The 2 simple images below 📷 speak volumes about the speed at which AI transforms.

In July last year, Satya wrote his famous LinkedIn article on the next platform shift (🔗 link in comments). He highlighted how, for the first time, we’re creating a user interface that interacts with machines (reasoning engines) in natural language. Previously, we had to learn programming languages to communicate with machines.

Fast forward a year, and Satya’s AI Tour Keynote in London reveals the leap. 🗣️
What was once a 2-tier architecture now looks like a 3-tier structure. So, what’s changed? 🤔

The new layer, “Memory & Context,” enhances the natural language interface and reasoning engine. Last year’s 2-tier system focused on these as core elements, but the memory addition in this year’s diagram drives AI towards autonomous, human-like agents 🤖 capable of continuous learning and task planning.

This memory allows the system to recall past interactions and use them for future tasks, making the AI more intelligent and adaptive. 🌐🧠

In just a year, we’ve moved from a transactional model to a more advanced rationalization engine.

I have no doubts – 2025 will be the year of “AI Agents”!💡