Grok, Meta AI — assistants where you already are

Some assistants you go to; some come to you. This lesson covers two of the second kind — and completes a pattern you’ve now seen three times, so you’ll recognise it in every future launch: the assistant embedded where the attention already is.

Grok — the assistant on X

Built by xAI, Grok’s calling card is its connection to X: it can draw on what’s being posted right now, which suits “what are people saying about —?” questions the way answer engines (E2·L5) suit sourced research. Its default tone runs more informal than most. Two cautions follow directly from your training: what people are posting is a signal about conversation, not a verified account of events — treat trending claims as leads to verify, not facts to repeat; and an informal, funny register can make hallucinations more persuasive, not less (E1·L7’s tone warning, doubled).

Meta AI — the assistant in your chat apps

Meta AI sits inside WhatsApp, Instagram, and Messenger — for many people in many countries, the first AI they’ll ever use, because it arrived in the app they already open fifty times a day. It handles the everyday strong-zone tasks (drafting, explaining, ideas, image generation) with zero setup. The judgement to carry: a group chat is a social space — think before pulling an AI into conversations containing other people’s messages, and know your region’s feature set varies.

The embedded-assistant checklist

For any assistant that lives inside a platform — these two, Gemini in Google, Copilot in Office, and whatever ships next year — ask the same four questions: What can it see in this app? What does the platform do with these chats? Is convenience choosing the tool, or am I (E2·L4)? And does my use involve other people’s content?

Check your understanding

1. Grok reports 'what's trending' about a breaking story. The right treatment is: Trending equals confirmed A lead about the conversation — verify events through sources before repeating Ignore it entirely; platform data is useless

2. Before using an assistant inside a group chat, the distinctive extra question is: Which model tier is it running? Does my use involve other people's messages, and is that appropriate here? Is the assistant's tone formal enough?

Recap

Grok: real-time platform sensing, informal tone, verify before repeating. Meta AI: zero-setup AI inside chat apps, with a social-space judgement to apply. Next: a genuinely different tool — the assistant that answers ONLY from your documents.