Changes at Google DeepMind: Demis Hassabis from CEO to Chair, Jeff Dean departs

Wow. Jeff and Sanjay both departing. Truly end of a golden era.

There is an entire cohort of work-optional very senior engineers for whom one of the last reasons for hanging on was "at least Jeff and Sanjay are around".

It seems like the real news is Jeff and Sanjay are leaving Google, and Demis is effectively replacing Jeff as Chief Scientist for all of Alphabet.

The bigger deal is the departure of Jeff and Sanjay, rather than Demis moving into a different role.

So, in last several months, all the prominent names Google lost: Demis Hassabis (technically still with google but these things are usually presented with a spin), Jeff Dean, Sanjay Ghemawat, Oriol Vinyals, Quoc Le, Noam Shazeer, John Jumper, Jonas Adler, Alexander Pritzel, David Silver, Denny Zhou, Fernando Pereira, Alex Turner

And all the prominent names Google gained: NULL

Combined with no gemini frontier GA release in about 14 months. You have to have created an environment pretty hostile to innovation for this to happen

> Lastly, after an incredible 27-year run, Jeff Dean is at a moment where he wants to try something new, and we’re excited to support him in that. Jeff and Google Senior Fellow Sanjay Ghemawat are launching an independent public benefit corporation to accelerate discoveries in ML, science, and engineering.

Oof. Good for Jeff and Sanjay (who just joined Twitter), bu that is a big loss for Google. Google stock is down 5%. It might not be much of an exaggeration to say these two are worth ~$200 billion.

Just some random tidbit: When I was in high school I bought every computer gaming magazine I could get my hands on and could afford with my allowance.

So when Deepmind first made headlines I recalled an article in the UK magazine, Edge, which had an article about a game called Republic being developed by a team of former Bullfrog employees lead by one Demis Hassabis.

Out of interest, I just checked Internet Archive, and lo and behold I found it [0]

I see Wikipedia [1] also mentions that he was the lead programmer on Them park and worked with Peter Molyneaux at Lionhead during the development of Black & White.

It doesn't add much to the story under discussion, but it makes me think at the time I wanted to be a game programmer but my parents encouraged me to go study engineering instead.

[0]: https://archive.org/details/edge-issue-078-november-1999/pag... [1]: https://en.wikipedia.org/wiki/Demis_Hassabis

I think there's one weirdly simple reason DeepMind isn't doing as well as OpenAI and Anthropic. I may be wrong on this.

OpenAI and Anthropic went from tiny startups to huge companies. As a consequence the stock options/RSU's offered to the employees paid off a far higher percentage ROI than any stock options a DeepMind (and thus Google) employee would get (since Google is already huge). This disincentivizes people who truly believe in the economically transformative power of AI to work at Google since their benefits will be capped by Google being large + having public company obligations.

DeepMind had a generational run as a pure AI research lab. AlphaGo, AlphaZero, protein folding, tensor improvements, weather forecasting, GNoME and so much more. Google leadership saw all this and went “now go generate a multi trillion dollar commercial business and beat OpenAI and Anthropic” and the results were, predictably, failure. Such a shame.

Today I saw one of the most talented engineers I know and worked with leave DeepMind and now I probably know part of the "why". Dark clouds hovering over Google's AI game.

https://xcancel.com/JeffDean/status/2085034604172603724

> Announcing Discovery Loop!

> I am very excited to announce that, along with my longtime friends and collaborators @Sanjay_Ghemawat, @OriolVinyalsML and @quocleix, we are founding Discovery Loop (@DiscoLoopAI), a Public Benefit Corporation whose mission is to automate machine learning, science, and engineering to accelerate discoveries and progress. The four of us have worked together for 14 to 30 years, and have helped build some of the world’s most used products, infrastructure and AI models, and we’re excited to turn our attention to this ambitious endeavor.

gemma-4 is an incredibly good model, beating several others five time its' size. Hope Jeff's departure won't impact their next-gen products! And best of success for their new ventures!

Everyone commenting here seems to take this to mean he's leaving Google, when that's not what the article says.

With the top AI personnel leaving Google, I wonder what will happen to their Gemini. Right now they still have a lead in inference hardware with TPUs but with both Anthropic and OpenAI developing their own chips, I wonder how long until either one will catch up.

Who knew the real revenue unlock wouldn’t be based on how much paranoid red-teaming the model underwent to resist users jailbreaking its ‘alignment’, and instead more on whether the model is post-trained to use ‘sed’ and ‘git’? Poor Gemini

Scientist-heavy orgs that want to solve everything in token space may overtook tool use; meanwhile Anthropic has been super focused on MCP, Claude Code etc for over a year

WoW, so basically all senior members of GDM are now mostly gone? I thought Google was lagging in coding AI, but this suggests bigger issues.

I am not fan of the future of AI but this I am not sure how I feel about Google's (also Amazon recently laid off it's AGI team) AI initiatives blowing up before all the new AI Labs.

I don't get why Google failed here? maybe after a decade someone will write about it candidly.

What happened to the "strike team" that Sergey Brin was leading to try and improve Gemini's coding performance?

Deepmind always worked on some of the coolest architectures. Following things like AlphaStar, AlphaGo, and more were extremely exciting and felt like the hacker persona of machine learning. I hope Google can take advantage of this awesome team. They've done incredible work.

I found this article more thorough:

https://www.reuters.com/business/google-shakes-up-ai-leaders...

> Google DeepMind: We are building strong momentum: Flash is in high demand, our Cyber model is live, and Gemma models have surpassed 900M+ downloads

Considering these are the best stats they could find, gemini usage+general situation must be really, really bleak.

