OpenAI gangsters: Silicon Valley's new tech class

Hello, here is the technology outpost, today I will continue to share with you the new insights related to the field of artificial intelligence.

Since its birth, Silicon Valley has been deeply buried with entrepreneurial blood, and has always had a strong "gangster" culture, and the leader of each round of scientific and technological innovation leads new trends and cultivates a large number of innovative talents.

The earliest Fairchild company is the cradle of the rise of chip companies in Silicon Valley, and the PayPal gangsters led by Musk and Peter Thiel have achieved a new round of glory in Silicon Valley.

Today, OpenAI is completely out of the circle by chatGPT, and the talents who have gone out from this company have long been all over Silicon Valley.

Just as every round of technological innovation will create new entrepreneurs and rich people, the current AI technology change has sown new seeds, the future leader of the Silicon Valley artificial intelligence industry, the next round of technology investors are likely to be born in this group of people.

The technology outpost will sort out with you today, and a few of the more important figures in it, see in advance what future celebrities may be!

1. OpenAI seeds talent in Silicon Valley

OpenAI has only 375 full-time employees, and former employees from the company are found in Silicon Valley's largest corporations and startups.

According to Linkedin's data, more than 20 former OpenAI employees work at Tesla and at least 16 at Google. In addition, popular companies such as Amazon, Meta and Microsoft have also recruited many former OpenAI employees. Another popular company to recruit former OpenAI employees is Anthropic (tech outposts have previously done research, click to jump to read), an artificial intelligence startup founded in 2021 by former OpenAI founders Dario Amode and Daniela Amodei, where 15 former OpenAI employees work according to Linkedin.

Below are profiles of 16 former OpenAI employees who are building the next generation of AI tools and contributing to the spread of machine learning expertise.

2.Pieter Abbeel

Abbeel joined OpenAI in June 2016 as Director of Research and left OpenAI about a year and a half later to found Covariant, a startup backed by Index Ventures and Amplify Partners whose vision is to provide intelligent robotics solutions for manufacturing, with the core technology of using deep deep learning technology to control robots to automate robots. Covariant's precision handling technology allows robots to grasp objects with precision, which can be used in logistics to perform sorting tasks. Abbeel is also a professor of artificial intelligence at the University of California, Berkeley, and much of his research has been in deep reinforcement learning, a learning method that teaches computers to perform tasks such as autopiloting and playing board games.

3.Daniela Amodei

Daniela Amodei has been with OpenAI for less than two and a half years, but has done a lot of credit during her tenure. She left Stripe in 2018 to join OpenAI, where she served as a project manager for OpenAI's natural language processing and music generation teams, and as OpenAI's vice president of human resources, responsible for recruitment and human resources.

According to her Linkedin profile, in May 2020, Amodei was named vice president of security and policy. Eight months later, she left OpenAI with her brother Dario Amodei to co-found Anthropic (tech outposts have been researched before, click to read).

4.Vicki Cheung

Cheung, a founding engineer at OpenAI and previously Duolingo, joined the company in late 2015 as the second employee after Greg Brockman, who is now OpenAI's president and chief technology officer.

As head of infrastructure for the startup, she manages OpenAI's lab and research resources, as well as technology development. Leaving OpenAI in 2018, Vicki Cheung joined Lyft. Later, she co-founded startup Gantry with another early OpenAI employee, Josh Tobin.

Vicki Cheung is currently the CEO of Gantry, a company dedicated to improving the safety of large-scale AI systems by researching areas as diverse as language understanding, computer vision, and robotics.

5.Ian Goodfellow

Goodfellow, one of the most influential figures in deep learning, joined OpenAI in 2016 as a research scientist and left after just over a year.

In 2014, as a research scientist at Google, Goodfellow helped create the Generative Adversarial Network, a method of training machine learning models that then produce raw images, videos, and speech. After leaving OpenAI in 2017, Goodfellow returned to Google for two years before moving to Apple in 2019 as head of machine learning for the Special Projects Group. Last May, Goodfellow resigned from Apple to protest Apple's policy of requiring employees to return to the office. Soon after, he joined Alphabet's Deepmind division as a research scientist.

6.Jesse He

After working as a researcher at Amazon Web Services and Microsoft, Han joined OpenAI in 2021 as a technologist at OpenAI while also pursuing a PhD at the University of Pittsburgh, according to Linkedin.

According to papers he published during his work at OpenAI, he built an AI model designed to solve challenging mathematical problems. Han left OpenAI last year and is now CEO of Multi Technologies, a company dedicated to using state-of-the-art technology to provide customers with the best applications and solutions, enabling people with little coding knowledge to create their own AI programs.

7.Ariel Herbert-Voss

Herbert-Voss dropped out of Harvard's computer science doctoral program in 2019 to join OpenAI.

During his research at Harvard University, he studied the vulnerabilities of AI systems with the aim of making them more secure. He co-authored a thousand-cited paper with colleagues at OpenAI at the time, including Daniel Ziegler and Dario Amodei, titled "Language Models Are Few-Shot Learners." The paper shows that if researchers train a language model on data with enough computing power, the model can perform a wide range of tasks and be controlled by a straightforward prompt, an idea that paves the way for products like ChatGPT, which are now popular.

Herbert-Voss serves as Vice President of Research at OpenAI, where he focuses on AI and machine learning research. He left OpenAI in April 2022 to start a cybersecurity startup, which is currently undisclosed.

