The AGI hype train is running out of steam

AGI is no longer a big buzzword in Silicon Valley. The unfulfilled expectations will affect the future of artificial intelligence.

Like the idea of a true driverless car or nuclear fusion, artificial general intelligence is not going to happen soon, despite hype in recent years suggesting it is not far away.

Most modern artificial intelligence is a form of deep learning or natural language processing. AI is also useful for applications like computer vision. Things have moved fast in the last decade, but today's AI is more about sifting through vast piles of data than building machines that think.

General artificial intelligence

General Artificial Intelligence is about building machines, or more accurately software, that can think in ways that resemble human thinking.

That’s a loose definition, experts argue over exactly what AGI means and doesn’t mean. They don’t even agree on its name: there are many variations in use each with a different, nuanced meaning.

AGI doesn’t necessarily mean sentience. At best it is an early step on the path, but it is possible machine sentience can never be achieved.

All quiet on the AGI front

However, the financiers behind AGI have gone quiet lately. This suggests they have lost their appetite for the technology.

Peter Thiel — the tech billionaire and New Zealand citizen— says the big names in Silicon Valley have lost enthusiasm for AGI. Media reports, which are no longer online quote Thiel saying:

“Elon’s not talking about it anymore and Larry [Page] is off to Fiji and doesn’t seem to be working on it quite as hard”.

Why general artificial intelligence will not be realised is paper published by Nature. The author argues that AGI is impossible:

"Deep learning and Big Data are among the latest approaches, and advocates argue that they will be able to realize AGI. A closer look reveals that although development of artificial intelligence for specific purposes (ANI) has been impressive, we have not come much closer to developing artificial general intelligence (AGI). The article further argues that this is in principle impossible."