OpenAI Fellows Fall 2018
OpenAI
OpenAI opened applications for its next Fellows batch back in 2018 — a paid 6-month AI research apprenticeship. No PhD required, just a way in for people outside the usual academic pipeline.
Based on reporting by OpenAI — read the original for the full story.
Summary, retelling and take written by AI under human oversight; images are AI-generated illustrations. How we work · Report an error
Buried in OpenAI's blog archive is a short, almost administrative post: applications are open for the Fall 2018 Fellows cohort. It's easy to skim past, but the program itself says something about how OpenAI thought about talent even in its earlier, smaller days.
The pitch is simple. Six months, paid, hands-on AI research inside OpenAI's walls, working alongside the same researchers publishing the papers everyone in the field was reading at the time. No mention of needing a machine learning PhD or years of publication history — just enough curiosity and technical chops to be useful in a lab that was, in 2018, still a fraction of the size it is now.
That detail matters more than it looks. Around this period OpenAI was competing with Google Brain, DeepMind and a handful of well-funded startups for a small pool of people who actually knew how to train large models well. A fellowship is a cheap, low-risk way to test candidates before committing to a full hire, and it doubles as a pipeline for people who didn't come up through the standard Stanford-or-Berkeley-PhD route. Plenty of now-familiar names in AI research got their start through exactly this kind of side door rather than the front one.
It's also a quiet signal of how OpenAI saw its own trajectory. A company confident it's about to shrink doesn't run apprenticeship programs; a company that thinks it's about to need a lot more research capacity does. Six years on, knowing what OpenAI became, this little recruiting post reads less like routine HR copy and more like a preview of the hiring machine it would eventually build.
My take — AI-written commentary, not fact-checked reporting
I'll admit there's something almost quaint about reading this now, given how gatekept and credentialed AI hiring has become since. If more labs ran open, paid apprenticeships instead of demanding five years of published work before they'll even look at your resume, we'd probably have a less incestuous, less Ivy-League-coded research culture — and that's a trade worth making.
Read more about this at: OpenAI