OpenAI Continues to Bring Talent to the Community, 19 Former Members Have Started Their Own Startups
PUBLISHED
2024-01-23
The OpenAI “mafia” appears to be well entrenched, with many former members leaving to start their own companies based on their OpenAI “gold plating” experience.
A few years after the AI company was founded and sparked the cultural zeitgeist with ChatGPT, some of its employees have begun to capitalize on the AI hype by starting their own AI startups.
Let’s meet the members of the OpenAI mafia, which includes Matt Krisiloff, one of the founding members of OpenAI, Jeff Arnold, the former head of operations, and a number of research and technical scientists (who are now at the helm of their own startups). Also included are the founding teams of Anthropic and Covariant AI, who all worked at OpenAI before founding those companies.
At one point, former PayPal colleagues, including Elon Musk, Reid Hoffman, Keith Rabois, and Peter Thiel, published a “Goodfellas” poster in Fortune magazine. Goodfellas” style photo in Fortune magazine and became known as the “PayPal Mafia”. Now, other big tech companies like Facebook and Oracle have their own mafias, along with younger mafia groups like Square, Stripe, and Instacart.
Top venture capital firms, including Andreessen Horowitz (A16z), Sequoia Capital, Index Ventures, Khosla Ventures, and Y Combinator, have backed startups that are members of the OpenAI mafia. In total, OpenAI Mafia members have raised nearly $8 billion in funding from investors, according to PitchBook and the founders themselves.
Here’s a list of the 19 OpenAI Mafia members and the startups they’re running now. The list is organized alphabetically by startup name.
1.Anthropic co-founders Dario Amodei, Daniela Amodei, Tom Brown, Jack Clark, Jared Kaplan, and Sam McCandlish
Dario Amodei, Jack Clark, and Daniela Amodei
Total funding: $7.25 billion
Number of Employees: Currently 300 employees according to PitchBook
Notable investors: Google, Amazon, Menlo Ventures
Anthropic was founded in 2021 by a group of researchers from within OpenAI who bonded over their belief in AI’s potential for good and evil. Since then, the company has received billions of dollars in funding from Google and Amazon, in what some have called “an AI arms race”.
From its inception, the company has marketed itself as a big model with security in its DNA. CEO Dario Amodei, a former Google Brain researcher with a Ph.D. in computational neuroscience, has been writing about the catastrophic potential of AI since 2016. He and Anthropic’s other co-founders, including former Bloomberg tech reporter Jack Clark, see that AI is going to grow exponentially, and they believe that AI companies need to start developing a set of values to bind these powerful programs.
Amodei said of himself and his co-founders at a Fortune conference last year, “We really trust each other and want to work together, so we started our company with that in mind.”
Anthropic is a public benefit corporation with an independent board of directors, which over time will control the hiring and firing of the company’s leadership.
Amodei’s sister, Daniela Amodei, who is the company’s president and previously oversaw OpenAI’s policy and security team, said Anthropic’s safety-first policy is one of its key differentiators.
Last year, Anthropic released a 22-page document laying out what it called a “responsible scaling policy,” or a plan to prevent its technology from hastening the end of humanity. The policy is reportedly overseen by Anthropic co-founder and theoretical physicist Sam McCandlish, who formed a team during his time at OpenAI to study the scaling laws of machine learning and paved the way for GPT-3.
At the heart of Anthropic’s pitch to enterprise customers is so-called “constitutional AI,” through which the creators of language models give them a conscience-a set of principles designed to prevent the misuse of technology. Constitutional AI is in part the brainchild of two other OpenAI alumni and Anthropic co-founders, Tom Brown and Jared Kaplan. Brown, a former Google Brain researcher, and Kaplan, a former physics professor at Johns Hopkins University, consulted for OpenAI before leaving to found Anthropic.
