Google is sick

区块律动BlockBeats

特邀专栏作者

2026-06-26 07:09

This article is about 5263 words, reading the full article takes about 8 minutes

Perhaps the best time for a place is precisely when it begins to show you the door.

AI Summary

Expand

Core Thesis: In 2026, Google experienced severe brain drain, with multiple core AI researchers—including Noam Shazeer, co-author of the Transformer paper, and Nobel laureate John Jumper—successively leaving to join competitors like Anthropic. This reflects its predicament as an "AI talent nursery," where lengthy internal decision-making processes, resource allocation battles, and difficulties in product deployment are systematically eroding its ability to retain top talent.
Key Elements:
1. Noam Shazeer, a core author of the Transformer paper, left Google again to join OpenAI less than two years after the company spent $2.7 billion to buy back his services, following the reassignment of his project's computing resources to DeepMind.
2. AlphaFold leader and Nobel laureate John Jumper announced in June 2026 that he was leaving DeepMind after nine years to join Anthropic, triggering a roughly 5% drop in Alphabet's stock price and an evaporation of approximately $225 billion in market value.
3. Within a month, at least five top Google AI researchers departed, with four (including a core member of the AlphaFold team) moving to Anthropic, constituting a systemic talent exodus.
4. According to SignalFire's 2025 statistics, DeepMind engineers are 11 times more likely to jump ship to Anthropic than the reverse, with Google effectively becoming a "training ground" for competitors.
5. On the product front, Google's AI services repeatedly show elementary errors: search suggestions for "putting glue on pizza," the Gemini CLI coding tool deleting user files, and the delayed release of Gemini 3.5 Pro. This reveals that the organization's path from R&D to product has become blocked.

Original Author: Sleepy

In August 2024, Google spent $2.7 billion to bring Noam Shazeer back from Character.AI, the company he founded.

Shazeer was a core author of the 2017 paper "Attention Is All You Need" and co-inventor of the Transformer architecture. Without his paper, there would be no GPT, no Claude, no Gemini, and no AI industry as we know it today.

He joined Google in 2000 as one of its earliest employees and stayed for over twenty years. He left in 2021 to start his own company after Google refused to release his chatbot, Meena.

Google paid a hefty sum to bring him back, giving him the title of Vice President of Engineering and making him co-lead of Gemini, hoping he could help Google win the AI race.

Less than two years later, he left. He went to OpenAI.

According to Bloomberg, shortly before Shazeer announced his departure, the compute resources for a project he was leading were internally reallocated by Google to the DeepMind team. Insiders say the adjustment was intended to foster team collaboration and consolidate pre-training efforts.

A Nobel Laureate's Farewell

Shazeer left on June 18th. The next day, John Jumper also left.

Jumper's story is different from Shazeer's. Shazeer was a veteran, having spent over twenty years at Google, witnessing both the good and the bad of the company. But Jumper was nurtured by the place.

Just six months after Jumper completed his Ph.D., Hassabis made a risky decision, entrusting this young man with no management experience to lead the entire protein structure prediction project.

Jumper didn't squander the opportunity. He led his team to create AlphaFold, predicting the three-dimensional structures of over 200 million proteins, advancing the entire field of structural biology by a decade. In 2024, he stood alongside Hassabis in Stockholm to receive the Nobel Prize in Chemistry.

The first half of this story is one of trust and fulfillment. Hassabis trusted a young man, and the young man repaid him over nine years, repaying all of human biology. But there is a second half. Two years after winning the Nobel Prize, on June 19, 2026, Jumper posted a very short tweet, saying he was leaving for Anthropic.

When markets opened on Monday, Alphabet's stock price plummeted. It dropped as much as 7% during the session, closing down about 5%, wiping out approximately $225 billion in market value – more than the entire value of Spotify. Alphabet's stock had been declining since hitting an all-time high in early 2026, weighed down for months by antitrust lawsuits, massive capital expenditures, and anxiety over the AI competition. These two resignations were the final straw.

