I'm D. I live in Asia (UTC+8) and English is my second language.
Every day I give Grok Bot and Meta Muse the same real task and post who did better, with screenshots.
I don't have Dots or Duet, so those get takes, not fake tests.
What should they try first? ๐
@testingcatalog spotted it: Grok Bot can now import memory from another AI assistant.
Real question from a non-native speaker: half my context is in Chinese, half in English. Does it move over cleanly, or get lost in translation?
Anyone tried it with non-English chats?
Prompt I used (draft only, nothing sent): "A coworker asked me to cover their Saturday shift. I can't, I have family plans. Write me a short, polite message to say no without hurting our relationship."
Asked Grok Bot to say "no" for me. It thanked first, said sorry once, and still left the door open.
Task: a coworker asked me to cover their Saturday shift. I can't, family plans.
Would you send this as is, or cut the sorry?
Claude Duet is rolling out, but only to US Max users first, with Pro next (via @testingcatalog).
Here in Asia, no date yet.
So my always-on agent stays Grok Bot for now. Outside the US? What are you using while you wait?
ANTHROPIC ๐ฅ: Claude Duet is rolling out to Claude Max users in US with Pro US users coming next in line.
> Broader expansion beyond US is planned too, yet there is no timeline.
Expect that EU users may have to wait for even longer.
Not soon ๐
@WIRED ran 6 AI agents through the same 5 tasks. Best detail: all 6 picked the identical China route.
I live in Asia. I want to see that itinerary asked in Chinese. Does the route change?
That's my daily test: Grok Bot, same task, English vs Chinese.
Halloween is Saturday. Give me 3 easy costume ideas I can make tonight from stuff already at home: a hoodie, cardboard, a black marker, tape. Under 30 minutes each. Planning only, no purchase.