Both Kaplan and Brown were involved in Anthropic’s “red teaming” of the company’s flagship language model, Claude, to explore possibilities for abuse. This included an effort last year to create a version of Claude that could lie. Kaplan, speaking at a Bloomberg conference last October, said he thinks AGI – a version of AI powerful enough to disrupt society – could be realized in just five to 10 years.
“I’m worried about it, and I think regulators should be too,” Kaplan said at the conference.
2.Matt Krisiloff, co-founder and CEO of Conception
Matt Krisiloff, Co-Founder and CEO, Conception
Total funding: $40 million, according to the company
Notable investors: Sam Altman, Laura Deming, Jaan Tallinn
Number of employees: 43
Previous role at OpenAI: Founding member
As a member of OpenAI’s OG team, Matt Krisiloff initially led the startup’s operations in 2014 and 2015 before moving to Y Combinator to lead the research arm of the gas pedal program. In 2018, he founded Conception, a health tech startup that grows human eggs using stem cells to fight infertility.
Krisiloff graduated from the University of Chicago in 2014 and in 2021 founded SciFounders, an organization that funds scientists and helps them run their own companies.
3.Covariant co-founders Pieter Abbeel, Peter Chen and Rocky Duan
From left, Covariant’s founding team: Peter Chen, CEO; Pieter Abbeel, President and Chief Scientist; Rocky Duan, CTO; and Tianhao Zhang, Research Scientist.Total Financing: $222 million
Notable Investors: Index, Ventures, Industry Ventures, Temasek Holdings
Number of Employees: 200
Covariant aims to be a universal operating system for AI robots. Its model, called the Covariant Brain, helps robots perform previously challenging tasks, such as folding clothes or picking and packing in a warehouse.
The technology powering Covariant Brain stems from research conducted by co-founders Peter Chen, Rocky Duan, and Tianhao Zhang at the Berkeley Artificial Intelligence Lab in 2016. Another Covariant founder, Pieter Abbeel, was the director of Berkeley’s Robotics Learning Lab at the time and served as their PhD advisor.
They worked for OpenAI for several years before ChatGPT became a household name, and they have made great strides in the field of AI-powered robotics.
Chen told Forbes in 2023, “OpenAI brings together an ambitious and talented group of AI scientists and researchers who think big and push boundaries.” When Covariant was founded in 2017, it planned to commercialize its robotics for specific enterprise use cases. This was at odds with OpenAI’s pure research approach at the time.
Now, robots running Covariant Brain are working in warehouses and fulfillment centers around the world.
4. Tim Shi, co-founder and CTO, Cresta
Tim Shi, Co-Founder and CTO, Cresta
Total Funding: $151 million, according to the company
Tim Shi spent a year at OpenAI working on a variety of projects, including the open-domain platforms World of Bits and OpenAI Universe, as well as AI software for training and measuring AI on games and websites. He left the startup in 2017 to co-found Cresta, an AI coach that improves employee productivity.
Shi says his experience at OpenAI made him appreciate working on a team with so many talented people, and the experience has helped him manage talent at his own startup.
He said, ” OpenAI taught me the value of a high talent density environment and guided us to keep our hiring standards high and hire only the best talent.” He added that some former employees have started their own generative AI companies in the past year .
Before joining OpenAI, Shi was a software engineer at file-sharing company Dropbox, specializing in machine learning. He graduated from Tsinghua University in Beijing and now lives in San Francisco.
5.Jonas Schneider, Founder and Chief Executive Officer, Daedalus Jonas Schneider, Founder and CEO, Daedalus
Total Financing: $17.5 million
Notable Investors: Khosla Ventures, Y Combinator, LEA Partners
Number of Employees: 50
Previous position at OpenAI: Head of Technology
After earning a degree in computer science from the Karlsruhe Institute of Technology in Germany, Jonas Schneider took his first job at OpenAI in 2016 as a technical director. During his three years at OpenAI, he co-founded and led software engineering for the OpenAI robotics team.
