news

한어Русский языкEnglishFrançaisIndonesianSanskrit日本語DeutschPortuguêsΕλληνικάespañolItalianoSuomalainenLatina

Zhidongxi (public account:zhidxcom）
authorvanilla
editHeart

Zhidongxi reported on August 6th that just now,John Schulman, co-founder of OpenAI and head of post-training of ChatGPTAnnounced his resignation on social media XJumped to OpenAI competitor AnthropicEngage in technical research.

▲John Schulman announced his move to Anthropic

In addition, according to The Information, anotherGreg Brockman, co-founder and president of OpenAIExtending vacation time.Peter Deng, Product Director at OpenAIBrockman responded by saying that he wouldVacation until the end of the year, “We still need to build safe AGI”.

▲Brockman responded to the revelation of his resignation

So far, OpenAI’s11 Co-foundersOnly CEO Sam Altman, Wojciech Zaremba, head of the OpenAI language and code generation team, and President Greg Brockman, who is on a long-term leave, remain.three people。

1. "The mission is far from complete", and the holiday will last until the end of the year

Greg Brockman, who is responsible for turning breakthroughs from AI researchers into big models and products like ChatGPT, was also a key ally of Altman and played a key role in Altman's return to the company. He told employees he planned to return after an extended leave, according to The Information.

After being exposed for his resignation, he responded on social media X, saying that heWill be on vacation until the end of the year"This is the first time I can relax since I co-founded OpenAI nine years ago. The mission is far from complete,"We still need to build safe AGI。”

▲Brockman responds to resignation rumors

But netizens did not seem to buy Brockman's response. Someone mentioned Andrei Karpathy, who once left after a long vacation, "We know what happens after a long vacation... ahem."

▲ Netizens commented on Brockman

Another netizen compared it to Musk: "Imagine Elon telling everyone: I'm going to take a 4-month vacation. The world will really end."

▲ Netizens commented on Brockman

2. Join OpenAI competitors, "hope to further focus on AI alignment"

John Schulman leads a process called post-training for improving the large language models that power ChatGPT and other products. He also recently took over the remainder of the Super Alignment safety team, which is focused on preventing AI from causing societal harm.

Schulman announced his resignation on the social platform X. He said he hopesFurther focus on AI alignment, returned to actual technical work, and decidedPursuing this goal at AnthropicHe also stressed thatLeaving OpenAI was not due to lack of support for alignment research。

▲John Schulman officially announced his resignation

The following is a complete compilation of John Schulman’s official announcement:

I shared the following notes with my colleagues at OpenAI today:

I have made the difficult decision to leave OpenAI. This choice stems from my desire to further my focus on AI alignment and to start a new chapter in my career, returning to actual technical work. I have decided to pursue this goal at Anthropic, where I believe I can gain new perspectives and conduct research with people who are deeply engaged in the topics that interest me most. To be clear, I am not leaving because of a lack of support for alignment research at OpenAI. On the contrary, company leadership has been very committed to investing in this area. My decision is a personal one, based on where I want to focus my efforts in the next phase of my career.

I joined OpenAI as a member of the founding team nearly nine years ago after grad school. Aside from internships, this is the first and only company I’ve ever worked for, and it’s been incredibly fun. I’m grateful to Sam and Greg for recruiting me from the beginning, and to Mira and Bob for having the confidence in me to take on huge opportunities and help me succeed at every challenge. I’m proud of what we’ve accomplished together at OpenAI: we’ve built an unusual, unprecedented company with a mission to do good.

I believe that OpenAI and the team I am a part of will continue to thrive even without me. The post-training work is well underway with a group of great people. I get a lot of credit from ChatGPT. It's been great to see Barret, along with Liam, Luke, and others, build the team into the very capable group it is today. I'm excited to see the coordination team come together and work on some promising projects. Under the leadership of Mia, Boaz, and others, I'm confident the team will be very capable.

I am grateful for the opportunity to be a part of such an important moment in history and proud of what we have accomplished together. Even if I were employed elsewhere, I would still support all of you.

After Schulman officially announced his resignation, Altman praised Schulman repeatedly in the comment section, calling him a talented researcher who has deep thoughts on products and society. He also recalled the scene of their first meeting: In 2015, Altman and Schulman chatted for 15 minutes in a cafe in Berkeley, which became a big part of OpenAI's initial strategy.

▲OpenAI CEO Sam Altman’s comment on Schulman’s resignation blog post

However, netizens did not buy into Altman’s “affectionate confession” because the capitalization of “OpenAI” in previous tweets written by Altman himself had always been irregular, but this time it was completely correct.

▲ Netizens joked that the comments were not written by Altman himself

Some netizens also joked that they have entered a cycle of "waking up every day to see the OpenAI leader resign."

▲ Netizens joked about the frequent resignations of OpenAI executives

Jan Leike, the former head of security and super alignment at OpenAI who announced his joining Anthropic in May this year, also commented on the tweet: “Very happy to work together again.”

In his new team at Anthropic, Leike focuses on scalable supervision, weak-to-strong generalization, and automatic alignment research.

▲Comments from Jan Leike, former head of security and super alignment at OpenAI

Leopold Aschenbrenner, a former OpenAI employee, congratulated John Schulman on his departure in a tweet and said he was looking forward to his new research. Aschenbrenner also worked in OpenAI's super alignment department before and was fired by OpenAI in April this year for leaking company secrets.

▲Comment by former OpenAI employee Leopold Aschenbrenner

3. Executives left one after another to start their own businesses, leaving only 3 of the 11 co-founders

The company has been in a state of leadership instability since Altman was fired and rehired in November. The major personnel changes come at a time when OpenAI is experiencing strong growth and is facing significant losses.

OpenAI just lost a co-founder in May.Former Chief Scientist Ilya SutskeverHe officially announced his resignation and announced in June the establishment of a new company called "Secure Super Intelligence", which focuses on pursuing AI capabilities through revolutionary engineering and scientific breakthroughs while prioritizing safety.

anotherAndrej Karpathy, Co-FounderHe resigned in February and announced last month that he had founded an AI+education company, Eureka Labs, which will serve as a new type of "AI native school" to tutor students.

So far, OpenAI’s11 Co-foundersOnly CEO Sam Altman, Wojciech Zaremba, head of the OpenAI language and code generation team, and President Greg Brockman, who is on a long-term leave, remain.three people。

OpenAI also recently hired its first chief financial officer and chief product officer and moved another safety chief, Aleksander Madry, into other positions.

According to foreign media reports, OpenAI may transform from a non-profit company to a public welfare company, similar to Anthropic, which is currently training its next flagship model. OpenAI also recently launched a search engine and is developing an AI Agent that can automatically perform computer tasks for customers.

Conclusion: Top AI startups caught in executive resignation storm

Star AI startups are in turmoil. OpenAI is not an isolated case. Last year, Stability AI, which is famous for its graphic design, was also caught in a storm of executive resignations, and then experienced the CEO's resignation and the company's attempt to sell itself. This year, many AI startups that are at the top of their fields have sold themselves to large companies, such as Character.AI to Google, Adept AI to Amazon, and Inflection AI to Microsoft.

This is a microcosm of the Silicon Valley big model industry entering a period of reshuffle. The same is true for the top unicorns. How more AI startups with less fame and financing can survive and make profits has become the focus of the industry.