8.Jacob Jackson

Jackson first joined OpenAI as an intern as an undergraduate at the University of Waterloo. Before graduating from college, he founded Tabnine, a tool that helps programmers complete tasks faster by providing them with auto-completion suggestions as they write code, greatly improving their development efficiency. The Israeli company Codota acquired the company in 2019 for $15 million.

Jackson returned to OpenAI after graduation before leaving again in 2020 to start working on his own visual search engine, founding Same Energy, a company focused on transforming the energy industry by leveraging artificial intelligence and sophisticated algorithms.

9.Diederik (Durk) Kingma

According to Kingma's personal website, Kingma was part of the OpenAI startup team in 2015. While working at OpenAI, he led an algorithm-focused team, including generative models like the original GPT and Glow that could generate images and edit images. He worked as a researcher at Alphabet's Deepmind in 2014 and 2015, generating work related to generative models. His 2014 paper, "Adam: A Simple and Effective Way to Optimize Arbitrary Functions," describes optimizer techniques that help improve the accuracy of machine learning models. Almost all generative models, including language models like GPT-3, are now using its optimizer.

In 2018, Kingma returned to Alphabet and joined Google Brain, focusing on generative models.

10.Matt Krisiloff

Krisiloff became part of the founding team of OpenAI shortly after being named research director at Y Combinator in 2015, left OpenAI in 2016 and left Y Combinator in 2018 to start Conception, a biotech startup that studies how to generate human eggs from stem cells for fertility.

The startup has raised $20 million from backers, including OpenAI CEO Sam Altman and Skype co-founder Jaan Tallinn.

11.David Luan

Luan joined OpenAI in 2017 as a director of the company and was later named vice president of engineering before leaving in 2020 to lead Google Brain's work on large-scale machine learning models. The following year, Luan left Google to found Adept.

Adept is a startup that develops artificial intelligence software to automate computer tasks. According to The Information, the two-year-old company is recently seeking a valuation of up to $1 billion, up from $250 million last year.

12.Vedant Misra

Misra joined OpenAI from HubSpot in 2019 and spent two years at HubSpot-owned Kemvi. Kemmi is an automated sales and marketing copywriting company.

At OpenAI, Misra led the team to build multimodal neural networks, which are artificial intelligence such as training on multiple types of data such as text and images, which can be used to analyze visual and auditory data for natural language processing, machine translation, machine learning, and more.

Misra left the company in 2021 to come to Google Research, where Misra focused on Minerva, an artificial intelligence model for quantitative reasoning, and a large language model similar to GPT-3.

13.Ashley Pilipiszyn

From 2018 to 2021, Pilipiszyn served as Technical Director of OpenAI, where she was responsible for maintaining OpenAI's developer ecosystem and leading the launch of products including the GPT-3 API and Dall-E.

Pilipiszyn left in 2021 to found Titan Space Technologies, a company dedicated to developing space technology, focusing on building space platforms and developing technologies that enable space exploration. The startup, which raised a $10.5 million seed round, announced last year that Hewlett-Packard Enterprise's space computers would use their technology on the International Space Station, PitchBook reported.

14.Jonas Schneider

Schneider joined OpenAI in 2016 as the technical lead of its robotics team, and according to Linkedin, he developed software for some projects, including a robotic arm that can manipulate the Rubik's Cube, but OpenAI disbanded its robotics team in 2020 to focus on generative AI models.

Schneider left OpenAI in 2019 to found Daedalus, a startup that designs factory robots that can be reprogrammed through simple programs to perform different production tasks. According to Pitchbook, Daedalus, incubated at Y Combinator, has raised more than $14 million from investors including Khosla Ventures.

15.Aravind Srinivas

Srinivas first joined OpenAI in 2018 as a research intern for three months while pursuing a PhD in computer science at the University of California, Berkeley. According to Linkedin, after graduation he returned to OpenAI in 2021 as a research scientist, focusing on language and generative models.

Srinivas left OpenAI in August 2021 to found Perplexity AI, a startup that develops search engines that let people ask questions through chatbots similar to ChatGPT. Perplexity raised seed funding in October, and according to people familiar with the matter, Perplexity's investors include Google's head of artificial intelligence, Jeff Dean, Meta's chief AI scientist Yann LeCun, angel investor Elad Gil, former Github CEO Nat Friedman, and former OpenAI research scientist Pieter Abbeel.

16.Kenneth Stanley

Stanley, an expert in neural networks and evolutionary computing, joined OpenAI in 2020 after more than a decade in academia and as head of research at Uber's autonomous driving division. Stanley joined after Uber acquired Geometric Intelligence, a machine learning startup he founded.

At OpenAI, Stanley leads a team focused on how to continually acquire new capabilities for algorithms, such as new scenarios for robot walking. He left in June 2022 to start a now-undisclosed company, about which Stanley declined to reveal.

17.Daniel Ziegler

Ziegler worked as an engineer at OpenAI from 2018 to 2021, according to data on Linkedin, and he left OpenAI in a similar role at Redwood Research, an AI security nonprofit.

Ziegler's research focuses on ensuring that AI systems operate in the way their creators intended, and he and his team are trying to solve problems including how adversarial training can be used to ensure that AI systems don't generate solutions that are bad for people.