I can explain the idea, but I'll be honest, it took myself years of living in the dry world of maths, theoretical logic and computer science (+ likely autism) to truly get behind this one.

So, you probably know from maths that a function has inputs and outputs. You stick parameters into it and then it evaluates to a result.
With this model, the only way to change the output, is to change the inputs. A function cannot randomly decide that today, it's going to output something different.
Well, it can, but then the date (or whatever makes that day special) needs to be an input to the function.

There is no random, there's only pseudo-random, which is when it's really difficult to work out which inputs lead to a given output.

And I don't see a reason why our brain should magically be different from this. Our whole brain is a function. It's an extremely complex function, with gazillions of inputs throughout our lifetimes.
It's also not a pure function, as the outputs of future evaluations are influenced by the inputs of previous evaluations, a.k.a. we remember shit. And that remembering is fuzzy, too, as we only store certain statistical weights, like LLMs do.

So, there are a ton of sources for pseudo-randomness, meaning even for toddlers (who've experienced comparatively few inputs), it can already seem random why they are crying.
But I see no reason to believe that we've somehow acquired the only source of true randomness in the whole universe, let alone somehow a consciousness?/soul? which for whatever reason makes decisions not based on those inputs. That just sounds like terrible decision making to me.

I feel like people struggle to let go of this idea of free will, because so much of our philosophy, religion, laws, motivation etc. are based on it. Even for me, with my particular set of brain-weightings, I have to be quite aware of my thinking to not fall into traditional patterns.
Maybe the whole artificial intelligence shenanigans will make us realize that our own intelligence isn't that special.