What's New in AI Coding: August 6, 2025
The first issue of this newsletter, with useful links to read!
Welcome to this AI Coding Daily newsletter on Substack.
After the initial success of my new YouTube channel with 1,000+ subscribers in no time, I decided to start this more “intimate” way of sharing my experiments with AI coding.
My current plan is to share the links to my own content over the last week, as well as useful news/tips from the AI community I’ve read/watch.
Let’s dive in!
My YouTube Videos
What I published over the last week:
Claude Code 5-hour Session: How Many Tokens Used/Left? (5 min)
CCUndo for Claude Code: Reject or Revert Code Changes (3 min)
Cursor Rules and AI Guidelines: More TESTING Instructions? (1 min)
Industry News and Tutorials
There’s soooo much stuff happening. How to keep up?
No worries, I will try to help. Here are the main things I personally have read/watched over last week.
New Claude Code features: Microcompact, Subagents, PDF support
cui: Claude Code Web UI. A modern web UI for Claude Code agents. Start the server and access Claude Code anywhere in your browser.
Qwen3 Coder has a 5.32% diff edit failure rate in Cline, based on real-world data. It's right alongside Claude Sonnet 4 and Kimi K2 as an excellent model in performing diff edits.
Cursor 1.3 is out! You can now collaborate with Agent in your terminal, clearly see context window usage, and make faster edits.
Claude Code best practices: 26-minute video by Cal Rueb from Anthropic
Reddit user: “Canceling Claude Code due the MASSIVE performance drop in the last 2 weeks” (and others in the comments report similar issues)
Revenue of OpenAI vs Anthropic for 2025: Anthropic's API $3.1bn revenue is dominated by coding, with two top customers, Cursor and GitHub Copilot, generating $1.4bn alone.