First things first
(insert everything)
The day you come here and say: "I love something" is the day you gonna surprise me.
Now about AI, the history of AI shows the "winters" it had so far. Twice, they were advertised as a breaking change before it fails to meet expectations and people lose interest on it.
But the main take on the AI winters is that the demand was never dropped. Humanity really wants to make machine as sentient as humans.
But the failures comes more because lack of technology or knowledge.
That garantees that, while a sentient machine dont appear, any new hardware development will constantly bring AI back to news.
About our current AI age, I think we are nearing a plateau. Big tech managed to massively increase the resources over the last iteration of GPT, but the results werent proportional to the investment. Meaning maybe the Big Tech can come up with a more accurate GPT but will hardly introduce a new breaking change over what we have already.
And, as you said, the best those models can do is infer the next word in a sentence in a way that can make a coherent answer for a question asked.
I believe soon the AI development will enter a standby phase, like a third AI winter, till people can bring the so called quantum processor.
The quantum processor would be a breaking change over the current processors used for Machine Learning and Deep Learning, and would allow data cientists to try new models and maybe bring the machine closer to sentience.
AI will never completely die. But its very likely it will reach a wall, because humanity is always retrying with any new changes in technology.
Though, there is no discussion this current "AI sprint" brought a whole lot of toolings that surely gonna be a breaking change to many tech fields from now on.
But unfortunelly for you, your AI girlfriend that you are paying monthly the premium sub wont say she want to have sex with you and will do old man jokes about your small penis.
Cest la vie