Every year, AI becomes significantly stronger. But we keep raising the bar for AGI - "real" Artificial General Intelligence.
<< Yes, AI is smart, but cannot solve math problems. So it's not real intelligence.>>
<<Yes, it solves simple math problems, but not difficult problems. >>
<<Yes, it solves complex math problems, but cannot reason. >>
<< Cannot write poems. Cannot draw. Cannot compose music >>
Well, now it can.
One year ago, ChatGPT even passed the Turing test for AI. Still nobody noticed.
The Turing test is the single most important test for AI. The imitation game. Designed over 50 years ago. The single simplest measurable test of real intelligence. Never passed - at least until last year.
Read more about the Turing test: https://blog.stefanmorcov.com/2022/12/can-ai-pass-turing-test-lets-ask-chatgbt.html
OpenAI ChatGPT passed the Turing test in April 2025
But we simply raised the bar even higher. "There is no AGI".
No comments:
Post a Comment