Yao Shunyu left OpenAI for Tencent to pursue artificial general intelligence, which is the tech industry's way of saying "I want to build God but make Him better at targeted advertising." The man spent however long at OpenAI watching Sam Altman raise billions to train models on the entire internet, then decided his best shot at AGI was working for a company whose main product is a messaging app with a payment system attached. Brilliant pivot.
Tencent calls this a talent acquisition. OpenAI calls it addition by subtraction. When your Chief AI Scientist's primary qualification is "used to work somewhere successful," you're not poaching talent. You're buying a used luxury car with 200,000 miles and telling yourself it still has that premium feel.
China wants the next "super-app" because WeChat apparently wasn't super enough. They want an app that does everything: messaging, payments, AI companionship, probably predicts when you'll die. The American approach is ten apps that each do one thing poorly and share your data with each other. The Chinese approach is one app that does everything poorly and shares your data with one entity. Efficiency.
Retail traders will interpret this headline as bullish for anything with "AI" in the name. They'll buy shares of a company that makes synthetic leather because the ticker sounds vaguely technological. Meanwhile, Yao Shunyu will spend the next three years in Shenzhen trying to explain to executives that AGI requires more than just adding a chatbot to QQ. He'll build something impressive. They'll ask why it can't recommend better restaurants. He'll remember his OpenAI salary with the fondness of a man who chose ideology over stock options.
The funniest part? Both countries think they're winning this talent war. America loses an AI scientist and celebrates having one less person who might regulate them. China gains an AI scientist and celebrates having one more person who definitely won't regulate them. Yao Shunyu gets paid either way to pursue something that doesn't exist while everyone pretends we're two quarters away from Skynet.
Artificial general intelligence remains exactly as close as it was yesterday: theoretically possible and practically fantasy, now with better funding in Mandarin.
Photo by on Unsplash
Leave a Comment