Das Produkt ist noch nur ein Ordner, aber die Investoren jagen sich um, um Geld zu investieren: Nur weil sie "ex-OpenAI" ist?

[Introduction] Ten years ago, the term "Mafia" belonged to PayPal. Ten years later, it belongs to OpenAI. Those who left OpenAI didn't go far; instead, they are rebuilding another OpenAI outside. They invest in and endorse each other, weaving an invisible power network from algorithms to capital. The story of AI, seemingly about technology, is actually about the inheritance of power. When algorithms learn to imitate humans, humans are also using algorithms to replicate their own empires.

More than a decade ago, the most famous "Mafia" in Silicon Valley was PayPal.

That group of young employees who left - Elon Musk, Peter Thiel, Reid Hoffman... They rewrote the financial order with the Internet.

More than a decade later, this title has a successor.

Their headquarters is not in a bank but in a computing power cluster; their weapon is not capital but algorithms.

Dario Amodei, Mira Murati, Ilya Sutskever, William Fedus... These people who left OpenAI are building a new invisible network with investment, technology, and reputation.

They invest in and endorse each other, almost surrounding the entire artificial intelligence landscape from language models, automation systems to "AI scientists".

In Silicon Valley, they have a new name - "The OpenAI Mafia." A group that can define the future of intelligence without violence or bloodshed.

Leaving OpenAI, Worktrace AI Ignites the "Mafia" Legend

In the summer of 2025, a new company called Worktrace AI was quietly established in Silicon Valley.

It neither had a dazzling product launch nor an exaggerated promotional video. The official website only has a simple introduction:

Worktrace AI helps enterprises identify and automate repetitive tasks in employees' daily work.

It sounds like an ordinary enterprise automation company, but its founder has caught the attention of the entire AI circle - Angela Jiang.

Reports say that this startup is in talks for a $10 million seed - round financing, with a valuation of about $50 million.

What's even more eye - catching is the investment lineup behind her - it's almost a reunion of OpenAI's top management:

Former CTO Mira Murati, ChatGPT team leader Nick Turley, Chief Strategy Officer Jason Kwon, ChatGPT behavior team leader Joanne Jang, as well as institutions like OpenAI Startup Fund, 8VC, and Conviction.

Who is She? From Behind the Scenes of ChatGPT to the Founder of Worktrace

Angela Jiang is a Chinese product manager who graduated from the Wharton School of the University of Pennsylvania, majoring in a double degree in economics and computer science.

Before joining OpenAI, she worked in product positions at Google and Dropbox successively.

At OpenAI, she was the third product manager, participating in the core product coordination and design work before and after the launch of ChatGPT.

That was a crucial node when ChatGPT moved from the laboratory to the public. She was responsible for connecting engineering, strategy, and user experience, turning research results into applications that could reach hundreds of millions of people.

Later, she transferred to OpenAI's public policy team, researching the social impact and regulatory issues of artificial intelligence.

At the end of 2024, she left the company. Half a year later, she returned with Worktrace AI - this time, instead of making AI talk, she wants AI to "work".

If ChatGPT learned to imitate human language, what Worktrace AI aims to do is to make AI learn to imitate human labor.

This idea has become a new trend in Silicon Valley:

AI learning and replicating human tasks. - AI is learning and replicating human tasks.

From OpenAI to Worktrace: An Invisible Capital Chain

Angela Jiang is not an isolated case. In the past year, OpenAI has seen a wave of departures.

An invisible network composed of former executives, researchers, and product leaders is expanding.

Former CTO Mira Murati founded Thinking Machines Lab, completed a $2 billion seed - round financing, and the company is valued at $12 billion.

Former Chief Scientist Ilya Sutskever founded Safe Superintelligence Inc. (SSI), valued at about $32 billion

ChatGPT researcher William Fedus co - founded Periodic Labs and completed a $300 million financing in September.

These companies have different directions - some focus on AGI safety, and some target scientific research automation. But the investors, advisors, and founders behind them can almost all be connected on OpenAI's organizational chart.

The core connections, funds, and reputation that originally belonged to OpenAI are being rearranged and replicated outside the organization by these former employees, piecing together a new, invisible empire.

So, the media has given it a somewhat joking name: "The OpenAI Mafia."

Mafia - Style Expansion: Another Way to Replicate OpenAI

After PayPal, Silicon Valley is witnessing such a "family effect" again.

