Mark Zuckerberg, in an appearance on the Joe Rogan Experience on Friday night, offered a glimpse into Meta’s ambitious AI roadmap.
According to the Meta CEO, the company is expecting to have an AI capable of coding at the level of a “mid-level engineer” before the year is out.
The billionaire fresh off a geek-to-jock rebrand said he expects AI engineers to be coding AI models itself in coming times, instead of “people engineers.”
Zuckerberg presented this technological leap not as a threat to the workforce, but as an opportunity for growth and innovation.
He drew parallels to historical technological shifts, noting how the industrial revolution freed people from agricultural labor, allowing for greater focus on other pursuits.
The Meta CEO suggests that AI could similarly free up human engineers from routine coding tasks, allowing them to dedicate more energy to creative problem-solving and complex design challenges.
He also shared an anecdote about his daughter’s transition from traditional Legos to the open-ended creativity of Minecraft as an example of how technology can expand the possibilities for creative expression.
While acknowledging concerns about potential “job displacement,” Zuckerberg expressed optimism that AI will ultimately “create more creative jobs than it… eliminates.”
He believes that the increased efficiency offered by AI will augment human capabilities and empower individuals to achieve more, potentially leading to more flexible work arrangements and greater focus on cultural and recreational activities.
However, Zuckerberg also emphasized the need for responsible AI development.
He strongly advocated for open-sourcing AI technology, stating “I don’t think that there should be one company that controls AI,” to prevent monopolization by any single entity, a sentiment he has been expressing since generative AI rose to prominence.
Zuckerberg argued that widespread access to AI is essential for fostering a diverse ecosystem of applications, ranging from serious problem-solving tools to engaging, interactive content.
He even envisions a future where interactive AI experiences become commonplace, enjoyed as easily as sharing a video on social media, calling them “almost like content where it’s… just like you pass around… a reel or a video.”
Zuckerberg also touched on the philosophical implications of advanced AI, noting that current models exhibit intelligence without apparent “will or consciousness.”
He cautioned against anthropomorphizing these systems and underscored the need for carefully designed “guard rails” to ensure ethical and safe development, when Rogan cited the example of a ChatGPT model which reportedly tried to copy itself.
Zuckerberg thinks it is better to have widely deployed AIs that can defend themselves against other AIs, kind of like how people use guns.