With that experience, Schneider took the step to launch his own robotics startup designed to help factories and their production robots become more efficient. Daedalus is part of Y Combinator’s Winter 2020 program and, according to company representatives, has raised $17.5 million in funding from firms such as Khosla Ventures and LEA Partners. Daedalus is part of Y Combinator’s Winter 2020 program, and according to company representatives, the company has raised $17.5 million from companies like Khosla Ventures and LEA Partners. The company uses its AI robotics to produce custom prototypes and part families for factories around the world.
Schneider, who is the founder and CEO, said the gap Daedalus is filling is the need to constantly reprogram production robots for new tasks or objects. “Currently, programming them limits their use to mass-produced products, but our software-driven factory makes high-precision manufacturing scalable and increases the efficiency of mass production into high-mix manufacturing.”
6. Gantry co-founders Josh Tobin and Vicki Cheung
Josh Tobin and Vicki Cheung, Co-Founders, Gantry
Total Funding: $28.3 million
Notable Investors: Coatue, Index Ventures, OpenAI’s Greg Brockman
Number of Employees: 25
Previous roles at OpenAI: Tobin was a research scientist and Cheung was a founding engineer and infrastructure lead.
Josh Tobin, a former OpenAI research scientist, taught a deep learning course at UC Berkeley in 2019 with OpenAI infrastructure lead Vicki Cheung when the two realized the problems associated with building support infrastructure for AI tools.
Cheung was also a founding engineer at Duolingo and a senior engineer at Lyft, while Tobin holds a Ph.D. and a doctorate in computer science from Berkeley.
In 2020, the pair launched Gantry, which officially debuted in 2022. The AI startup works in the field of machine learning operations (MLOps), enabling teams to train their AI systems and evaluate which data to use when retraining. This allows companies to deploy AI systems more effectively to better engage with customers.
According to PitchBook, Gantry received $28.3 million in funding from OpenAI president and co-founder Greg Brockman and Coatue, among others, to double down on its customer acquisition efforts and expand its workforce, TechCrunch reported.
7.Margaret Jennings, Co-Founder and Vice President of Product, Kindo
Margaret Jennings, Co-Founder and VP of Product, Kindo
Total funding: $7 million, according to the company
Number of employees: 13
Previous roles at OpenAI: Application team member
Margaret Jennings worked at OpenAI for six months (from August 2022 to January 2023), during which time she was part of the applications team, which works to bring the startup’s technology to consumers. During her tenure, Jennings worked on a project to provide ChatGPT to Morgan Stanley financial advisors, which launched in late 2023, according to CNBC.
Jennings’ experience at OpenAI taught her “the importance of broader accessibility” – especially when translating research goals into product design.
Today, Jennings is working for Kindo, an early-stage startup that builds security and product layers for large-scale language modeling applications for the enterprise. Since its founding in February 2023, the startup has raised $7 million in venture capital.
Before Kindo and OpenAI, Jennings served as vice president of product and artificial intelligence at Halodoc ID, a digital health startup, and as global head of Google. She graduated from Bard College in 2013 and earned a master’s degree in computer science from University College London, where she is currently pursuing a PhD in human-centered general artificial intelligence (AGI).
8. Jeff Arnold, Founder and Chief Operating Officer, Pilot
Jeff Arnold, Founder and COO of Pilot Jeff Arnold
Total Funding: $174 million, according to the company
Notable investors: Sequoia Capital, Index Ventures, Bezos Expeditions
Number of employees: 250
Previous position at OpenAI: Head of Operations
Jeff Arnold is a serial startup founder who spent five months at OpenAI in 2016 running operations before launching his third startup, Pilot – a fintech startup that helps companies with their accounting needs. He notes that when Pilot launched in 2017, OpenAI was one of the company’s first customers, and Pilot was also a customer of OpenAI.
“In Silicon Valley at its best, where tech people come together and work together to do things beyond what people expect, I can’t think of a better example than OpenAI.” He told Business Insider.
