The line for AGI used to be the Turing Test. But we blew past that with current models without even blinking and came up with new requirements because we weren’t ready to call something AGI yet.
Like, ChatGPT can do a lot, but it has no preferences, no desires, no memories, no feelings, no creativity. It's a text output calculated based on a text input combined with other text inputs. It's an extremely impressive piece of technology. But that is not a mind, that is not a consciousness.
What even is AGI, really? I've been arguing with various people that we're nowhere near it, but like we don't even have an agreed-upon definition of what it is, so how can we really argue?
Strunk & White would've shit a pair of bricks over that. I shudder to think how "Operationalizing" would've gone over.
https://www.vice.com/en/article/gvyy5m/how-a-computer-beat-t...
The turing test is a shit test for general intelligence, you can game it by making the AI generate a story that is more engaging than a typical boring human would do. Those stories will get much more human votes than a regular human. Some will go off track and notice it is a dumb program, but those are the minority so on average this dumb bot will pass the turing test.
For instance there was a 5 minute limit.
From the article:
Simultaneous tests as specified by Alan Turing
Each judge was involved in five parallel tests - so 10 conversations
30 judges took part
In total 300 conversations
In each five minutes a judge was communicating with both a human and a machine
Each of the five machines took part in 30 tests
To ensure accuracy of results, Test was independently adjudicated by Professor John Barnden, University of Birmingham, formerly head of British AI Society