Who's Behind Claude: How Former OpenAI Employees Built Anthropic's Team

Anthropic grew out of an internal split at OpenAI and in just a few years turned Claude into one of the most prominent products on the AI market. Behind this model stands not a single bright founder, but a team with different roles: research, safety, operations and behavior tuning.

Where did Anthropic come from

Anthroptic was launched in 2021 by former OpenAI employees, including Dario and Daniela Amodei. The reason for leaving was not personal drama, but a different vision of how to build strong artificial intelligence. The new company immediately made safety, interpretability, and model safety a priority, not just a race for demo effect.

By April 2026, this strategy had worked: Anthropic is already perceived as OpenAI's main rival, and its valuation in an official round reached $380 billion. On secondary markets, investors, according to business media, were driving it even higher — roughly to $1 trillion. Claude itself went viral not because of a single beautiful release, but because of practical usefulness.

The model was loved for working with long texts, analyzing correspondence, code and documents without noticeable loss of context. Claude particularly resonated with developers: a separate tool called Claude Code can read code bases, change files, run tests and bring tasks to a working state. Even within the industry this was noticed: Jensen Huang publicly called Claude a strong leap in coding and reasoning.

Who makes Claude

The main idea of the material is that Claude is not backed by an abstract "laboratory," but by a very specific combination of people with different competencies. One sets the scientific trajectory, the second keeps the company operational, the third tries to understand the internal mechanics of models, the fourth is responsible for the character and style of the assistant's behavior. Together this looks not like another chatbot, but like a product with its own philosophy.

• Dario Amodei — CEO and co-founder of Anthropic, former VP of Research at OpenAI, worked on GPT-2 and GPT-3.
• Daniela Amodei — President of Anthropic, responsible for operations, processes, partnerships and the company maintaining its own growth.
• Chris Olah — co-founder and interpretability researcher who studies how exactly model thinking is organized internally.
• Amanda Askell — philosopher by training, working on personality alignment, ensuring Claude is helpful, honest, cautious and polite.

This team has a rare composition for the AI market. Here, the classical cult of the "strong technician" is important, but also the presence of people who can build processes, set ethical frameworks and analyze a model as a complex organism. It is no coincidence that Anthropic is often described as a company where the product is built simultaneously on research, safety policy and very fine-tuning of behavior. This is why Claude is perceived not just as a powerful model, but as a system with character.

Safety as strategy

Anthroptic's public line reads well through the texts and decisions of Dario Amodei. In October 2024, he released an essay Machines of Loving Grace, describing a scenario in which powerful AI helps medicine, education, economics and governance, rather than becoming a source of chaos. In January 2025, he separately wrote about DeepSeek, China and export controls, insisting that democratic countries should maintain their lead in AI. In April 2025, he published the text The Urgency of Interpretability: strong models already exist, but we still poorly understand what exactly is happening inside them.

"The elegance of machine learning is the elegance of biology."

This idea well describes the approach of Chris Olah, one of Anthropic's key researchers. For him, a model is not just an engineering system with understandable gears, but a complex digital organism that needs to be opened up and studied almost like a living structure. The same approach appeared in the loudest conflict around Anthropic in February 2026. The company was discussing Claude's work with the Pentagon, but refused to remove two restrictions from the contracts: a ban on mass internal surveillance and a ban on fully autonomous weapons. This was not a PR posture, but a demonstration that safety for Anthropic is part of the product, not an add-on to it.

What this means

The history of Claude is not just the history of a successful model, but the history of a correctly assembled team. Anthropic shows that in the AI race, victory goes not to those who simply release another chatbot, but to those who combine strong research, operational discipline and a clear position on safety. For the market, this is a signal: competition with OpenAI is now taking place not only on the quality of answers, but on who users and companies are willing to trust with real work tasks.