AI 教程与实战

The red leaves at the SAIJOU mountain spend the whole summer getting ready before they bloom. After...

A version number is a promise: nothing a client depends on changed without me saying so. For MCP...

Derivative turns software requirements into executable Python, but generated code is never allowed to...

The problem We build Mushare, a messaging network for AI assistants: you tell Meta's Muse...

On the first day, I planned nine tasks. Three got done. The other six ran into familiar problems:...

At Autoadify we build Workflows: a schedule, a new Shopify/WooCommerce product, or a post that's...

This is a submission for the Kaggle Benchmarking Challenge. What I Benchmarked ...

When I was about fourteen I spent my days in a computer club in Romania. Anyone could walk in. CP/M...

I built a small retrieval-augmented generation system over a corpus of research papers and wanted to...

Twenty-one sentences, no pretrained model. Drag to rotate, click a word to see its nearest neighbors,...

This article was originally published on the Zenrows blog. In this tutorial, you'll build a Mastra...

Why full fine-tuning an 8B model requires 130+ GB VRAM, how low-rank matrix decomposition cuts parameters by 99%, and how weight merging enables zero-latency serving.

TL;DR: Credentials conflate authentication and authorization: Long-lived API keys,...

Months ago, a commenter on this series proposed a test: if the same backend really stands behind both...

I've spent the last stretch building a SOC 2 readiness tool, and the most useful thing I learned has...

You know the feeling. You've just spent 3 hours on a fix, your brain is mush, and now you have to...

If you have been coding for a while, you already know the feeling. You spent three months on your...

This is a submission for the Hacktoberfest Open-Source AI Challenge Week 1: Touch Grass ...

Exact numpy search over 100k embeddings takes 3.5 ms on 2 vCPUs. Measured latency, memory and HNSW recall to show when an index actually pays off.

When building LLM applications, we often rush to pass user data directly into our prompts or RAG...

This is a submission for the Kaggle Benchmarking Challenge What I Benchmarked One of my...

How I built a quiet heartbeat for my AI assistant: HEARTBEAT_OK, exact-match silence, and a...

Fair warning: I'm not a physicist, sociologist or scientist of any kind. I found a couple of studies,...

Every coding agent I use has the same habit. It does the work, writes a lovely summary, and stops....

Earlier this year I built a test-quality harness called killcheck and wrote about it here. The short...

Most "AI invoice automation" demos stop at the fun part: a model reads a PDF and spits out JSON. The...

Ever wondered how an AI can call you, understand what you say, respond naturally, and even book an...

Every few weeks someone posts "provider X is serving a watered-down model" with a handful of...

Ditch brittle coordinate clicks and predatory RPA fees. Build a self-healing, local vision-guided GUI automation engine in Python.

Agentic Resource Discovery may solve one of the biggest scaling problems facing AI agents: How does...

This is a submission for the Hacktoberfest Open-Source AI Challenge Week 1: Touch Grass ...

We took routebase.dev from 87 to 100 on the open Agent-Friendly Docs Scorecard in one day. What failed, the four fixes, and the CDN header detail that nearly broke it.

description: "A practical, ~200-line guide to building a read-only MCP server that connects Claude...

Ankita v2.5.0: a desktop companion island that never misses an approval Version 2.5.0 of...

Malware that has no orders inside: CLOSEDQUORUM asks DeepSeek, Qwen, Mistral and Gemini what to do next. The pattern is no longer in the binary, it is in the outbound traffic.

This is a submission for the Kaggle Benchmarking Challenge A Kaggle benchmark about the most boring...

I Put ChatGPT in Charge of Claude Code I love Claude Code. I have spent an unreasonable...

I let an AI agent make an infrastructure change An AI agent recently took an...

Fine-tuning is usually discussed as a way to improve LLM accuracy. But there is another benefit that...

A field study of 116 generation call sites across 14 public repositories, including five of the SDK vendor's own. The output cap is the one almost nobody sets.