Skip to content

Details

2025 was the year AI stopped being “just a model” and started behaving like a system: multimodal foundation models with explicit reasoning modes, tool use, and increasingly autonomous agents became the default across the industry. In this talk, I’ll walk through the most consequential releases and breakthroughs of 2025, focusing on what changed at the foundation-model layer and what that unlocked in agents, voice, video generation, and robotics.

We’ll track the flagship arcs from the top AI labs and platforms—OpenAI’s GPT‑5 era and the rise of productized agents, Google’s Gemini 2.5 to Gemini 3 evolution, and xAI’s rapid Grok iterations. Along the way, we’ll examine the year’s standout modality leaps: video generators that moved toward controllability and synchronized sound, voice stacks that became platform primitives for real-time assistants, and robotics “foundation models” that push agentic reasoning into the physical world.

The goal is having a clear overview of the accelerating pace of AI development, what these shifts imply for developers, product teams, and anyone building with AI in 2026. Attendees will leave with a clear mental timeline of 2025’s defining moves, a practical framework for evaluating “agentic” claims, and concrete takeaways on where the AI stack is headed next.

Speaker: Junling Hu
https://www.linkedin.com/in/junlinghu/

Join online (Link to be added here soon)

7-7:05 pm Meet & Chat
7:05-8:00pm Talk
8:00-8:30 Additional Q&A

Data Mining
Data Science

AI summary

By Meetup

Online talk for developers and product teams; reviews major AI breakthroughs and provides a practical framework to evaluate agentic AI and next steps.

Members are also interested in