Sam Altman’s $7 trillion AI plan: But then, how else can we save the world?

livemint.com:

In a world where billionaires launch into space for fun, Sam Altman of OpenAI fame has upped the game of one-upmanship by throwing a number into the universe that has made economists and mathematicians do a collective face-palm: $7 trillion. That is not a typo. It is seven trillion US dollars, a number so large that it’s practically a black hole in the world of finance that makes most tech-startup pitches look like lemonade-stand business plans.

Sam Altman has said he needs $7 trillion. Why? According to reports, he is on a mission to build a global network of AI chip factories, data centres and power plants. Apparently, training the next generation of AI models requires more computing power because they may need to solve a Rubik’s cube while juggling flaming torches, riding a unicycle and translating Tolstoy’s War and Peace into Klingon.

These models will attempt to understand the nuances of human language in all its glory, including why we say “parkway" when we drive on it and “driveway" when we park on it. Hence, all the costs. Is it needed? Some experts argue that this level of investment is key to keeping up with the exponential growth of AI capabilities.

Others suggest Altman accidentally added a few extra zeros to his funding estimate. So, how much is $7 trillion? It’s enough to do plenty else. I could suggest a few ideas.

The money could fund America’s space agency Nasa for more than 250 years at its current budget. Cost: $7 trillion. Outcome: We’d probably find aliens before Altman’s AI becomes self-aware.

Or we could build around 4,600 Burj Khalifa skyscrapers like the one in Dubai. Cost: $1.5 billion per skyscraper. Outcome: The world’s most expensive game of Jenga.