Every tech baron worth their Patagonia vest is talking about AGI these days – albeit with mixed feelings. Some await our robot overlords with ravishing excitement; others anticipate a digital apocalypse.
The divergence stems from several motivations: personal perspectives, vested interests, and the ambiguity of what exactly artificial general intelligence entails.
Demis Hassabis, CEO of DeepMind, describes it as “knowledge on a human level” – and his opinion carries a lot of weight. Hassabis has turned London-based DeepMind into one of the world’s leading AI labs, with building AGI as its core mission.
“The progress is pretty incredible.
This week, the former chess prodigy and video game pioneer revealed his own expectations for AGI’s arrival.
“The progress over the last few years has been pretty incredible,” Hassabis said on Tuesday at the Wall Street Journal Future of Everything Festival. “I see no reason why that progress should slow down. I think it can even accelerate. So I think we could be just a few more years, maybe within a decade.
He’s left some wiggle room, but clearly doesn’t see AGI as a distant prospect. But what about his fellow tech greats? This is what they predict.
Geoffrey Hinton – Turing Award winner and ex-Googler
Geoffrey Hinton is so concerned about AI that he left Google to warn about the risks of the field. In the wake of his departure, Hinton made a new prediction about when AI will surpass human intelligence. OminousThe deep learning legend dramatic accelerated its original prediction of 30-50 years.
“I’m predicting five to 20 years now, but without much confidence,” he said on Twitter. “We live in very uncertain times. It’s possible I’m completely wrong about digital intelligence catching up with us. No one really knows why we should worry now.”
I now predict 5 to 20 years but without much confidence. We live in very uncertain times. It’s possible I’m completely wrong about digital intelligence catching up with us. No one really knows why we should be concerned now.
— Geoffrey Hinton (@geoffreyhinton) May 3, 2023
“By 2029, computers will have human-level intelligence,” he said. “That leads to computers having human intelligence, we put them in our brains, connect them to the cloud, and expand who we are. Today, that’s not just a future scenario. It’s here, in part, and it’s going to accelerate .
Ben Goertzel – CEO at SingularityNET and Chief Scientist at Hanson Robotics
Ben Goertzel, a divisive figure in tech circles, helped popularize the term AGI. He is also prone to bold statements about the future of technology. At a conference in 2018he added a few more.
“I don’t think we need fundamentally new algorithms,” he said. “I think we need to connect our algorithms in different ways than we do now. If I’m right, we already have the core algorithms we need… I believe we’re less than a decade away from creating human-level AI.
Before you flee to the fallout shelter, it’s worth noting that Goertzel isn’t the most sincere predictor. “It will take place on December 8, 2026, my sixtieth birthdaye birthday,” he added. “I’ll put it off until then to have a great birthday party.”
Jürgen Schmidhuber — co-founder of NNAISENSE and Director of IDSIA
When it comes to technical predictions, Schmidhuber looks beyond AGI and to “the singularity.” Broadly speaking, this refers to a time when AI is becoming so uncontrollably advanced that it is irreversibly changing humanity. What could go wrong?
“[The Singularity] is only 30 years away, if the trend doesn’t break through, and there will be fairly cheap computing equipment that has as many connections as your brain but is much faster,” he told futurism in 2018.
Yoshua Bengio – professor of computer science at the University of Montreal
Like his friend and fellow Turing Award winner Yann LeCunn, Bengio prefers the term “human-level intelligence” to “AGI.” Either way, he’s skeptical about predictions for its arrival.
Herbert A. Simon — AI pioneer
Herbert A. Simon, a founder of AI, has the earliest prediction on our list. The Nobel Prize winner once went so far as to estimate that AGI would come along 1985. Frankly, he made that bet way back in 1965. “Machines will be able to do any job a human can do in 20 years,” said Simon.
Simon’s guess shows how fickle the predictions of gaming technology can be. If you think there is room for improvement, please let us know through the usual channels, but do it quickly, because the clock is ticking fast. Possible.