These entrepreneurs from OpenAI are re - binding each other with algorithms and capital, replicating a new AI power structure outside of OpenAI.

So, the term "The OpenAI Mafia" is no longer a joke but a mirror reflecting the reality of Silicon Valley:

They left OpenAI but are still jointly shaping the future of AI.

The Rise of Algorithmic Capitalism: Valuation First, Then Product

In the early 21st century, the "PayPal Mafia" was a legend - they were the "Avengers" of Silicon Valley.

Tesla, SpaceX, LinkedIn, YouTube, Yelp... Almost all the billion - dollar companies you can name were either founded by former PayPal employees or received their investments.

As history always repeats itself, we are witnessing a similar story now.

Just as the PayPal Mafia shaped the Web2 era, the OpenAI group is quietly reshaping the next big thing - the AI field. Some of the companies they founded have already exceeded the $10 billion valuation mark.

In the past year, a group of executives, researchers, and product leaders from OpenAI have left to start their own businesses - and they almost all have one thing in common:

The companies haven't launched products yet, but their valuations have soared to hundreds of millions or even billions of dollars.

In this wave of enthusiasm, capital no longer seems to care about "what they are doing" but focuses more on "who they used to be".

Angela Jiang's financing is not an isolated case.

In Silicon Valley in 2025, a new form of "algorithmic capitalism" is taking shape: there is no sign of a product, but the valuation has skyrocketed.

Capital no longer asks "what are you doing" but first looks at "where are you from".

In the eyes of capital, "being from OpenAI" seems to have become a new golden seal.

"OpenAI Bloodline" Becomes the Pass for Financing

In the past, startups had to rely on prototypes or user growth to convince investors. But in the AI industry, this path has been completely rewritten.

Worktrace AI has not launched a usable product yet, and there is only a simple introduction on the official website.

Nevertheless, it can still negotiate a $10 million seed - round financing with multiple funds and has a valuation of $50 million.

The reason is simple: its founder is the third product manager of OpenAI.

The same scenario is being repeated with Mira Murati, Ilya Sutskever, and William Fedus.

The Logic of Algorithmic Capital: Trust Replaces Technology

Behind this is a set of logic recompiled by capital. Resumes have become a credit endorsement, and identity has replaced product verification.

Investors believe that these people from OpenAI understand the boundaries of large - scale models better and are closer to defining the next - generation intelligence.

What they are betting on is not a specific technology but a trusted algorithmic authority.

The OpenAI Startup Fund frequently appears on these financing lists, and the mutual investment and support among former executives further strengthen this system.

Capital is circulating in a closed circle: people leave OpenAI, and money flows back from OpenAI.

So, the product is no longer the starting point of entrepreneurship but just the result - what is really being traded is trust and bloodline.

A group of OpenAI people are investing in each other.

In this new capital game, algorithms have become a belief, and identity has become currency.

AI entrepreneurship no longer starts from the laboratory but from connections - this is the new rule of "algorithmic capitalism".

Those Who Define Intelligence Can Define the Future

From PayPal to OpenAI, the protagonist of Silicon Valley's "Mafia" has changed. The former monopolized payments, and the latter is monopolizing intelligence.

They use algorithms instead of currency, identity instead of credit, and trust to build valuations. Behind the seemingly "decentralized" AI entrepreneurship wave, a new center is taking shape:

A power network connected by the same group of people, the same set of ideas, and the same resumes.

Angela Jiang makes AI learn to work, Mira Murati makes AI think, and Ilya Sutskever makes AI safe.

They seem to be going in different directions, but they are holding the future in their hands together.

Silicon Valley talks a lot about innovation, but when all new companies come from the same source, the boundary between "innovation" and "inheritance" begins to blur.

AI is not only imitating human labor but also replicating human power structures.

And the new "Mafia" teaches machines to understand the world and also teaches the world to obey machines.

References:  

https://www.linkedin.com/in/jiangangela/ 

https://angelajiang.carrd.co/# 

https://scholar.google.com/citations?user=gcS68oUAAAAJ&hl=en 

https://x.com/rowancheung/status/1627669926902173700 

https://algarch.com/blog/the-openai-mafia-why-ex-openai-is-the-new-golden-resume-line 

https://archive.ph/0bRQu#selection-1549.0-2123.142 

This article is from the WeChat official account "New Intelligence Yuan", edited by Qingqing, and published by 36Kr with permission.