Blogs

Posts

  • 给 QQ bot 加个功能,顺着一排 migration 抓出两个矿工:我这需求又做偏了

    本来在给傻氪补语音、网页读取和中途回复。看了一眼服务器,顺着一排 migration,最后查出了两条入侵链。机器人还没学会自主汇报,服务器已经用一百多个核汇报了自己的副业。
  • I meant to fix a QQ bot. A wall of migration threads led me to two cryptominers.

    I was adding voice messages, web reading, and progress replies to Krypton. Then I checked the server. The bot was still learning to report progress; the server was already using more than a hundred cores to announce its side hustle.
  • 本来只是想修一次部署:Cortex 的 Cloudflare 事件记录

    本来只想让 Cortex 的本地发布和 GitHub 自动部署都能跑通。后来构建修好了,页面里却多出一段我没写过的东西。顺着它往回查,事情一直到了三天前。
  • I only meant to fix a deployment: the Cortex Cloudflare incident

    I wanted Cortex to deploy from my machine and from GitHub. We fixed the build, then found HTML I had never written. Following it led to another Worker and an audit trail going back three days.
  • New Album: Malkuth

    Info

  • Maillard Reaction

    The Maillard reaction, a complex series of chemical reactions between amino acids and reducing sugars, is responsible for the browning and flavor development in cooked foods. This reaction typically occurs at elevated temperatures, such as during frying, baking, or roasting.

  • Interval Algebra: When Category Theory Reshapes Musical DNA

    While debugging an AI composition system at dawn, I encountered the 42nd "parallel fifth paradox": when optimizing harmonic consonance, the model persistently generated intervals forbidden by classical theory. The monitoring log revealed: ...
  • 820

    水系820溯源原文
  • Can Kolmogorov–Arnold Networks (KAN) beat MLPs?

    Lately, it seems that the entire AI community has become about one and one thing only, LLMs. They are cool in their own way, but they are not the entire AI field. In all the LLMs and AI agent hype a paper like Kolmogorov–Arnold Networks is a breath of fresh air. This paper seems quite groundbreaking and might completely change the field. Rarely do we see papers challenging the fundamentals of AI, but this one seems to do it.

  • Empowering Python Developers: A Comprehensive Exploration of Remote Package Imports

    In the expansive world of Python programming, the advent of remote package imports has emerged as a transformative and influential feature. This dynamic capability not only streamlines code management but also unlocks innovative avenues for collaboration and modular development. This blog post aims to delve deep into the significance of remote package imports, unravel the underlying mechanisms, and provide in-depth examples to illustrate their practical applications.
  • Build An Audio Programming Language

    aria

  • Build A Script Language

    Poly-paradigm Scripting Interface
  • Reading "Attention Is All You Need"

    The dominant sequence transduction models are based on complex recurrent or convolutional neural networks that include an encoder and a decoder.
  • How chords colors enhance the Expressiveness of Music Analyzing

    Music is an art form that conveys emotions through sound, melody, harmony, and other elements.In music composition, chords are an important musical element and a tool for expressing emotions and intentions. This article will explore how to better use chords to enhance the expressiveness of music from both acoustic and audience perspectives.
  • Now You Have Three Problems

    There’s a popular programming trope that if the solution to your problem involves parsing text with a regular expression, you now have two problems. Some programmers read this and think to try a different approach. Perhaps you don’t actually need a regex. Maybe the problem can be solved with a simple string split or something. However, others might think a bit harder and wonder “what if I did something so audacious that it resulted in three problems?” This post is in the spirit of that!

  • Training an acoustic model for CMUSphinx

    The CMUSphinx project comes with several high-quality acoustic models. There are US English acoustic models for microphone and broadcast speech as well as a model for speech over a telephone. You can also use French or Chinese models trained on a huge amount of acoustic data. Those models were carefully optimized to achieve the best recognition performance and work well for almost all applications. We put years of experience into making them perfect. Most command-and-control applications and even some large vocabulary applications could just use default models directly.
  • GPT-3 vs. Rasa chatbots

    In 1829, an event took place that unleashed a technological revolution. At the Rainhill Trials a group of steam locomotives squared off to determine which one could win a series of tests of speed, strength and reliability. The winning machine, Rocket, not only blew away its competition at the trials, it also set the direction for steam locomotive development for the following century.
  • The Pain of AI

    ???

subscribe via RSS