High demand means nothing. A model being live is nothing to brag about. And gemma downloads also can be from auto CI pipelines etc. Nothing concrete

From https://www.ft.com/content/61d41764-f2f7-4906-a112-ff3073972...

> The moves suggest that DeepMind will be absorbed into Google’s broader business. [...] > They added that while Google DeepMind would not become purely commercial, it was integrating further into Google’s business.

I guess might be quite a change (but writing was on the wall, the Deepmind -> Google Deepmind part was the first step)

The first sentence is quite telling

>We’ve got amazing talent, world-class compute and products…

Products are third on the list. Google is an incubator for talent first and foremost. Products are an afterthought

A few points:

1) the simplest explanation whenever a bunch of people leave [x] at a company at the same time, is that [x] is becoming less important to that company moving forward. It's not proof, but you know, everyone is trying really hard to say that's not what's happening here. Sometimes the simplest explanation is correct.

2) rumor is that Sergey Brin, co-founder of Google, got bored during pandemic lockdown and eventually returned to Google in a lower profile role, related to AI. I have to think that he still has a lot of say in what goes on in Google relative to his interests.

3) Yahoo Finance article says Hassabis "has long prioritized research over profits", so his replacement might indicate that Google has decided it is time for this stuff to pay for itself?

End of an era for Alphabet.

At least judging from the headlines, it seems Google is far behind on AI. Fable/GPT 5.6 and recently Kimi get a lot of attention, solve long-standing math problems and lead in coding. It seems that Google's supposed advantages of very deep pockets, original talent, vast amounts of data and unmatched distribution are really not making that much of a difference. What is going wrong?

I wonder if this has anything to do with it

https://www.lesswrong.com/posts/iKm2FhpWkuuBojm82/why-i-left...

Why form a "Public Benefit Corporation"? There must be some kind of access available that a regular for-profit corporation is excluded from. Politics perhaps? AI tells me PBCs can be shielded from shareholder lawsuits. Also "Founders can maintain vision control even as outside venture capital enters the cap table".

Seems like a good way to spend investor dollars without consequences while maintaining control. Maybe a bit cynical but i can't figure out why they'd form a PBC over anything else.

This seems bad for AI safety/risk. Does DeepMind have any checks on model alignment now? What's stopping them from using AI for military/surveillance purposes?

Publicity benefit corporation HAHAHA typo, think they mean public benefit corporation. Anthropic however is DEFINITELY a publicity benefit corporation.

If all the other AI companies are promising AGI by Q4 of next year, what else can you do to satisfy shareholders than also jump on that same bandwagon?

My read: managing bureaucracy is a boring job esp. if you’re financially in a good spot.

Some think Gemini is falling behind in benchmarks so there was a shakeup at the top. I don’t agree with it.

That said, winning this LLM race is difficult given the number of talented people working on it across the world.

It's chuckle-worthy that Axios webiste reads:

> ... independent publicity [sic] benefit corporation in which Google ...

A Freudian slip? https://archive.is/SxFjr.

Wow we just had a post on here a couple weeks ago about Jeff Dean supporting dissident voices at Google.

Woah, Jeff and Sanjay are big losses. The hits keep coming for google ai

Chinese labs are the proof that there is no need of big names, but of the right mindset and agility. It's those last things that Google truly misses, but now they are missing for a long time, and outside the AI divisions too, in almost every department of the company.

Jeff Dean leaving Google should be the headline but big news all around for Google AI nonetheless

Demis stepped down? Who cares? Jeff and Sanjay are leaving Google. We are in uncharted territory.

The Gemini 3.5 Pro delay has been catastrophic for Google. I'm really curious what has gone on behind the scenes there.

But missing out right as AI coding agents become genuinely deeply capable and useful is just an immense failure.

Considering Hassabis co-founded Deepmind to pursue AGI, I take this pivot as a tacit understanding on his part that Deepmind isn't on a path towards AGI in the near future. They might at some point, but they cannot be close in his estimation, otherwise why leave now?

If you lost your only 2 Senior Fellows, you deserve to be replaced yesterday

Makes sense to me... leaving to start their company with Google being an investor.

Clayton Christensen [1] says (paraphrasing) it's a good idea to spin-off (or invest) in a startup that you've a say on, before the ecosystem sprouts a seemingly-non-entity and gently disrupts your market.

Interestingly (diff strategies i guess) Apple tends to acquire (Q.ai) rather than invest/venture (as google in OP).

[1] The Innovator's Dilemma: When New Technologies Cause Great Firms to Fail.

Between Shazeer and these guys

Looks like Gemini's sub-par performance is claiming heads

Google's stock dropped 5% right away, 200billion of value

Hassabis is being promoted to Chair of Deepmind

Dunno what the longer term effects of this and the other departures will be, but in the world of vibe-finance Alphabet dumped 4.15% of share value as of this minute.

Welp, the market didn't like that. GOOG dropped about 5% on that news.

Google loves to kick senior execs upstairs.

Whatever happened to Prabhakar Raghavan? Got kicked upstairs and we barely hear from him nowadays.

It's only a matter of time before Demis leaves and joins Anthropic or OpenAI.

can anyone with corporate background decipher if that is good for demis or not?

This move was expected when Google shifted focus to Gemini-only.

Same day Google is rolling out layoffs.

Absolute earthquake at Deepmind the last few months...

That's..huge. Hassabis has been the most important figure of the last 15 years in AI.

Even in indirect ways. OpenAI itself was founded because Musk got fixated on "stopping" Hassabis.