🤖 Generative Programmer | Issue #12
AI-Assisted Development, Prompting Techniques, Agent Frameworks
📣 Welcome back! Here’s what’s new in GenAI since last time.
🔗 Share: X · LinkedIn · BlueSky · Email and let’s dive in.
🔥 Top Picks This Issue
🧬 On the Biology of a Large Language Model - Anthropic research
📊 Agentic Software Is Not a Chatbot with Loops - Mike Erlihson
🧠 Avoiding Skill Atrophy in the Age of AI - Addy Osmani
🛠️ Writing Cursor Rules with a Cursor Rule - Adithyan
🧙♂️ How to Think About Agent Frameworks - LangChain
🌊 Technology Radar April 2025 - Thoughtworks
✍️ AI by Hand - Tom Yeh
…and if you dare, 65 more updates to explore 👇
🎯 Featured Article
🌟🌟🌟 On the Biology of a Large Language Model 🧬 Anthropic's research into Claude 3.5 Haiku's internal mechanisms using circuit tracing reveals groundbreaking insights into how large language models perform complex tasks. Key findings include the model's ability to perform multi-step reasoning, plan future outputs and utilize both language-specific and abstract multilingual circuits.
📰 Long-Form Reads
What Agentic Software Really Means: Beyond LLM Wrappers and Hype - Mike Erlihson - Agentic Software Is Not a Chatbot with Loops!🔥🔥🔥
Radar Trends to Watch: April 2025 -- O'Reilly - Mike Loukides - Developments in AI, Biology, Operations, Security, and More 👌
The State of Reinforcement Learning for LLM Reasoning - Sebastian Raschka, PhD 🧠
The “S” in MCP Stands for Security - Elena Cross - The USB-C for AI agents desperately needs security before we plug our data into every model under the sun. 🔓
The MCP Authorization Spec Is... a Mess for Enterprise - Christian Posta
Model Context Protocol has prompt injection security problems -
-MCP promises a new standard for LLM tools, but apparently forgot to pack security in.🥀 The Bitter Prediction - The Ivory Tower - A developer's journey from AI skeptic to enthusiast ends with the sobering realization that tools like Claude Code are so frighteningly competent they might soon render human programmers obsolete.
💰 A new framework for AI agent pricing - by Manny Medina - Manny Medina
How Cursor (AI IDE) Works - Shrivu Shankar (X post)
Stop overbuilding evals - Doug Turnbull
MCP Servers: The New Security Nightmare - MCP Servers are bringing back command injection vulnerabilities like it's 2005 🤦♂️
Avoiding Skill Atrophy in the Age of AI -
From APIs to Capabilities to Support AI Agents - Christian Posta
Building a To-Do List App with Prompt Driven Development - A step by step guide by Andrew Miller
Writing Cursor Rules with a Cursor Rule - Adithyan 🧙♂️
Human in the Loop Software Development Agents - Jirat Pasuksmit from by Atlassian (X post)
From requirements to prototype with Gemini Code Assist
Build Your Own Code Interpreter - Cookbook - Mandeep Singh 🔧
Claude Code Best Practices - Boris Cherny from Anthropic
Principles for Building One-Shot AI Agents for Automated Code Maintenance - Rob Szumski
Secure execution of code generated by LLMs
How to think about agent frameworks - LangChain's manifesto subtly throwing shade at OpenAI's oversimplified take. 🔥🔥🔥
AI Trends Disrupting Software Teams - (this one is by me) - This article outlines five AI-driven trends already reshaping software development, and offers guidance for how different roles can adapt, survive, and thrive in this fast-evolving roller coaster. 🚂
Thoughtworks Technology Radar April 2025 - This edition focuses heavily on the practical evolution of generative AI tools especially coding assistants 🤖🌊
5 AI Use Cases for GTM Market Research - Maja Voje
Senior Developer Skills in the AI Age - Manuel Kießling
What are AI Agents? why do they matter? -
The Art of Validating Non-deterministic AI Responses - Swetha Veluri
Vibe Coding is not an excuse for low-quality work -
Principles for coding securely with LLMs - sean goedecke
Build safe and responsible generative AI applications with guardrails - AWS Machine Learning Blog
🐦 Social Highlights
Claude Code Best Practices at Anthropic - Alex Albert from Anthropic reveals a collection of practical coding patterns using Claude.
Three AIs, one prompt. Who did it best?
Netlify Hits 1 Million AI-Generated Apps 👏
X: Models don't know how to use cursor anymore - they're all being optimized for terminal. That's why 3.7 is and o3 are so awful in Cursor, and so amazing outside of it.
The Ultimate List of Prompts to Style Images using ChatGPT
OpenAI Releases New Models Trained for Developers - Loraine Lawson
How AI Agents Are Quietly Transforming Frontend Development - Alexander T. Williams / The New Stack
What’s new in Cursor_0.49: rule generation…
🛠️ Tech & Tools
Build and deploy Remote MCP servers to Cloudflare
Doriandarko/claude-engineer - Doriandarko
OpenAI's Codex CLI - openai
LLM Guard - The Security Toolkit for LLM Interactions
Serverless Framework's MCP Server
Agent2Agent Protocol - Google's Agent2Agent Protocol promises to revolutionize AI communication 🙄 (X post)
AGNTCY.org - An open source collective for inter-agent collaboration.
GitHub - wild-card-ai/agents-json - The agents.json specification is yet another attempt to standardize AI agent interactions
Teaching LLMs how to solid model - Will Patrick boldly predicts that LLMs will soon transform mechanical engineering by designing CAD models too 🔧
Agent Development Kit - Google announces Agent Development Kit
A curated list of resources about AI agents for Computer Use
Pocket Flow - Turns GitHub repo into Easy Tutorial with AI
Fleur MCP Fleur /flœʁ/ is the app store for Claude
📚 Learning Picks
📋 The AI & ML Resource Library - a solid collection of learning resources.
Foundational Large Language Models & Text Generation - a pleasant-to-read whitepaper on Kaggle
👏 A practical guide to building agents - OpenAI's guide to building agents (X post)
🎭 AI Humor
👋 See You Next Week
Thanks for reading Generative Programmer! 🙏
If you like the content please support my newsletter by subscribing and sharing with your network: X · LinkedIn · BlueSky · Email. Have a great rest of the week!