Google DeepMind

“Demis Hassabis stepped down as the CEO of Google DeepMind.”

“Both teams got a 35 out of 42. There was a bunch of back and forth because OpenAI validated with former IMO gold medalists while Google's result was graded by the official rubric that year, which is internal to the organization.”

“I think that the AI company CEOs understand that they're on the path of building wildly smarter than human systems, like super intelligent AI systems. They understand that we don't really have a like clear thought through plan for how to manage the risks from …”

“We would literally take a large language model and we would turn it into the robot brain, which this is like very simply stated, but honestly at the time was like a kind of a crazy idea to do because everybody was just trying to make robot brains in many other…”

“Demis Asabas is stepping down as chief executive for Google DeepMind to become chairman and chief scientist. Google CEO Sundar Pichai said in a post on X. Google DeepMind technology chief Koray Kavukcholu is taking on responsibility for all AI model developmen…”

“So I actually think it's good for AI that Demis will be focusing more on the scientific side, probably more on the world model side and not deeply in LLM. So you'll have basically the talent that open AI and anthropic and potentially what's left of the AI orga…”

“The point is if these systems start improving themselves, like less than two years from now, the more cynical take is that if you're one of the 1000 people who have the most coveted of all skill sets, which is in training, pre-training, the sort of raising of …”

“Sources tell me he's already been disengaged from day-to-day management for a while now, but he was a firewall between DeepMind and the rest of Google, even as the two got pulled closer over the last couple of years.”

“He announced that he's stepping down as CEO of Google DeepMind. He's also stepping up into the role of Google DeepMind chairman, and he will also be Alphabet's chief scientist.”

“Google DeepMind CTO, and Google's chief AI architect, Kore Kovu, Kovu, I can't pronounce it, I'm sorry is being promoted to SVP and will lead DeepMind, so Kore is now running DeepMind, but interestingly, there's no new CEO of DeepMind”

“Google, the deep mind organization has, you know, made some real contributions to the space, but has been lagging. It's been a point of embarrassment. For, I would say, for Google, for, I would say, a year plus. They've always felt behind, even though they cre…”

“Stock's down. It was down five percent around the open. It's recovered a little bit, just down three and a half percent.”

“The last substantial piece that Demis put out was called The Framework for Frontier AI and the Dawning of New Age, in which he argues for a frontier AI standards body that evaluates frontier models before they're released for risks associated with superintelli…”

“Obviously the acquisition was for two billion dollars and went over to DeepBind, and then I decided to give up all the acquisition money to start trajectory.”

“You see people like Google DeepMind, Microsoft, other places like Meta, either building their own lab or running experiments at someone else's lab to get that data back.”

“And I think of the prominent AI leaders, I think Demis does a pretty good job of actually Translating the technology into more quantifiable and tangible ways for, like, your every person, because I think there's a lot of focus on, like, benchmarks and a lot of…”

“After AlphaProof, kind of like, we didn't see a lot of the formal math you know, results or kind of progress from Google DeepMind, and that's actually because of reasons that are not necessarily technical.”

“Gemma four is just out. It's the most capable open model we've released so far. We already tried to compact as much intelligence per parameter as we could, bring all of these multimodal capabilities.”

“The Gemma team is actually relatively small. We have, like, two or three PMs. We have one marketing person, and then there are, like, engineers and researchers working on shipping this.”

“I mean, if you look at Gemma, you compare to how we were one year ago, I would say Gemma four is matching state of the art from one and a half years ago for most. Things.”

“Multimodal wise, the smaller models can understand audio Images and short videos, so, 30 to 62nd videos and audios.”

“The other thing we do not support yet is video with audio, so we can understand, like, video input or audio input separately, but if you want to pass, like, in the same, from both the visual part and the audio part, we still need to do some improvements around…”

“I think especially like the model quality is still a bit worse from what you would get from the normal autoregressive model.”

“For Gemma four, we had 50 To 60 partners. And some of them were like, oh yeah, we're going to try and fine tune the 27 B model for this vision task. And they were like, oh, actually the model works too well out of the box. We don't need to fine tune it. Yeah. …”

“So for example, within the team, we are building skills to do experiments and ablations and evaluations and how the research team can use all of these agentic tools as part of their research process is also quite interesting.”

“I do think the next generation of fine tuners will not be, I mean, will be people that are not coding at all, right? Like one year ago, we had to write like our own Colab with Transformers or Oncelot or whichever library of your choice. I do think as we like k…”

“If you want to do, like, deeper research in the architecture, my hunch is that most likely this will not be, like, automatable, at least in the next one or two years.”

“You can now simulate real places by grounding Genie three experiences with street view imagery.”

“It could be that there's still one or two big ideas left that need to be cracked. I don't think it's more than one or two, if there are out there. And I think You know, my betting is about fifty-fifty if that's the case.”

“But I think for now there's the assumption we make is that, you know, a year later after one of our leading, you know, pro models or frontier models goes out half a year later, you'll have them in the, The really tiny, almost edge models.”

