Durk Kingma: OpenAI Co-Founder Joins Anthropic, Reinforcing AI Safety Focus

Why the OpenAI-to-Anthropic pipeline remains so strong

In a series of posts on X, Kingma revealed that he'll be working mostly remotely, from the Netherlands (where he's based), but didn't say which Anthropic team he'll be joining -- or leading. "Anthropic's approach to AI development resonates significantly with my own beliefs; looking forward to contributing to Anthropic's mission of developing powerful AI systems responsibly," Kingma wrote on X. "Can't wait to work with their talented team, including a number of great ex-colleagues from OpenAI and Google, and tackle the challenges ahead!" Kingma holds a PhD in machine learning from the University of Amsterdam and spent several years as a doctoral fellow at Google before joining OpenAI's founding team as a research scientist. After leading OpenAI's algorithms team, Kingma joined Google Brain (later merged with DeepMind) in 2018, where he focused on generative models for text, image, and video.

Yet another OpenAI co-founder joins Anthropic, this time Durk Kingma

Kingma left OpenAI in 2018 to work as a research scientist at Google DeepMind. Durk Kingma joins the OpenAI to Anthropic pipeline, becoming the latest co-founder to join the Claude chatbot maker. Kingma, who has been working as a research scientist at Google's Brain (which merged with DeepMind) since 2018, announced the move in a post on X. "Anthropic's approach to AI development resonates significantly with my own beliefs," he said. "Looking forward to contributing to Anthropic's mission of developing powerful AI systems responsibly." While he said he will be mostly working from the Netherlands, where he is based, Kingma did not give details about his new role. Kingma has a PhD in machine learning from the University of Amsterdam and worked as a doctoral fellow at Google before founding OpenAI with 10 others. However, as of now, only three original OpenAI co-founders, Sam Altman, Greg Brockman and Wojciech Zaremba remain at the company. After his stint at OpenAI ended, Kingma became an angel investor, investing in and advising AI start-ups, before returning to Google as a research scientist in large language models later that year. Kingma becomes the latest Anthropic hire with a past at OpenAI. John Schulman, also an OpenAI co-founder, joined the rival company in early August. In a post on X, Schulman said he left OpenAI to pursue "hands-on technical work". "I've decided to pursue this goal at Anthropic, where I believe I can gain new perspectives and do research alongside people deeply engaged with the topics I'm most interested in," Schulman said. Earlier this year, former OpenAI safety executive Jan Leike left the company to join Anthropic in a similar role. Leike, who was part of a team that focused on the safety of future AI systems, expressed his disagreement with the company leadership's priorities and said that these issues had reached a "breaking point". Founded in 2021 by seven former OpenAI employees, the AI start-up Anthropic attempts to position itself as more safety-centred than OpenAI. In June, it launched its Claude 3.5 Sonnet model, which it claims operates at twice the speed of its predecessor and is ideal for complex tasks, boasting a "marked improvement" in grasping nuance, humour and writing high-quality content "with a natural, relatable tone". The company has gained a large amount of attention in a short space of time and is backed by various tech giants including Google and Amazon. Don't miss out on the knowledge you need to succeed. Sign up for the Daily Brief, Silicon Republic's digest of need-to-know sci-tech news.

