The Cursor 2.0 update tells us that agents are here to stay
Everything that changed in Cursor 2.0 and why it matters for software engineers
Cursor 2.0 shipped last week! If you’ve been reading the newsletter for a while, you know I’m a HUGE Cursor fan.
The first AI tool I tried was GPT-2, then ChatGPT, then I was an early user of GitHub Copilot. Copilot was pretty cool but it didn’t stick for me. Then came Cursor. Cursor changed the game for me (and plenty of developers around the world.
No longer was I copy/pasting errors and code from my machine into ChatGPT. Cursor made it easy to work with AI without altering my flow.
The Agent made that even more intensely game-changing. But somewhere along the way, Claude Code came out and I stopped using Cursor as much. I think this week’s release is going to bring me back.
So what changed? Here’s a quick rundown:
Composer (custom model)
Agent-focused UI
Parallel agents
Browser usage (THIS IS HUGE)
Paid subscribers get access to the full archive of deep dives like this one, and many expense it to their company’s education budget.