梦兽编程™AI_SUITE [ SECTION / AI ]
LLMs
Frontier LLM technology, inference optimization, and training.
31 POSTS01
AI / 2026.03.17 / 5 MIN
AI Programming Hits the 'Review Bottleneck'? Mistral Drops an Open-Source Bomb
Mistral AI releases Leanstral, the first open-source Lean 4 code agent that lets AI both write code and prove its correctness. Let's talk about formal verification and AI …
↗
02
AI / 2026.03.09 / 5 MIN
AI Agents Finally Make Literate Programming Worth Trying
Literate programming has been around for 40 years but never caught on — because maintaining parallel narratives of code and prose is exhausting. AI agents change that …
↗
03
AI / 2026.03.09 / 5 MIN
Locking Down Your Local AI Agent: An Agent Safehouse Review
A macOS tool that sandboxes AI agents like Claude Code and Codex, preventing accidental file deletions and credential leaks once and for all.
↗
04
AI / 2026.03.02 / 6 MIN
Karpathy's Latest Work: Complete GPT in 200 Lines of Code - The Most Adorable AI Tutorial
Andrej Karpathy has done it again! This time he implemented a trainable, inferable GPT model in just 200 lines of pure Python with no dependencies. This might be the most …
↗
05
AI / 2026.02.28 / 9 MIN
Neural Network Reverse Engineering: When Machine Learning Becomes a Puzzle Game
Jane Street released a special ML puzzle: given the complete neural network weights, find an input that makes it output non-zero. This isn't a typical black-box attack - …
↗
06
AI / 2026.02.18 / 9 MIN
GLM-5 Just Dropped: $1/M Tokens Doing What $5/M Does — Is Your AI Subscription Still Worth It?
Zhipu AI's GLM-5 scores 77.8% on SWE-bench with a 744B MoE architecture, MIT license, and $1/M token pricing — just 3 points behind Claude Opus 4.6 at 1/5 the cost. We …
↗
07
AI / 2026.01.11 / 8 MIN
I Spent $400 on Claude Code, Then Switched to Codex
After a month and $400, I went from being a Claude Code fanatic to a Codex user. This is a real user experience report.
↗
08
AI / 2025.12.20 / 9 MIN
Karpathy's 2025 LLM Year in Review: Six Paradigm Shifts That Blew My Mind
Andrej Karpathy's 2025 year-end review: From RLVR to Vibe Coding, exploring six major paradigm shifts in the LLM field and new directions in AI trends
↗
09
AI / 2025.12.18 / 5 MIN
Rust + AI Are Quietly Killing Python's Biggest Advantage (And Most Devs Haven't Noticed)
Python didn't dominate AI because it's fast, but because it tolerates human uncertainty. But when AI starts writing code, that advantage is slowly disappearing.
↗
10
AI / 2025.12.17 / 4 MIN
How I Use Skills to Turn My AI Assistant into a Loyal Workmate
Step-by-step guide to using Skills so Codex and Claude shift from generalists into assistants that follow your playbook.
↗
AI / 2026.03.17 / 5 MIN
AI Programming Hits the 'Review Bottleneck'? Mistral Drops an Open-Source Bomb
Mistral AI releases Leanstral, the first open-source Lean 4 code agent that lets AI both write code and prove its correctness. Let's talk about formal verification and AI …
↗
02
AI / 2026.03.09 / 5 MIN
AI Agents Finally Make Literate Programming Worth Trying
Literate programming has been around for 40 years but never caught on — because maintaining parallel narratives of code and prose is exhausting. AI agents change that …
↗
03
AI / 2026.03.09 / 5 MIN
Locking Down Your Local AI Agent: An Agent Safehouse Review
A macOS tool that sandboxes AI agents like Claude Code and Codex, preventing accidental file deletions and credential leaks once and for all.
↗
04
AI / 2026.03.02 / 6 MIN
Karpathy's Latest Work: Complete GPT in 200 Lines of Code - The Most Adorable AI Tutorial
Andrej Karpathy has done it again! This time he implemented a trainable, inferable GPT model in just 200 lines of pure Python with no dependencies. This might be the most …
↗
05
AI / 2026.02.28 / 9 MIN
Neural Network Reverse Engineering: When Machine Learning Becomes a Puzzle Game
Jane Street released a special ML puzzle: given the complete neural network weights, find an input that makes it output non-zero. This isn't a typical black-box attack - …
↗
06
AI / 2026.02.18 / 9 MIN
GLM-5 Just Dropped: $1/M Tokens Doing What $5/M Does — Is Your AI Subscription Still Worth It?
Zhipu AI's GLM-5 scores 77.8% on SWE-bench with a 744B MoE architecture, MIT license, and $1/M token pricing — just 3 points behind Claude Opus 4.6 at 1/5 the cost. We …
↗
07
AI / 2026.01.11 / 8 MIN
I Spent $400 on Claude Code, Then Switched to Codex
After a month and $400, I went from being a Claude Code fanatic to a Codex user. This is a real user experience report.
↗
08
AI / 2025.12.20 / 9 MIN
Karpathy's 2025 LLM Year in Review: Six Paradigm Shifts That Blew My Mind
Andrej Karpathy's 2025 year-end review: From RLVR to Vibe Coding, exploring six major paradigm shifts in the LLM field and new directions in AI trends
↗
09
AI / 2025.12.18 / 5 MIN
Rust + AI Are Quietly Killing Python's Biggest Advantage (And Most Devs Haven't Noticed)
Python didn't dominate AI because it's fast, but because it tolerates human uncertainty. But when AI starts writing code, that advantage is slowly disappearing.
↗
10
AI / 2025.12.17 / 4 MIN
How I Use Skills to Turn My AI Assistant into a Loyal Workmate
Step-by-step guide to using Skills so Codex and Claude shift from generalists into assistants that follow your playbook.
↗