Isn't this how human brains work? We're just large probabilistic neural inference machines. A network of neural weights guided by past training.
To be honest I think the debate about what "sentience" is is inconsequential navel gazing. There are many different kinds of intelligence - even within humans. What matters is how useful that intelligence is as applied to solving real world problems. Like it or not, LLMs produce intelligence which is very, very useful for 1.5B people and rapidly growing.
I think this ultimately boils down to the classical economic debate of marginal utility. There isn't an objective way to value a product or service. Each person decides for themselves what said product or service is worth based on their needs and preferences. Intelligence works the same way. We don't have the right to tell someone that their perception of the value of that intelligence is wrong. They alone determine that.
One thing we can all agree on, is that the capability of this intelligence is expanding rapidly. In a few short years, we went from Will Smith spaghetti hands to full length movies and strikingly realistic images. For 60 years, passing the Turing Test was considered Star Trek level science fiction. Last year GPT 4.5 passed the Turing Test. AI is already being used to convince people over audio that they are real, and very soon, this will occur over video.
I think people are being too dismissive of this intelligence. It doesn't need to be perfectly humanoid to be considered intelligent.