坦和萨丕尔-沃尔夫假说。下面是完整对话，原样保留。 我在想一个问题。人类在想问题的时候，或者仅仅只是日常的时候，脑袋里所有的想法，都是以语言的形式在存在。例如，他饿了，他会想，「饿了」，这个以语言或文字为载体来的，中国人会想，饿了，美国人会想，hungry，这些都是不同语言但都是文字为载体的想法。但语言的表达是有限的，存在语言无法准确描述一个事物或感受的情况。例子1，有一个概念，一个人懂了，但他从未想与别人解释这个东西，他就从来没想过或组织过语言来看这个东西要怎么说，因此这个东西在他大脑里是存在的，但并没有对应的语言。只有他想要表达的时候，他才会把这个东西组织成语言。因此这个状态下，概念在，但没有对应的语言。例子2，假设有一个人，生下来之后就没有生活在人群中，因此没有任何人教他语言教他说话，因此他不会任何语言，那他大脑里的想法是以什么形式来存在呢？例子3，动物，不会任何的语言，例如一只猫或狗，他们也有大脑，也应该有想法，但不会语言，那在他们的脑海里，念头或意识以什么形式存在？我们太习惯我们脑海里的想法都是以文字形式存在了以至于都没有意识到这些 Cursor 这是一个非常深刻的问题，涉及认知科学、语言学和哲学的交叉领域。你的直觉是对的——思维并不完全依赖于语言，语言只是思维的一种「编码方式」，而非思维本身。

前几天和 Cursor 里的 AI 聊了聊「抽象」和「无厘头」、为什么现在的搞笑越来越难解释、以及它和传统「情理之中意料之外」的段子有什么不同。下面是完整对话，原样保留。

Nowadays, LLMs have become so impressive that I often have mixed feelings about their capabilities. I’ve always wanted to showcase my conversations with LLMs on my blog. However, the Hugo theme I use to generate my blog’s static site did not have a built-in way to display conversations like those on ChatGPT or other LLM websites. It was actually one of my conversations with ChatGPT that inspired me to develop this feature. I used Cursor to modify the Hugo theme’s template files and add a custom chat component. Below is that conversation.

Apple has released mps backend ^1^ which boosts Macs that have AMD GPU or M series processor runs an LLM locally. This tutorial gives step-by-step instructions to run the ChatGLM2-6B model on a 16-inch MacBook Pro (2019) with 32G memory and AMD Radeon Pro 5500M 4 GB GPU. Build the enviroment Install openMP curl -O https://mac.r-project.org/openmp/openmp-12.0.1-darwin20-Release.tar.gz sudo tar fvxz openmp-12.0.1-darwin20-Release.tar.gz -C / The contained set of files is the same in all tar balls:

Abstract The objective of this project is to establish a functional whole-house networking system within a multiple-bedroom apartment, which presents two primary challenges. The first challenge concerns the well-documented issue of internet firewalls, which requires additional measures to enable access to fundamental web services like Google, YouTube, Netflix, and OpenAI. The second challenge relates to the diminishing Wi-Fi signal as it penetrates through the walls of the bedrooms, rendering a single router insufficient for the entire house. The aim of this project is to address these issues comprehensively and provide a comprehensive solution.

Install Nvidia driver on Ubuntu can be very tricky. Sometimes it goes extremely smooth while sometime it just doesn’t work. Follow these following steps, it should make it easier. Download Nvidia driver to your Ubuntu machine from its website. It is a .run file. Install make and gcc. sudo apt install make sudo apt install gcc Disable Nouveau kernel driver. # Create a file sudo nano /etc/modprobe.d/blacklist-nouveau.conf with the following contents:

Why virtual environment is needed? It is very common in deep learning that each project has different dependency requirements, which makes environment management important. Conda can help us set multiple virtual environments and keep them independent from each other. Thus, all of your ML projects can run smoothly. Set up virtual environment for pytorch using conda Install CUDA and cuDNN This step is only necessary for those who has a GPU. Before downloading CUDA and cuDNN from NVIDIA, check the version requirements on PyTorch website.

Everyone is an island, and so do the blogs. It makes us cherish things we know and people we love. Here are some blogs I found very interesting. This post will update continuously. https://charlesliuyx.github.io This blog was writen in Chinese, topics including machine learning, blockchain, Chinese ancient poems, etc. https://yiqinfu.github.io A PhD student’s blog which contains a lot of very deep and interesting thoughts. Born and raised in China, Yiqin (pronounced ee-ching) Fu spent many of her formative years in the U.S. and the U.K. She is studying towards a Ph.D. in political science at Stanford University, after having worked as a research associate at Yale Law School’s Paul Tsai China Center in New Haven, Connecticut and Beijing, China. Yiqin holds a B.A. in Philosophy, Politics, and Economics from the University of Oxford and is broadly interested in innovation, U.S.-China relations, and comparative political and electoral systems.

Why virtual environment is needed? It is very common in deep learning that each project has different dependency requirements, which makes environment management important. Conda can help us set multiple virtual environments and keep them independent from each other. Thus, all of your ML projects can run smoothly. Set up virtual environment for pytorch using conda Install Anaconda or Miniconda depends on your taste. Go to anaconda or miniconda’s website and download the .sh file for your system. Go into the terminal use “chmod +x filename” command to the .sh file then run it. Always click yes to finish installation.

What’s the goal? To run YOLOv5 on jetson developer kit to achieve real-time object detection. The following instructions has been tested on: Jetson nano Jetson Xavier NX What you need: A Jetson nano (If you are using other Jetson developer kit from Nvidia, I am sure you can take this as reference as well.) A SD card and power suply which meets the needs. Another computer or laptop with internet connection and built-in SD card slot or an adapter. A monitor with HDMI port, a USB mouse and a USB keyboard. Firsty, to start the system on Jetson developer kit, write image to microSD card.