I didn't think I'd be saying this... but I'm switching back to ChatGPT, and you should too. For months, I've been all in on Claude, but GPT5.5 is just too good to ignore. My stack just got a significant overhaul. Here's how I'm repositioning for maximum AI productivity (copy this): The first major shift I'm making is switching my primary OpenClaw agents from Opus to GPT5.5. This switch should be more efficient (have to wait and see), and I've just generally found GPT better for big decisions, brainstorming, and business planning. Because of this, all "heavy operating" Claw agents will be test-running GPT 5.5 for the next couple of weeks. I'll also still run some open-sourced agents for cost efficiency on my sub-agents. Secondly, I plan to use Codex much more going forward. Tbh, I've relied on Claude Code for almost all of my coding tasks, but GPT5.5 looks promising for agentic coding, and I just generally want to test it. Lastly, I've switched about ~50% of my total image generation needs to the new ChatGPT Image 2 (before it was 100% Nano Banana). So far, I'm impressed with Images 2, and I'm excited to see how GPT5.5 performs over the next few weeks with these changes. Where I'm still using Claude: • Creative tasks (writing, Claude Design) • Unique tools/features (Claude Skills, Live Artifacts, Cowork, etc.) • Claude Code - it's been my daily driver for months; I'll definitely still be using it as I'm already quite comfortable with it This setup could significantly boost your AI productivity and could possibly save you money on token usage - be sure to save this!(Miles Deutscher)