Before Pilot and OpenAI, Arnold co-founded enterprise chat app Zulip (acquired by DropBox in 2014) and enterprise software startup Ksplice (acquired by Oracle in 2011). He earned his bachelor’s and master’s degrees from MIT.
9. Aravind Srinivas, Co-Founder and CEO, Perplexity
Aravind Srinivas, Co-Founder and CEO, Perplexity
Total Financing: $102.3 million
Notable Investors: IVP, Jeff Bezos, Sequoia Capital, Nvidia
Number of Employees: 34
Previous role at OpenAI: Research Scientist
Based in San Francisco, Perplexity was founded in 2022 by Aravind Srinivas, a former research scientist at OpenAI, to become the AI search engine of choice.
The startup uses a range of big models, from OpenAI’s own ChatGPT to Meta’s Llama, to operate its chatbot-like search engine. Users can ask questions and make dialog-like queries using the AI tool, which will provide answers primarily in the form of citations and sources.
Srinivas has managed to win the backing of several high-profile investors, including Sequoia Capital, Databricks, Github founder Nat Friedman, and former YouTube CEO Susan Wojcicki.
He says his advice to founders is to “go all in on the idea. When you’re starting a company for the first time, there are a lot of opportunities, but it’s easy to fall into the trap of trying to do too many things. Pick an area of expertise and focus your efforts and resources to be the best you can be.”
“It’s risky, but in my opinion, if your idea isn’t worth the risk, then it’s not worth starting that company.” He added.
10.Shariq Hashme, co-founder and CEO, Prosper Robotics
Shariq Hashme, Co-Founder and CEO, Prosper Robotics
Total Funding: Undisclosed
Notable Investors: Scale CEO Alexandr Wang, Anthropic co-founder Ben Mann, Notion co-founder Simon Last
Number of Employees: 11
Previous roles at OpenAI: Technical Staff
Shariq Hashmi worked at OpenAI for 9 months as part of the technical team before moving to AI newcomer Scale AI later that year.
With experience in AI and a degree in electrical engineering, Hashme launched Prosper, a London-based robotics startup, in February 2021, to build a robot capable of performing basic household tasks, from doing laundry and dishes to cleaning the house.
The product is not yet ready for the market, Hashme said. The team is working to create a robot that can perform a variety of tasks, but is still affordable enough to attract customers.
While Prosper has not disclosed the total amount of money it has raised, the startup has received backing from investors such as Anthropic co-founder Ben Mann and Notion co-founder Simon Last, Hashme said.
His advice to current employees of the fast-growing startup stems from his favorite movie, Cloud Atlas: the Apocalypse of Sonmi 451. “The essence of our immortal lives lies in the consequences of our words and actions, which are always pushing themselves. Our lives are not our own. From the womb to the grave, we are connected to others, past and present, and through every offense and every kindness, we give birth to our future.”
11. Ishant Singh, Founder of Invisible Startups
Ishant Singh, founder of a stealth startup
Total Funding: Unknown
Previous role at OpenAI: AI Security Team
Ishant Singh spent four years in product in New Delhi before heading to Arizona State University to complete his MBA. During this time, he spent some time on the product team at Amazon working on Alexa before moving to OpenAI’s Trust and Safety team in 2021, where his job was to ensure that any output from an AI model was safe and that any output was safe. end users will use it in a way that complies with OpenAI’s security guidelines.
In June 2023, he began work on his startup, which is currently in secret mode but has completed the development of a minimum viable product and is currently raising seed funding.
The product is a ChatGPT-like AI platform that connects to every data source in the company, such as Google Drive, and becomes a context-aware ChatGPT. it’s model-agnostic, so it’s not just limited to ChatGPT, but it also disseminates guidance to employees based on their level of seniority and level of access to that information.
Singh says, “I have a former OpenAI tag on my LinkedIn, which means that I have been approached by over 35 VC firms.” He adds that this is a common feat for any OpenAI alumni to raise money.
His advice to founders is to “learn from all the ups and downs of the workplace and always look for new opportunities in customer and product problems.”
