← Lex Fridman Podcast

Career advice in AI: Where to work? | Lex Fridman Podcast

Lex Fridman · 12m
aicareer developmenttech culturesilicon valleyresearchwork-life balancetechnology careers

Résumé

In this discussion about career paths in AI, the speakers explore the complex trade-offs between academic research and industry positions. They highlight the challenges faced by researchers, including low pay in academia and the increasingly secretive nature of frontier AI labs. The conversation reveals that while industry jobs offer better compensation and meaningful impact, they often come with intense work cultures characterized by the '996' model - working from 9am to 9pm, 6 days a week. The speakers note that while AI research labs like OpenAI and Anthropic are driving significant technological progress, this comes at a potential cost of employee burnout and limited public visibility of work. They emphasize that career choices in AI are highly personal, depending on individual preferences for publishing, compensation, and work environment. The discussion also critiques the Silicon Valley 'bubble' mentality, warning that intense focus and isolation can lead to disconnection from broader human experiences and perspectives.

Points clés

  • Consider the trade-offs between academic research and industry positions, weighing factors like compensation, publication opportunities, and personal fulfillment
  • Be aware of the intense '996' work culture in AI companies, which can lead to burnout and potential health issues
  • Recognize the value of maintaining perspective outside of tech echo chambers, including reading diverse literature and understanding different cultural contexts
  • Understand that AI career paths are not permanent - mobility and adaptability are key

Citations notables

"I feel like my friends who are professors seem on average happier than my friends who work at a frontier lab"
"The only thing that is forever is that nothing is forever"
"This leaprogging nature and having multiple players is actually an underrated driver of language modeling progress"