“For now, what we've decided is that our edge models, the things we want to use for Android, and glasses, and robotics it's best that they're open models. Because they're vulnerable anyway on the, once you put them out on the surfaces, so they might as well be …”

“And I think we're about 10 years away probably from something like a virtual cell, like a full virtual cell”

“The lesson I've learned from all the alpha projects we've done, specifically alpha go and alpha fold is I think the techniques we have and the problems I look like to look for are great in, if this, if the situation can be described as massive combinatorial se…”

“We've always defined, we've been very consistent how we define AGI as basically a system that Exhibits all the cognitive capabilities the human mind has. And that's important because the brain is the only existence proof we have that we know of in, maybe in th…”

“I mean, I think, look, I've got a probability distribution around the timings, but I would say there's a very good chance of it being within the next five years.”

“I think compute is the big one, not just for the obvious reason of scaling up your ideas and your systems as you know, the scaling laws as they're called, you know, keeping on building bigger and bigger architectures with more and more parameters. And as you d…”

“Well, I think on the positive side and the things obviously I've, I spent my whole career and life building towards AGI is, I think it will be the ultimate tool for science and medicine. So in terms of advancing scientific discovery finding cures to diseases, …”

“We're working on that right with our partners at Commonwealth fusion.”

“I think we paid a bit more of a fixed cost upfront, but we designed the Gemini models to be very multimodal from day one.”

“Like in a safe setup, you know, where they can like, because right now we definitely, we don't, we're not confident about, you know, them doing the right things all the time and measuring like how much they can push and how long they can push a task is very di…”

“Not easy to pinpoint like specific things, but again, like, you know, this is just like my personal opinion and maybe I have colleagues and like the other people like sharing this with me, but I think we're underestimating how hard like jagged intelligence is …”

“Less than two years after the messy shotgun marriage that created Google DeepMind, Demis' team had closed the technical gap.”

“Future AIs would behave in the same way. Task, for example, to help solve energy scarcity by inventing a superconductor, an AI might draw up a reading list, conduct experiments, invent novel materials, and so on. Pursuing its goal over the space of a year or m…”

“In this particular case, you know, the hardest task we were working on, we got like to five percent performance with Gemini, 1.5 flash. This was a while ago. And then when we added on the reasoning strategies, we went from five percent to 95%.”

“I mean, especially if you're talking about these big pre-trains, it's really just these four. No one else is really at that scale.”

“We want to build interdisciplinary teams. How do you create a culture where you can have a bioethicist next to a computer scientist and a neuroscientist? Because we think that's really where the magic happens and unlocks the work.”

“Over the past few years, we've been doing a lot of work around one science area, learning science. How do people learn and can we improve it? And then this year was really, Dennis was like, okay, Gemini is good enough. It's time to infuse Everything we've done…”

“And then you've got all these other, you know, special kind of ambitious projects working on things like Genie build world models. You've got work going on to build special things for Waymo and enhance the models that Waymo's that lead to Waymo, the Waymo driv…”

“You now have three and a half million researchers accessing it in over a 190 countries.”

“What they found was on average they were sending 10 Saving 10 hours per week per teacher, and their post-it notes were how they were using their time, which was, I'm getting time back with my family. I can now do lesson plans for different learners of differen…”

“The alpha genome, the alpha gnome, the one that you've just mentioned, our material science, was really exciting because we basically went from 40,000 known stable crystals to 400,000 plus. That are now being tested with research, in research and in labs.”

“And so at Davos a few weeks ago, I think both Demis and Dario said some variation of, yeah, like if we were the only two groups building this technology, We would find a way to get together and figure out how to slow down this frantic pace. Like it's going too…”

“From our point of view, we have no plans at the moment to do ads, if you're talking about the Gemini app, right specifically.”

“Actually it turns out you can wring more more juice out of the existing Architectures and data. So there's plenty of room, I think, and we're still seeing that in both the pre-training, the post-training, and the thinking paradigms, and also the way that they …”

“I'm definitely a subscriber to the idea that maybe we need one or two more big breakthroughs before we'll get to AGI, and I think they're along the lines of things like continual learning, better memory, longer context windows, or perhaps more efficient contex…”

“I don't think AGI should be sort of turned into a marketing term or for commercial gain. I think there is always been a scientific definition of that. My definition of that is a system that that can exhibit all the cognitive capabilities humans can. And I mean…”

“I think an AGI system would have to be able to do all of those things to really fulfill the original sort of goal of the AI field. And I think, you know, we're five to 10 years away from that.”

“Why this is why we worked with Gemini as being multimodal from the beginning, able to deal with, you know, video, image and eventually converge that all into one model, that's our plan, is that it'll be very useful for a universal assistant as well.”

“And it's been incredibly gratifying to see, you know, three million researchers around the world use it in their important research.”

“I think in future Almost every single drug that's discovered from now on will probably have used AlphaFold at some point in that process, which is, you know, amazing for us, and, you know, really, that's what we do, all the work we do for.”