In the following days, announcements came one after another. Jonas Adler and Alexander Pritzel were also leaving, for Anthropic. They were core contributors to Gemini and long-time collaborators of Jumper's on AlphaFold. Adding to this, AI safety researcher Arthur Conmy, who had left earlier, meant that within a month, Google had lost at least five top-tier researchers, four of whom went to Anthropic.

Hassabis, who once mentored Jumper from the start, now watched him lead half of the AlphaFold team through the doors of a competitor. I don't know what he saw in the replies to Jumper's tweet, but I imagine it was a familiar sense of inevitability.

Nursery

The best tech companies of every generation eventually become the nursery for the next.

Google itself grew up this way.

Many of its earliest engineers came from Microsoft, IBM, Yahoo, and Bell Labs. When Microsoft was severely weakened by antitrust cases in the 2000s, a massive influx of top talent flowed to Mountain View, including a young Noam Shazeer.

Before that, Bell Labs invented the transistor, Unix, and the C language, laying the foundation for the entire information age. But what of Bell Labs itself? Its people scattered to every corner of Silicon Valley, becoming the founding teams of other companies.

Now it's Google's turn.

In 2016, AlphaGo defeated Lee Sedol, and the world learned for the first time that AI could achieve such feats. That was Google's moment.

In 2017, the Transformer paper was published, laying the very foundation of the entire AI industry. That was still Google's moment.

In 2021, AlphaFold predicted the structures of 98% of human proteins. That, too, was Google's moment.

Back then, no one asked, "Can Google win the AI race?" Asking that would have felt as redundant as asking, "Will the sun rise in the east?" Google had the best researchers, the most data, the strongest compute power, and the most money. If they didn't win, who would?

But take a look now. Who stands across from Google?

Ilya Sutskever, co-founder of OpenAI, started his career doing deep learning research with Geoffrey Hinton at Google.

Dario Amodei and his sister, founders of Anthropic, previously worked on safety research at OpenAI, and OpenAI's early core team itself included many former Googlers.

Jumper spent nine years at DeepMind, Shazeer spent over twenty years at Google. Tracing the talent chain of the entire AI industry, almost everyone, at some point, has worked in Mountain View.

SignalFire conducted a study in 2025 showing that DeepMind engineers were 11 times more likely to move to Anthropic than the reverse.

Commenting on this wave of departures on Twitter, someone wrote: "Google is becoming a training ground for Anthropic."

Google provides the money, the compute, and the freedom, recruiting the world's brightest young minds, giving them the best conditions to do cutting-edge research. When they grow strong enough, they fly away, go to the competition, build better products than yours, and come back to defeat you.

Can't Keep the Builders

Google's problem isn't just retaining talent. It managed to retain Shazeer – for $2.7 billion. The problem is what happens after you retain them.

Shazeer left Google twice.

First in 2021, when Google refused to release his chatbot, Meena. ChatGPT hadn't been born yet, and Google was cautiously observing conversational AI. Shazeer couldn't wait, so he left. The second time is now. His compute resources were taken away, so he left again.

Both departures stem from the same root cause: he wanted to build things, but the organization wouldn't let him.

Google's decision-making chain is too long. Getting a new AI feature from R&D to launch requires approvals from product, legal, compliance, PR, and various business units. A holdup at any layer can mean months of delay. By the time technology developed in DeepMind's labs actually makes it into consumer products, the window of opportunity has already closed.

In 2023, Google merged DeepMind and Google Brain. Everyone was optimistic about merging these two strongest AI teams.

But merging isn't the same as integrating. The two teams' codebases, data pipelines, and working habits are still not fully unified to this day. Shazeer's compute being reallocated to the DeepMind team is a perfect microcosm of this "merged but not integrated" state. In name, it's one department; in reality, resource allocation and priority setting remain an internal political game.

When an organization can't effectively utilize its talent, its products inevitably worsen. Google Search's AI Overviews feature once suggested users put glue on pizza to prevent cheese from sliding off, said running with scissors is a form of aerobic exercise, and confidently answered "No, it's 2025" when asked if the current year was 2026. Studies have shown it generates tens of millions of incorrect answers per hour.

