A high school dropout who got hired at OpenAI says he used ChatGPT to learn Ph.D.-level AI (www.businessinsider.com)

Gabriel Petersson, a high‑school dropout who taught himself to code while working at startups, says he used ChatGPT to acquire “Ph.D.-level” AI knowledge and is now a research scientist at OpenAI on the Sora team. Petersson describes a pragmatic, top‑down learning loop: pick a concrete project, have ChatGPT generate code, iterate by fixing bugs with the model’s help, and recursively drill into specific components until the underlying theory clicks. His background includes building recommendation systems, scraping and integrations at small companies, and prior engineering roles at Midjourney and Dataland. The case highlights how large language models are democratizing advanced ML education and accelerating a skills‑first pathway into research roles. For the AI/ML community this implies faster, project‑driven skill acquisition, a shift in hiring signals from formal credentials to demonstrable impact, and more accessible bootstrapping of complex systems. Technically, Petersson’s method emphasizes interactive code synthesis + debugging and targeted conceptual deep dives enabled by LLMs — a practical template for applied ML learning. The trend dovetails with broader industry moves (Sam Altman, a16z, Palantir) questioning traditional education models, raising both opportunities for talent diversification and the need for rigorous verification of LLM‑generated knowledge.