As Meta AI moves into Threads messages, Satya Nadella urges companies to retain their metadata and model gateways; the two developments place control at different layers.
TechCrunch reports that Meta added Meta AI to Threads direct messages and began the global rollout on Monday. The assistant was already in messages on Facebook, Instagram and WhatsApp. Users can share posts, images, links and videos and ask follow-up questions. Those who do not want it can mute @meta.ai, choose Not interested or hide replies. Use in the public Threads feed remains limited to testing in a few markets.[1]
Microsoft chief executive Satya Nadella, speaking on CNN's Fareed Zakaria GPS, warned companies against becoming dependent on one AI provider. The recommendations reported by TechCrunch include retaining a company's own metadata and usage data, training its own models, operating an AI gateway that separates prompts from models, and avoiding single-vendor lock-in through integrated coding tools. TechCrunch notes that Microsoft invests in both OpenAI and Anthropic, and that the advice therefore also serves Microsoft's own interest.[2]
The developments address control at different scales: who decides when an assistant appears and which model runs underneath? In Threads, control sits in mute, Not interested and hide-replies routes, though Meta places the assistant on an existing messaging surface. Nadella's enterprise route is a layer separating data, prompts and model choice from the provider. Because he distinguishes consumer data sharing from enterprise risk, the stories identify separate control points in personal use and organisational architecture.[1], [2]
