Hey readers! 👋 This week’s AI coding landscape has been dramatically reshaped by OpenAI’s Codex release and the flurry of announcements from Google I/O, while Anthropic’s Claude 4 models are setting new benchmarks in coding capabilities — it’s like watching a tidal wave of AI development tools competing for your terminal. 🌊
🧪 BETA NOTE: This newsletter is still in beta, and we’re working out the kinks. We’d love your feedback to make it better! Just hit reply to share your thoughts.
This Week’s Highlights 🔥
Claude 4 — Anthropic has launched Claude 4 with two new models: Opus 4 and Sonnet 4, setting new benchmarks in coding and reasoning capabilities. (Ed: Huge news from Anthropic the same week as Google I/O. I’m already using Claude 4 Sonnet as my coding model and it’s leaps and bounds better) Anthropic
Claude Code is now generally available — Anthropic is bringing Claude to more of your development workflow—in terminals, IDEs, and via the Claude Code SDK. (Ed: Anthropic really trying to push Claude Code everywhere) @AnthropicAI
Aider LLM Leaderboards — Comprehensive evaluation of language models shows o3 (high) + gpt-4.1 achieving 82.7% accuracy across 225 coding exercises in multiple languages. (Ed: The stats are out for Claude 4, and they’re impressive!) Aider
Why Your AI Coding Assistant Keeps Doing It Wrong, and How To Fix It — Introduces the ‘Constraint-Context matrix’ to help developers frame tasks in ways that align with AI capabilities for better outcomes. (Ed: Interesting framing of how to think about AI-assisted coding) The Pete Blog
The Software Engineering Kawagara Moment — Explores how AI is transforming software engineering like an unstoppable wave, reshaping the industry and changing required skills for future engineers. Pasha
New Tools & Updates 🛠️
Amp Now Uses Claude Sonnet 4 — Amp has transitioned to Claude Sonnet 4, noting improvements in speed, stability, and error recovery. Amp Code
Exploring the JetBrains AI Assistant for Visual Studio Code — JetBrains expands beyond IntelliJ with a VS Code extension offering code completion and documentation generation. The New Stack(Ed: JetBrains released a VS Code plugin? Hell hath frozen over)
CodeRabbit PR reviews, more than meets the eye! — AI code review tool reduces time to first comment by approximately 90%. @coderabbitai(Ed: I know a team that’s pretty happy with CodeRabbit. Worth checking out)
Not using Augment Code yet? — Augment Code rolls out Claude Sonnet 4 with sharper edits and improved context handling. @augmentcode(Ed: Two weeks free? I like free stuff!)
Release v3.17.0 · cline/cline — Adds support for Claude Sonnet 4 and Opus 4, integrates with Nebius AI Studio, and redesigns the settings page. github-actions[bot]
All Hands AI — Organization focused on developing open-source AI software development agents. All-Hands-AI(Ed: Yet another tool, but devotedly open-source)
The Copilot Delusion — Critiques AI coding assistants as chaotic coworkers that disrupt coding processes rather than enhancing them. (Ed: A critique of AI-assisted coding if you want opinions from the bearish side of things) isaiahwp
Building with AI: highlights for developers at Google I/O — Google unveils Gemini 2.5 Flash Preview with improved coding capabilities and introduces new developer tools. (Ed: In case you missed it, here’s everything that Google sprayed AI on) Google
I decided to renew my @alexcodes_ai subscription — Xcode user praises Sonnet 4 + AlexSideBar as “shockingly good” for native Apple development. (Ed: Using Xcode? Here’s a viable option) @SlowbrewedS
the 35 min non-technical guide to codex — Greg Isenberg shares a concise, non-technical guide to understanding ChatGPT’s new AI agent software development tool. (Ed: Nice deep dive into Codex if you want to learn about it in video form) @gregisenberg
Everyday AI Podcast – An AI and ChatGPT Podcast — Daily livestream and podcast helping professionals enhance their careers through AI. (Ed: Haven’t listened to it yet, but looks like some interesting interviews) Everyday AI
😮 (!) A+ code model that runs on local — Devstral, a state-of-the-art open model for coding agents that can run locally, developed by Mistral AI and All Hands AI. (Ed: Domain-specific models are starting to heat up) @saranormous
Made with ❤️ for AI coding enthusiasts! Have questions, comments, or feedback? Just hit reply - we’d love to hear from you!
Your email address is safe with us. We never share your information with third parties.
Data Drift Press, 1388 Haight St #1199, San Francisco, CA 94117, United States
This newsletter was created using AI technology and reviewed by humans at Data Drift Press.