Gemini 2.0 and the evolution of agentic AI with Oriol Vinyals

In this episode, Hannah is joined by Oriol Vinyals, VP of Drastic Research and Gemini co-lead. They discuss the evolution of agents from single-task models to more general-purpose models capable of broader applications, like Gemini. Vinyals guides Hannah through the two-step process behind multi modal models: pre-training (imitation learning) and post-training (reinforcement learning). They discuss the complexities of scaling and the importance of innovation in architecture and training processes. They close on a quick whirlwind tour of some of the new agentic capabilities recently released by Google DeepMind.

Timecodes
00:00 Intro
02:30 Games and early AI agents
04:28 Weights
09:27 Architectures and the digital brain
10:24 Agentic behaviour
13:31 Digital body
14:09 Scaling
19:02 Data
20:59 Complex understanding and knowledge
25:14 Post training challenges
30:43 Reasoning
33:11 Planning
34:19 Systems 2
37:00 Memory
40:54 Gemini and agentic capabilities

Additional learning:

Thanks to everyone who made this possible, including but not limited to:

Presenter: Professor Hannah Fry
Series Producer: Dan Hardoon
Editor: Rami Tzabar, TellTale Studios
Commissioner & Producer: Emma Yousif
Music composition: Eleni Shaw

Camera Director and Video Editor: Bernardo Resende
Audio Engineer: Perry Rogantin
Video Studio Production: Nicholas Duke
Video Editor: Bilal Merhi
Video Production Design: James Barton
Visual Identity and Design: Eleanor Tomlinson
Commissioned by Google DeepMind

—

[JaimeChavezDJ]

Oriol is genius, but, for a moment, I'd love to acknowledge Hannah, specifically her JOY in receiving this geeky information (that we all love), making it so accessible, and orchestrating the flow of this conversation. Kudos. Keep radiating that jubilant smile... the breath of fresh air!

[gustinian]

Hannah is such a genuinely outstanding interviewer; she has that rare combination of charisma, intelligence, wit and infectious enthusiastic curiosity.

[John-sdli]

Best presenter I have ever seen, she really knew what she said and actively engage in the conversation. Oriol Vinyals is great, good scientist, he don't fail into the hype cycle like many AI's influencer (Stare at you Sam) and give us very clear picture of what is going on now.

[loucasi]

She’s so smooth in her interview style. Amazing work

[OyvindSOyvindS]

These are the best podcasts on the net. It’s so great to witness a host so knowledgable and intelligent, asking questions to get good answers, rather than trying to show off own knowledge.

[piasetzky]

It is a rare thing to watch such an amazing interviewer. Very interesting clip, yet even more thanks to the professor ;)

[lightheartedblog]

I like how Hannah breaks down complex topics into simplest of terms and her curiosity is so engaging!!!

[josephbarney]

I love these podcasts from Google. I make good use of Gemini on my phone and am happy each time it gets smarter. Great presenter and the scientist helping design Gemini is a genius!

[markfitz]

Get a pen out - take some notes - explore the definitions - a great primer for concepts and analogies. Beyond that, for AI enthusiasts like me, this is a cracking conversation that’s worth a few listens. Good job.

[PaulMarshall]

Really enjoyed the conversation and appreciated the insights into the advancements beyond the foundation models. Thank you

[Echolonious]

Ngl Google, just waiting for Live Video with Astra. Agents are awesome, but I make home use robots and repair cars.. so camera would be more versatile for a hands on help than agents.

This IS really cool and helpful for a vast amount of people; I commend you guys.

[JustinHalford]

A wonderful, clear, and compelling window into an exciting future. Kudos to both Hannah and Oriol.

[john_dren]

cant recall when's the last time I was this excited

[ulriknash]

Wow, the idea mentioned at around 22:45 is like the adding of sensors to the AI central nervous system. Right now, the sensory input for AI is the data on the internet. But why not gather data from live cameras, microphones, and other devices, and let the AI learn from this? Where these devices are pointed can be preset, or may become part of the learning process.

[RaviAnnaswamy]

Listening to this after Ilyas neurips talk this is so much more humble and detailed with tons of new insights and ideas that one can pursue.

Ilya might still create another groundbreaking gpt like innovation for sure but the level of innovation engineering and then integration reminds us google ecosystem is so vibrant which we had forgotten for some time
They were just steering their ship last couple of years and seem to be catching up if not overtaking on innovation scale

[EnGmA]

Thanks a lot I really enjoy these podcasts. 🙏

[TrishanPanch]

This is the single best podcast series anywhere.

[teydam_]

Amazing. Im a podcast-hater, but I listen to this and a few UCLA podcasts, so worth it.

[mikaelcodes]

Such a dearth of good interviewers in the world.

[BrianMosleyUK]

47:02 what do we mean by superintelligence really? Add strong reasoning to the amazing scale of memory and inference that we have now, and surely we are there? Perhaps the ability to continue learning as the test time compute generates new realisations that aren't in the training set?