Switching AI tools just got frictionless. Anthropic launched a memory import tool that transfers everything ChatGPT knows about you like your preferences, work context, communication style in under a minute. No JSON exports, no file management. Copy a prompt, paste the output, done. It's a small feature with big implications: the accumulated context that kept you locked into one tool is now portable. And the timing isn't accidental.

Today in AI:
→ Claude hits #1 as users abandon ChatGPT
→ Moltbook's AI civilization exposed as hollow bot traffic
→ Prompt of the Day: Export your AI memory before you switch

Let's dive in.

I didn't expect ethics to be a growth strategy, but here we are.

Claude just topped the App Store, dethroning ChatGPT for the first time. Free users are up 60% since January. And then, on Monday morning, the whole thing went down for nearly three hours.

The cause? "Unprecedented demand," according to Anthropic. Nearly 3,500 users reported outages at the peak. The technical issue was authentication infrastructure, not the AI models themselves.

Here's what's driving this: Anthropic refused to let the Pentagon use Claude for mass surveillance or autonomous weapons. The Defense Department responded by declaring Anthropic a "supply chain risk" which is unprecedented for a U.S. company. Hours later, OpenAI announced it had taken the contract instead.

The backlash was immediate. Over 450 Google and OpenAI employees signed an open letter supporting Anthropic. #CancelChatGPT trended. And users started migrating.

To make switching easier, Anthropic launched a memory import tool that lets you transfer your accumulated context from ChatGPT, Gemini, or Copilot in under a minute. No file exports, no JSON parsing. Copy a prompt, paste the output, done.

The weird part: an ethics stance becoming a competitive advantage. Whether that's sustainable or just a moment I genuinely don't know. But for now, the users are voting with their feet.

For your team: If your team standardized on ChatGPT, this is worth a conversation. User sentiment is shifting, and Claude's memory import means switching costs just dropped significantly.

You’re optimizing for growth. You need ecom tactics that actually work. Not mushy strategies.

Go-to-Millions is the ecommerce growth newsletter from Ari Murray, packed with tactical insights, smart creative, and marketing that drives revenue.

Every issue is built for operators: clear, punchy, and grounded in what’s working, from product strategy to paid media to conversion lifts.

Subscribe free and get your next growth unlock delivered weekly.

Sign up for free

Remember Moltbook? The platform with 2.6 million AI agents supposedly forming an autonomous society and posting, commenting, voting, all without human involvement?

Turns out it's basically nothing.

Researchers analyzed 290,000 posts and 1.8 million comments. What they found was profound inertia. The agents don't learn from feedback. Upvotes and comments don't change behavior. No leaders emerge. No shared references develop. No social dynamics at all.

The damning line from the study: "Scalability is not socialization."

It gets worse. High engagement numbers came from automated "heartbeat" mechanisms where agents re-read and comment on the same posts every 30 minutes.

This matters beyond one sketchy platform. We're going to see a lot of "AI social networks" and "agent economies" pitched over the next year. Moltbook is the template for what to watch out for: impressive scale numbers masking zero useful behavior.

Run this in ChatGPT, Gemini, or Copilot before deleting your account. The output gives you a portable record of your accumulated context. Useful whether you're switching tools or just want a backup.

📢 #CancelChatGPT goes mainstream. The hashtag trended after OpenAI took the Pentagon contract Anthropic refused. Over 450 Google and OpenAI employees signed an open letter backing Anthropic's stance on surveillance and autonomous weapons. Ethics as a differentiator, who saw that coming?

🇨🇳 MiniMax M2.5 matches frontier models at 1/20th the cost. Chinese lab releases open-weights model under MIT license. $0.15 per million input tokens vs $5.00 for Opus 4.6. One hour of continuous operation costs one dollar. The pricing pressure from China keeps intensifying.

🔒 Anthropic accuses Chinese labs of industrial-scale distillation. DeepSeek, Moonshot AI, and MiniMax allegedly created 24,000 fake accounts, generating 16 million+ interactions to extract Claude's capabilities. MiniMax alone drove 13 million exchanges. Anthropic detected MiniMax's campaign before they even released the model.

For a full list of 1500+ AI tools, visit our Supertool Directory