Another OpenAI founder moves to arch-rival Anthropic

Just two of the gang of eleven remain as safety concerns swirl Anthropic has hired yet another of OpenAI's founders, this time bringing on Durk Kingma in an unspecified role. Kingma revealed the news in a post on X, saying he was excited to join Anthropic because he believes in its approach to developing AI - a frequent refrain among the growing throng of OpenAI heavyweights who have left for the competition. Anthropic was formed by former OpenAI and Google staff, among others, and strives to provide artificial intelligence in a safe and comfy manner. Like OpenAI, Anthropic offers a family of models, and Claude is its answer to ChatGPT. "[I'm] looking forward to contributing to Anthropic's mission of developing powerful AI systems responsibly," Kingma posted. "Can't wait to work with their talented team, including a number of great ex-colleagues from OpenAI and Google." Only two of OpenAI's 11 founders are still associated with the Microsoft-backed AI house. Kingma left his role as boss of OpenAI's algorithms team in 2018 after just two years under CEO Sam Altman's direction - joining Google as a research scientist on Google Brain, which was folded into DeepMind in 2023. AI expert Gary Marcus, a long-time critic of OpenAI, has described the many senior leadership departures as a "slow-motion train wreck" at a business that's irrationally overvalued. "Investors shouldn't be pouring more money at higher valuations, they should be asking what is going on," Marcus observed last week in the wake of a new round of OpenAI departures, which included CTO Mira Murati. Kingma, who is based in the Netherlands, will largely work remotely for Anthropic. We asked Kingma for comment. Anthropic declined to elaborate on the news, only pointing us to Kingma's post on X. Kingma's departure from OpenAI and arrival at Anthropic was not immediate, but his reasons for choosing his new employer echo those of other founders who left: Concerns that Altman and his team weren't being as open as the company name suggests. Dario Amodei, who joined OpenAI after its founding to lead development of GPT-2 and GPT-3, left to form Anthropic, told Fortune last year he departed OpenAI because it was focused purely on making more powerful models without a proper focus on safety. "You don't tell the models what their values are just by pouring more compute into them," Amodei argued, noting that a number of OpenAI employees agreed with him on the need to put safety on equal footing with growing computing power. "We really trusted each other and wanted to work together," Amodei added, "so we went off and started our own company with that idea in mind." Amodei and his sister, Daniela (who served as VP of safety and policy at OpenAI before leaving), founded Anthropic along with various former OpenAI employees. Those that left with the siblings include Benjamin Mann; former El Reg reporter Jack Clark; and Jared Kaplan. Anthropic has also hired Jan Leike, OpenAI's former head of alignment, who departed OpenAI in May. Leike cited disagreements with OpenAI leadership as the reason for his departure, writing that "safety culture and processes have taken a backseat to shiny products" at the hype-loving AI biz in recent years. Industry talk now suggests that OpenAI is considering restructuring itself as a for-profit business with Altman at the helm - a possibility that's unlikely to ease concerns about his leadership qualities.

Who is Durk Kingma, Anthropic's latest transfer from OpenAI?

Durk Kingma, a co-founder of OpenAI, announced today that he is joining Anthropic. In posts on X, he shared that he will work mostly remotely from the Netherlands but did not specify which team he will be part of or lead. When asked for a comment, an Anthropic spokesperson referred to Kingma's posts. Kingma wrote, "Anthropic's approach to AI development resonates significantly with my own beliefs; looking forward to contributing to Anthropic's mission of developing powerful AI systems responsibly. Can't wait to work with their talented team, including a number of great ex-colleagues from OpenAI and Google, and tackle the challenges ahead!" Durk Kingma, also known as Diederik P. Kingma, is a prominent figure in the field of artificial intelligence and machine learning. He earned his Ph.D. cum laude in machine learning from the University of Amsterdam in 2017. His academic work, particularly in deep learning and generative models, has had a profound impact on the AI community. Kingma is best known for co-developing several groundbreaking techniques in AI, including the Adam optimizer, a widely-used optimization algorithm in deep learning, and Variational Autoencoders (VAE), a type of generative model that enables unsupervised learning and has applications in image generation and other AI tasks. These contributions have become foundational tools in the development of generative AI models. Before his work at OpenAI, Kingma spent several years as a researcher at Google Brain, where he focused on innovations in machine learning and generative models. He played a pivotal role in the creation of influential AI systems such as DALL-E and ChatGPT, which have helped revolutionize text-to-image generation and natural language processing. In 2015, Kingma co-founded OpenAI, a leading research organization in AI, where he led the algorithms team. However, in 2018, he transitioned to being a part-time angel investor and advisor to AI startups, and later rejoined Google Brain. After Google Brain's merger with DeepMind in 2023, Kingma continued to contribute to cutting-edge AI research.