Apparently what they really mean is that by 2030, we could have an AI that can translate between languages as well as a human can. That would be impressive, no doubt. But:
1. That's not the same as AGI, despite the article's thinking that it is.
2. AGI itself isn't the same as the singularity. The singularity is where an AI can improve itself, without any further human intervention. More, it's not just improving itself to asymptotically approach some limit; it's improving exponentially or hyperexponentially.
On both counts, while impressive, this isn't the singularity, even if an AI achieves human-level translation in 7 years.
"May".
Apparently what they really mean is that by 2030, we could have an AI that can translate between languages as well as a human can. That would be impressive, no doubt. But:
1. That's not the same as AGI, despite the article's thinking that it is.
2. AGI itself isn't the same as the singularity. The singularity is where an AI can improve itself, without any further human intervention. More, it's not just improving itself to asymptotically approach some limit; it's improving exponentially or hyperexponentially.
On both counts, while impressive, this isn't the singularity, even if an AI achieves human-level translation in 7 years.