惯性聚合 高效追踪和阅读你感兴趣的博客、新闻、科技资讯
阅读原文 在惯性聚合中打开

推荐订阅源

钛媒体:引领未来商业与生活新知
钛媒体:引领未来商业与生活新知
U
Unit 42
GbyAI
GbyAI
M
MIT News - Artificial intelligence
美团技术团队
罗磊的独立博客
雷峰网
雷峰网
量子位
博客园 - 【当耐特】
Last Week in AI
Last Week in AI
D
Docker
小众软件
小众软件
S
SegmentFault 最新的问题
Blog — PlanetScale
Blog — PlanetScale
阮一峰的网络日志
阮一峰的网络日志
宝玉的分享
宝玉的分享
T
Tailwind CSS Blog
WordPress大学
WordPress大学
V
V2EX
博客园_首页
腾讯CDC
The Cloudflare Blog
A
About on SuperTechFans
Cyber Security Advisories - MS-ISAC
Cyber Security Advisories - MS-ISAC

DEV Community

Authentication Security Deep Dive: From Brute Force to Salted Hashing (With Java Examples) Why AI Systems Don’t Fail — They Drift Spilling beans for how i learn for exam😁"Reinforcement Learning Cheat Sheet" I Replaced Chrome with Safari for AI Browser Automation. Here's What Broke (and What Finally Worked) How Python Borrows Other People's Work The $40 Architecture: Processing 1 Billion API Requests with 99.99% Uptime Vibe Coding: A Workflow Guide (From Zero to SaaS) Most webhook security guides protect the wrong side. The scary part is delivery. Headless CMS for TanStack Start: Build a Blog with Cosmic EU Age Verification App "Hacked in 2 Minutes" — What Actually Happened Comfy Cloud’s delete function does not actually remove files Running AI Models on GPU Cloud Servers: A Beginner Guide Event-driven media intelligence with AWS Step Functions and Bedrock I scored 500 AI prompts across 8 quality dimensions — here's what broke How to Call Google Gemini API from Next.js (Free Tier, No Backend Needed) The Portal Protocol: Reclaiming Human Connection in the Age of AI How to Fix Your Team's Scattered Knowledge Problem With a Self-Hosted Forum Intro to tc Cloud Functors: A Graph-First Mental Model for the Modern Cloud Designing Multi-Tenant Backends With Both Ownership and Team Access I Built a Neumorphic CSS Library with 77+ Components — Here's What I Learned PostgreSQL Performance Optimization: Why Connection Pooling Is Critical at Scale Cómo construí un SaaS multi-rubro para gestionar expensas en Argentina con FastAPI + Vue 3 🚀 I Built an Ethical Hacking Scanner Tool – Open Source Project I Replaced /usage and /context in Claude Code With a Single Statusline A Pythonic Way to Handle Emails (IMAP/SMTP) with Auto-Discovery and AI-Ready Design I Collected 8.9 Million Polymarket Price Points — Here's What I Found About How Markets Really Move EcoTrack AI — Carbon Footprint Tracker & Dashboard Everyone's Using AI. No One Agrees How. 5 self-hosted ebook managers worth trying in 2026 Building Your First AI Agent with LangChain: From Chatbot to Autonomous Assistant
Google Paid $2.7B to Keep Its Best AI Researcher. He Left...
Md Jamilur Rahman · 2026-06-21 · via DEV Community

Md Jamilur Rahman

In 2024, Google paid $2.7 billion to bring Noam Shazeer back. Shazeer invented the transformer architecture, the foundation behind GPT, Gemini, and Claude. Google said: "We can't lose this person."

June 2026. Shazeer is leaving Google for OpenAI.

This isn't one person's story. It's a pattern. Google DeepMind's best researchers are walking out the door to AI startups, one by one.


Who Left?

Noam Shazeer — The man who invented the transformer. Co-lead of Google's Gemini model. Google paid $2.7 billion to buy him back from Character.AI. Now joining OpenAI. (Reuters, CNBC, The Information)

John Jumper — Co-creator of AlphaFold. Won the 2024 Nobel Prize alongside Demis Hassabis for protein structure prediction. Spent nearly nine years at Google DeepMind. Now joining Anthropic. (CNBC, Business Insider, the-decoder.com)

Ruoming Pang — Meta hired him for $200 million to lead their Superintelligence team. He jumped to OpenAI after just 7 months. (observer.com)

4 Apple AI researchers — Including Apple's top Siri leader, all departed. (MacDailyNews)

9 of 12 xAI co-founders — Left Musk's AI lab. (tech-insider.org)


Why Are They Leaving?

One thing is clear. The problem isn't pay. Google pays well. The problem is autonomy.

The Los Angeles Times reported that Google researchers are quitting over a "compute crunch" — internal competition for GPU access that slows down research and frustrates top talent. (LA Times)

At startups? Start an experiment today. Fail tomorrow. Try again. Nobody's stopping you.

OpenAI and Anthropic have another card to play — equity. Google has RSUs, but startup equity can be worth far more. OpenAI has filed for a US IPO, targeting a valuation of up to $1 trillion. Early equity holders could see massive payouts. (Reuters, Cybernews)


What Is Google Doing?

Google isn't sitting still.

1. Throwing money at retention
They paid $2.7 billion to bring Shazeer back. Two years later, he left again. Money can't hold people.

2. Compute advantage
Google has massive GPU clusters. Startups don't. But the LA Times report suggests even Google's compute advantage is causing internal friction — too many teams competing for the same resources.

3. Product momentum
Despite the exodus, Google's AI products remain strong. Gemini 2.5 is competitive. DeepMind continues publishing breakthrough research. Accenture recently expanded its partnership with Google Cloud for enterprise AI using Gemini. (Axios)

The question is whether product momentum can outlast the talent drain.


Why This Matters

Think about it. Noam Shazeer invented the transformer. That's the foundation of the AI revolution. That person is now working at OpenAI.

John Jumper cracked protein folding. Anthropic just picked him up.

Google spent billions on their research. When those researchers mature, they leave for other companies. The question is whether this brain drain eventually shows up in Google's products — or whether Google's scale and compute advantage compensate for it.

Demis Hassabis is still there. But the people around him are thinning out.


The question is — which path are you on? Big corporate, scrappy startup, or building something of your own? 🙏


Sources: