What’s the real difference between Cursor and Windsurf at this point?

They both seem to use the same models and feel totally interchangeable VS clones.

Is there a reason to use one over the other? Anyone have thoughts or firsthand experience?

Engineer @farcaster, co-founder @bountycaster. Sharing what I learn about AI + development. Previously: Phantom, 0x. NYC 🗽

I've moved on from both of them (also aider) these days. I'm finding I get better results for less effort with agentic solutions. Currently, I'm playing around with ampcode over Claude Code or Friday but that's largely because its PAYG charging works better for exploration.

Established 1968. Cast about parenting, life in 🏴󠁧󠁢󠁷󠁬󠁳󠁿 South Wales 🏴󠁧󠁢󠁷󠁬󠁳󠁿, coding, being older than most around here and general nonsense.

Just a quick tip: try switching to the “slow” completion setting in Cursor/WS. It’s way less distracting than the default and makes the experience feel much smoother

No, I find it distracting. My workflow is becoming... 
If I'm in my editor, it's because I'm writing code myself. I don't want distractions from an AI agent.
If I want it written for me, I prefer to specify my requirement and leave the agent to get on with it.