John Carmack reckons we'll see an artificial general intelligence fully realized in 2030

The latest big-name personality to chip in on the artificial intelligence front is John Carmack, the dev who brought us Doom (and Quake) way back when.

Carmack reckons that an artificial general intelligence or AGI is likely to be realized around the year 2030.

AGI is a somewhat controversial idea, and even its definition can be argued over, and very much is. But the broad idea is that this is a 'real' AI - one that can reason and understand in a human way, as opposed to what we have now with AI (large language models, or essentially very fancy data scraping tools capable of doing a convincing impression of AI).

As The Register reports, Carmack shared his thoughts at an event for his AGI startup Keen, announcing it had hired Richard Sutton, who Carmack calls the "godfather of reinforcement learning" no less.

Sutton said that coding an AGI seems feasible using current tech, and that 2030 is certainly a possibility for a prototype AI that "shows signs of life" (again, a pretty broad and sweeping definition there).

The startup Keen is going to work towards this goal with a relatively modest funding budget ($20 million).

Carmack commented:

"There are fundamental research questions that need to be answered, and we have internal projects and angles of attack."

He added:

"We are six, seven, eight years out from something really big and important being publicly visible."

This isn't the first time the end of this current decade has been locked in as a timeframe for the emergence of AGI.

The prediction of AGI arriving by 2030 was previously made by thinktank Sustensis, which noted that it was in line with 61% of AI specialists that were polled in this belief.

As Sustensis previously observed:

"The probability that AGI may indeed arrive by 2030 increases even further if we include usable quantum computers, which should be available in 2-3 years' time, significantly increasing the processing power for some calculations. Therefore, we should consider 2030 as the AI's tipping point when it may be outside of human control."

And that latter prospect, of course, has been what has prompted a lot of worry - and what some have labeled scaremongering - around AI. Whatever the case, it does feel like the current progress of AI is pretty much unchecked thus far, and that some more checking needs to be implemented before we get too far down the road towards some kind of artificial general intelligence.