Anthropic Co-founder Tom Brown Shares 5 Career Tips After Self-Studying AI from Scratch

Tom Brown, cofounder of Anthropic and one of the first 20 employees at OpenAI, shared five career tips based on his journey from a B- in linear algebra to shaping the future of AI. His path wasn’t traditional—instead, he relied on self-study, networking, and persistence to break into the field. Brown began his career at Grouper, a group dating app, where he built connections with key figures in tech. One of them was Greg Brockman, OpenAI’s cofounder and president. Brockman’s regular posts on Grouper caught Brown’s attention, and the two developed a close relationship. That connection eventually opened the door to OpenAI. Brown emphasized the power of surrounding yourself with people you admire. “Surround yourself with people you want to be like,” he wrote. “You'll become more similar to them over time.” He also stressed the value of mentorship and learning alongside peers. “When learning, it's much easier if you have a mentor or two and a group of friends learning alongside you,” he said. Initially, Brown felt intimidated by AI, believing only elite minds could contribute. But he decided to take the leap. He spent six months deeply studying AI research to prepare himself. He used resources like Sheldon Axler’s “Linear Algebra Done Right,” the Google DeepMind e-book “How to Scale Your Model,” and the career guidance platform 80,000 Hours. He also took Coursera courses, completed Kaggle projects, and used a YC alum credit to purchase a GPU for hands-on practice. Once ready, Brown reached out to Brockman every month, offering help in any capacity. “I’d say, ‘I’d love to help out in some way. I’ve done distributed systems work. I’m happy to mop floors,’” he recalled. His persistence paid off—eventually, he joined a gaming project at OpenAI, and after nine more months, he began working on machine learning. In 2021, Brown left OpenAI with Daniela Amodei and Dario Amodei to cofound Anthropic. While he doesn’t believe his exact path is replicable today—especially given how saturated the AI field has become—he still offered advice for aspiring AI professionals. He encouraged taking risks and focusing on projects that excite you or would make an ideal version of yourself proud. Brown’s final tip? Just start. “The best way to get good at something is usually by doing it directly,” he wrote. “Try doing it first, then see where you fail. That will show you where you need to practice.” And to reduce the pressure of failure, he advised keeping personal expenses low, so you can afford to experiment and grow.