By David Stephen
Sedona, AZ — Imagine that a trained neuroscientist is leading a technology group, and that group is seeking to answer questions about machine consciousness then they hire a philosopher.
[Anything about consciousness and intelligence are brain problems, not philosophical label problems. Philosophers in the consciousness problem are labelists. Consciousness is not a problem for philosophers to solve. So, if a neuroscientist hires a philosopher to solve a brain problem, either the neuroscientist does not want the problem solved or the neuroscientist lacks a deep mind.]
The excellence of Google DeepMind, under the leadership of Demis Hassabis, was about enough before Google merged Google Brain and DeepMind into a unit and made Hassabis the czar.
As a niche leader, Hassabis seems fascinating, but for the entire juggernaut of Google AI, it appears that it is too much for Hassabis to steer, with vision, new ideas, industry leadership and less confusion.
There are choruses of complains about the tons of nonsense that comes with Google AI overviews. Something that Google, it seems, shamelessly preserves, and even with adjustments continue to be less than impressive, compared with any AI, available from other browsers or search engines, excluding the mighty chatbots.
How does an AI leader not have a way to do better than what Google AI overviews is? How does it happen on a leader’s watch, that for a second time, another upstart company would displace Google. First OpenAI, with ChatGPT, now Anthropic with Claude Mythos.
Google has all the compute, all the finance, all the talent, but it seems that vision, transformative paths, speed, change and accuracy have exceeded the organization.
It is true that DeepMind pushed forward focus on artificial intelligence, but the transformers breakthrough came from Google Brain. And after even another team did, DeepMind under the leadership of Hassabis could not beat ChatGPT to market, just to have a separate platform, experimental, away from the main page search.
It appears that Demis Hassabis has peaked. He seems to have achieved all there is to achieve. He comes off as hollow, lacking novel insights in interviews, sometimes resorting to anachronistic platitudes. Demis Hassabis is a unit sample — ongoing anthropology — study that an organization should never make a Nobel laureate a leader, ever.
And yes, the Nobel prize for the most part is a retirement award. Seemingly prestigious but handed at a stage, for many, that the hunger for new things has lapsed. And the individuals, on average, are mostly unlikely to ever have another breakthrough.
Also, the Nobel Prizes for AI, spread between physics and chemistry in the same year, some have said was a result of fear of missing out by the committee.
Even the prize in chemistry, appears to be for who got there first, on something that could have been reached, without Google DeepMind. The premature Nobel to Hassabis made the award seem uninspiring, made it appear like winning does not require a super stellar work.
There are lots of other proteins prediction open-sources aside from AlphaFold. There are several more excellence in computational sciences with AI that makes the chemistry prize seems rushed, especially after some open campaign, by anonymous quotes, before the awards, in FT. And AlphaFold progress so far, has not been validated by any major breakthrough, to cure diseases. The Nobel, particularly to Hassabis, was really a delinquency that will later become a public regret of disrepute for the Nobel medal. It is more like a political appointment for AI — the ruling party.
Demis Hassabis always discusses AI curing diseases. No one has ever asked him that being a neuroscientist, what is mental health in his own words, answering with components of the brain. Like, if he would have any credibility about curing disease, take it on from the home court, and at least make a difference in either neurology, psychiatry or the new third division of neuroscience, human intelligence.
Yes, human intelligence, everyone keeps saying AI would surpass human intelligence. But no one asks, what is human intelligence in the brain? What is a proposed mechanism? How can it be improved, to be competitive against AI?
If anyone on earth should be able to champion human intelligence answers, as a scientific cause, using unlimited resources, it should be Demis Hassabis, but it appears he lacks gas. Demis Hassabis is not a pivotal beginner for major great things that will become of tomorrow.
Noam Shazeer
There was a feature [on October 24, 2025] in The New York Times, A Teen in Love With a Chatbot Killed Himself. Can the Chatbot Be Held Responsible?, stating that, “Shazeer, who grew up in Philadelphia, showed unusual talents early: In high school, he won the gold medal in the hypercompetitive International Mathematical Olympiad. In 2000, he became an early employee of Google, where he would go on to develop the company’s autocomplete technology (the thing that finishes your queries when you search). In 2017, Shazeer and seven colleagues published the paper that set out the concept of the transformer, which sat at the heart of large language models like Meena.”
“The firm’s founding chief executive, Noam Shazeer, belongs on any short list of the world’s most important A.I. researchers. The former chief executive of Google, Eric Schmidt, once described Shazeer as the scientist most likely to achieve Artificial General Intelligence, the hypothetical point at which A.I.’s capabilities could exceed those of humans. In 2017, Shazeer was one of the inventors of a technology called the transformer, which allows an A.I. model to process a huge amount of text at once. Transformer is what the “T” stands for in “ChatGPT.” The research paper about the transformer, which Shazeer co-wrote, is by far the most cited in the history of computer science.”
“Shazeer had bluntly told Google’s chief executive, Sundar Pichai, that A.I. technology would “replace Google entirely.”
“Shazeer, who made some $750 million on the deal, became a vice president. He now co-leads Gemini, the company’s flagship chatbot.”
There is a recent [June 18, 2026] report on CTech, Two years after a $2.7 billion return to Google, AI pioneer Noam Shazeer is leaving for OpenAI, stating that, “Noam Shazeer, Google’s Vice President of Engineering and one of the leaders of the company’s Gemini AI project, has announced that he is leaving Google to join OpenAI.”
“This is Shazeer’s second departure from Google. The first came in 2021, when he left alongside Daniel De Freitas to co-found AI startup Character.AI. The pair returned to Google in August 2024 as part of a reported $2.7 billion deal between Google and Character.AI. The startup became known for allowing users to chat with AI-generated versions of celebrities, fictional characters, and custom-built virtual assistants.”
