Twelve dimensions across seven providers. Re-graded as announcements land, with the market's reaction shown alongside.
Meta · 2026-09-17 · Menlo Park
Connect is where Meta's ambient-AI thesis shows up: the next Llama, Meta AI across the app family, and the Ray-Ban / AR glasses as an on-device agent surface. Watch on-device AI, conversational UX, and whether open weights translate into a coherent platform play.
Watching META for the market's read on the keynote.
How much capability moves onto the glasses and phones at the edge?
Does the next Llama re-take an open-weights lead, and on what hardware?
Any sign of an actual agent runtime or developer platform beyond raw weights?
Tracking live. No announcements graded yet. Entries appear here the moment something lands.
Business Agent rolls out across WhatsApp, Instagram, and Messenger, marking Meta's push into agentic commerce.
Meta's plan to ship four inference-optimized MTIA chips in two years: purpose-built silicon over general GPUs.
Profiles the ~50-person TBD Lab at the core of Meta's superintelligence reorganization.
Each giant's AI strategy flows from its computing philosophy; Meta's 'personal superintelligence' is augmentation.
How Apple, Google, Meta, Microsoft, and Amazon are differently positioned across infrastructure, models, and distribution.
Camera-based AI on by default and stored voice recordings: the privacy stakes of glasses as an always-on AI surface.
The Meta AI app, built on Llama 4, leans on Facebook/Instagram profile data for personalized assistance.
The Llama API preview (OpenAI-SDK compatible) and Llama Stack partner integrations: Meta's developer-platform play.
Meta submitted an experimental chat variant to LMArena that differed from the public weights, a credibility hit.
Zuckerberg's four AI bets, and the argument that Meta spends enormously yet lacks a coherent direction.
Llama Stack standardizes agentic app development across on-prem, cloud, and on-device via a unified API.