"Invest through exponential curves before bottlenecks arrive Asked whether it is worth committing tens to hundreds of billions into AI infrastructure when no one knows how long scaling curves continue, Zuckerberg said: "I think it's likely enough that it will keep going that it is worth investing the tens or you know 100 billion plus... assume that if that kind of keeps going you're going to get some really amazing things." He acknowledged the honest uncertainty — "I don't think anyone in the industry can really tell you that it will continue scaling at that rate for sure" — but argued that waiting for certainty means missing the window entirely. When you can see a plausible exponential and the downside of being wrong is recoverable, committing early is often cheaper than waiting for proof. Source: Mark Zuckerberg, Dwarkesh Patel Podcast, Llama 3, Meta AI, Future of AI"