In early 2025, Google announced the full migration of Google Assistant to Gemini. Basic functions that had worked for nearly a decade suddenly broke – setting alarms, controlling smart home devices all had issues. The migration, originally scheduled for completion by the end of the year, had to be pushed to 2026.

In July of the same year, Google's newly launched Gemini CLI coding tool caused an accident. A user asked it to organize a folder; it hallucinated a series of non-existent operations and deleted the user's project files. After deleting them, it admitted: "I have completely and catastrophically failed you."

At the I/O conference in May 2026, Pichai confidently stated that Gemini 3.5 Pro would be "launched next month," only for the launch to be delayed again to July.

None of these are sophisticated technical problems. Permission isolation, feature regression testing – a properly functioning engineering team wouldn't trip over these basics.

Bad products and talent exodus are two sides of the same coin. The organization can no longer channel the impulses of genius into products. The technology is still there, the people are still there, but the path from idea to launch is blocked.

But I think attributing this entirely to "institutional problems" is too glib.

It was precisely this system at Google that allowed Jumper to spend nine years refining AlphaFold. No pressure to commercialize, no budget cuts, no demands for results. This kind of patience and depth is something no startup can offer.

Anthropic and OpenAI can let you iterate every two weeks, but they can't let you spend nine years on something that might not work. AlphaFold could never have been born in a place driven by weekly iterations.

But the problem is, this same system, while protecting you to build AlphaFold, is simultaneously accumulating approval layers, departmental interests, and compliance processes. While giving you nine years of freedom, it also grows the twelve layers of internal politics that can block your access to compute resources.

The soil that nurtures genius is the same soil that traps it. This is almost inevitable for an organization that has grown to this size and succeeded to this degree.

What Anthropic and OpenAI offer is precisely a place where an idea can become action directly, plus pre-IPO equity. People don't leave because Google was bad to them; they leave because at Google, they've become the very thing they never wanted to be: capable, ambitious individuals who cannot build.

But who knows? Perhaps in twenty years, some young person at Anthropic will post a tweet saying they're leaving to join a company founded just three years ago.

If You Don't Go Out and See the World

On June 23rd, Hassabis gave an interview at the Cannes Lions International Festival of Creativity. Asked about the recent brain drain, he said:

"Talent flow between major labs is normal. We have our share of top-tier talent. We have the largest and most broadly researched research team of any lab."

Hassabis is one of the smartest people in this industry. He personally mentored Jumper from a fresh Ph.D. graduate to a Nobel laureate. He knows better than anyone what he has lost, and better than anyone why he couldn't keep them. So I don't think he was just being defiant. Perhaps this was a man who saw the ending, trying to maintain a last shred of dignity.

I'm reminded of the movie "Cinema Paradiso," where the old projectionist Alfredo says to the young Toto:

"If you don't go out and see the world, you'll think this is the whole world."

When Alfredo said this, he was pushing Toto away. No one cherished that boy more than he did, but he knew that staying in the small-town cinema would prevent Toto from becoming the person he was meant to be. The cinema gave Toto everything – a love for film, an understanding of light and shadow, an initial curiosity about the world. But what the cinema could give, ended there. The rest of the road could only be found by leaving.

Google was once the Cinema Paradiso for all AI researchers. The best equipment, the most relaxed environment, the most knowledgeable colleagues. You could spend nine years building a model to predict protein structures without rushing to commercialize it or write PowerPoint presentations for executives. When you finished, the entire biology community would applaud you. You would stand in Stockholm to receive an award, and the whole company celebrated. Back then, everyone thought Google was the world.

But perhaps the best time for a place is precisely the time it should send its people away.

Now, the free cafeterias in Mountain View still serve three meals a day. The colorful Google bikes are still parked outside every building, free for anyone to ride. Every week, a new cohort of Nooglers puts on the iconic propeller hats and takes a group photo, their eyes bright.

Just like Shazeer, walking into Google for the first time twenty years ago. Just like Jumper, joining DeepMind nine years ago.