How does Noam Shazeer leave Google again, given what it took to land him back? The same feeling of a loaded truck on a speedway — that made Shazeer leave due to canceling Meena — could probably be responsible for leaving again.
It could also be a lack of vision, or an ability to try new solutions or make changes in small scales elsewhere, before bringing to mainstream.
Maybe Google should have several versions of Gemini, on separate websites or apps, such that it is easy to have teams’ experiment, especially to develop one version against major rivals or say to tailor against the competence of competitors, and have different teams lead those.
Maybe Sheezer would have stayed. The rules could be safety across all, but experiments with capabilities would be abundant. It could also be other things, since OpenAI has versions of models for public use.
The lack of direction from Demis Hassabis is obvious, regardless of the lead of Google in search, second place market cap, compute, data and all. In fact, for all the advantages of Google, for its AI to be this poorly managed, it shows that prescience has plummeted.
The story of Google AI in June, 2026 is that the subordinates of the leader are leaving, competitors are way ahead, the core product is subpar and the leader goes to all these places to talk, telling people what to do when AGI comes around, when he cannot find a way to even be slightly ahead.
John Jumper
There is a recent [June 23, 2026] analysis on Fortune, Defections from Google DeepMind prompt questions about Alphabet’s efforts to stay at the forefront of AI, stating that, “In recent months, a stream of well-known researchers have departed. Some, such as David Silver, one of DeepMind’s earliest employees and one of the top reinforcement learning experts in the world, have decided to launch their own startups (Silver’s is called Ineffable Intelligence).”
“Just days after the Shazeer news, Google DeepMind researcher John Jumper announced he too was leaving—in this case, to join Anthropic. Jumper shared the 2024 Nobel Prize in chemistry with Google DeepMind CEO Demis Hassabis for their work creating AlphaFold, the AI system that could predict the shape of proteins from their DNA sequences, solving a 50-year grand challenge in biochemistry.”
“He doesn’t strike me as the kind of person who is primarily motivated by money and I don’t think he’d leave Google DeepMind unless he thought the scientific opportunity at Anthropic was actually better—which is much worse news for Google DeepMind.”
How does Jumper, his collaborator, leave? Like with access to endless resources, and possibilities to pitch the board directly, for any innovation, Jumper still had to go? What obstructions did he face at a close range that would have made him leave? Leadership at Google DeepMind is impoverished. There’s no need for any leaks or gossip to see what is obvious.
What explanation did Demis Hassabis give?
There is a recent [June 24, 2026] report on Bloomberg, Google Poised to Lose Two More Senior AI Staffers to Anthropic, stating that, “Jonas Adler and Alexander Pritzel, both viewed internally as key contributors to Google’s Gemini AI model, are set to move to the Claude maker, said the people, who spoke on condition of anonymity as the information is not public.”
“Another researcher, Arthur Conmy, wrote on X Wednesday that he was set to join Anthropic to work on AI safety.”
Demis Hassabis responded “There’s a lot of talent movement between all the leading labs and we win our fair share of the top talent. We have by far the biggest and broadest research bench of any of the labs out there. It’s a ferociously competitive market right now, the most ferociously competitive it’s ever been in the tech industry.”
Oh really. Imagine that the Nokia or Blackberry mobile engineers were fleeing to the iPhone team and this was what the head of engineering had to say. Or, people are jumping ship and a captain says it is just competition, really, like some people find it fun to start and settle in a new organization again?
There is a lot of sleepy type thoughts with Demis Hassabis that it is almost unfathomable how Alphabet is watching their crown, crash, implode and burn. Well, maybe then, there could be more hope to solve human intelligence in the brain, since Google has not contributed to it mechanistically or directly.
Isomorphic Labs
In business, death is not the lack of life. Google is ahead, but those with less are doing better. Things are soaring beyond Google and getting ahead, while Google seems to look like a wasted opportunity.
There are several other verticals of Google that are actually worse. For example, Google News. The glitches are embarrassing. The rankings appear too obviously biased. The default of some stories [in certain topics] for months in Google News results seem like they are choosing winners.
Google news is not just destructive for local news publishers, it is an inefficient curation of news that appears to present unfair access to information. Fairness is not just political left or right.
Even where information is credible, it would look for something else, that is clearly disconnected from what was searched, so that Google can tell who is worthy or not, even for niche topics, even with caveats of postulates.
Google AI also summarizes some titles after a while for news for some publishers. Some stories appear and disappear. Some images disappear and the title takes the entire space. Sometimes too, even what should be listed for news does not. Searching for a topic on Google News and refreshing will bring different results sometimes, scrapping some results.
There are some other browsers and search engines that try their best to be fair, to represent information, as they are sourced, ensuring that people get news and stories, at least that are still credible, and relevant. Google tries to decide what say is accurate. Simply, Google News is information control, for who or what seems preferable. Meanwhile, in mental health and psychiatry, the amount of guesswork by experts, who have nothing to say on causation continues to be abundant, and favored by Google, maybe to feed on status quo.
Demis Hassabis should be retired to Isomorphic Labs. He can cure all diseases at his own pace there and can even solve all world problems and even make humans live forever. Whatever he wants, he should be where he does not use his lack ability to stoke important progress to hold back Alphabet.
Yes, surpassing AlphaGo champion with reinforcement learning was good, but it seems like a Sunday picnic compared with the greatness of the transformer architecture.
Google will sink with Demis Hassabis in charge of its AI. 100%. They can choose to adjust all they like, including experimenting with underlined links in AI overviews, they would lose so much ground, playing catchup, that they will eventually become a lesson in waste.
Demis Hassabis has tried for himself and Google has also tried for him. There is no shame in adopting a receptor agonist against a bad health outcome.