No one expected natural language to be solved in 2020. At the time the smart money was on 20 to 80 years. Many people are still acting like it's not.
Things happen not at all and then all at once.
We have the cheap humanoid platforms coming out of China and they cost six months minimum wage salary in the developed world. Once a model is developed that can use those platforms to match humans for simple tasks we will see the hollowing out of all unskilled physical labor overnight.
NLP still isn't solved. LLMs are a big step change, but NLP still isn't solved.
I can talk to a computer and it can talk back. The rest is commentary.
ELIZA passed the Turing test in the 60s. That's not a statement of technological progress towards AI, that's a statement on 1) the uselessness of the Turing test, which has been known for decades and 2) the gullibility of the human psyche when it comes to assigning intelligence to anything that can mimic human communication patterns.
It did not.
Even Turing didn't respect the "Turing test", and your willingness to ignore reality in order to defend such a useless metric on the path to AI marks you as one of the gullible ones.
I have no idea why you're bringing up the Turing test, then saying it doesn't matter when you used it wrong.
What I'm talking about is that I use Claude daily to find stupid bugs in my code. I can now spend my time on things that matter.
This was science fiction 5 years ago.