Without being a dualist, we can say from neurobiology, ecological psych, coord dynamics, neural reuse that meaning isn't simply upstream.
Technically it can't be because of the language problem is post-hoc.
You're an engineer so you have a synthetic view of meaning, but it has nothing to do with intelligence. I'd study how you gained that view of meaning.
A meaning ladder is arbitrary, quantum field dynamics can easily be perceived as Darwinism, and human speech isn't meaningful, it's external and arbitrary and suffers from the conduit metaphor paradox. The meaning is again derived from the actual tasks, scientifically no speech act ever coheres the exact same mental state or action-syntax.
Sorry you're using a synthetic notion of meaning that's post-hoc. Doesn't hold in terms of intelligence. Not even Barbour (who sees storytelling in particles) et al would assign meaning to Fermions or other state changes. It's good science fiction, but it's not science.
In neuroscience we call isolated upstream meaning "wax fruit." You can see it is fruit, but bite into it, the semantic is tasteless (in many dimensions).
Words are merely wax fruit metaphors for meaning, they aren't meaningful in and of themselves. That's how dictionaries exist. Any reality understood from words is mere simulation.
The idea of meaning is contradictory, it's not strictly an illusion. There's a huge difference. State changes mean differences, they don't ensure meaning. This is an obvious criteria. We have tasks and the demands are variable. We can assign meaning, but where is the credibility? Is it ever objectively understood? No. That's contradictory.
You have to look at mental events and grasp not only what they are, both material and process, how the come to happen, they're both prior and post-hoc, etc.
I study meaning in the brain. We are nit sure if it exists and the meaning we see in events and tasks are at a massive load. Any one event can have 100s even 1000s of meaningful changes to self, environment and others. That's contradictory. Searle is not even scratching the surface of the problem.
Thought is wordless. It's made in action-spatial syntax. As these are the defined states of intelligence, this would have been a far better approach to emulate. Words and images are the equivalent of junk code. Semantics can;t be specifically extracted from them.
There's much more meaning than can be loaded into statements, thoughts, etc. And conscious will is a post-hoc after effect.
Any computer has far less access to the meaning load we experience since we don't compute thoughts, thoughts aren't about things, there is no content to thoughts, there are no references, representations, symbols, grammars, words in brains.
Searle is only at the beginning of this refutation of computers, we're far more along now.
It's just actions, syntax and space. Meaning is both an illusion and fantastically exponential. That contradiction has to be continually made correlational.
Action patterns in syntax? They already exist, the binary chose to forgo that level emulation for arbitrary words, in arbitrary symbols, predicted, geometrically arranged in space as "meaning" as tokens.
I'd suggest comp sci caught the low fruit, whatever comes out of thekeyboard as a basis, non too smart.
We're probably post-narrative and post-lexical (words) but haven't become aware of what to possibly update these tools with. Post-truth is an abstraction rooted in the arbitrary.
Reality is specific. Actions, materials. Words and language are arbitrary, they're processes, and they're simulations. They don't reference things, they represent them in metaphors, so sure they have "meaning" but the meanings reduce the specifics in reality which have many times the meaning possibility to linearity, cause and effect. That's not conforming to the reality that exists, that's severely reducing, even dumbing down reality.
I think the best way would have been to assume thought is wordless (as the science tells us now), and images and probability (as symbols) are still arbitrary. That was the threshold to cross. Neither neurosymbolic, nor neuromorphic get there. Nor will any "world model" achieve anything as models are arbitrary.
Using the cybernetic to information theory to cog science to comp sci lineage was an increasingly limited set of tools to employ for intelligence.
Cybernetics should have been ported expansively to neurosci, then neurobio, then something more expansive like eco psychology or coodination dynamics. Instead of expanding, comp sci became too reductive.
The idea a reductive system that anyone with a little math training could A/B test vast swaths of information gleaned from existing forms and unlock highly evolved processes like thinking, reasoning, action and define this as a path to intelligence is quite strange. It defies scientific analysis. Intelligence is incredibly dense in biology, a vastly hidden, parallel process in which one affinity being removed (like the emotions) and the intel vanished into zombiehood.
Had we looked at that evidence, we'd have understood that language/tokens/embedded space couldn't possibly be a composite for all that parallel.
> They respond in much smarter ways than traditional multi-layer perceptrons or traditional RNNs like LSTM networks because their attention mechanisms helps them to make much better sense of the input they were provided by selectively paying more or less attention to different parts of the input (including their output generated so far). This is similar to what humans do.
Was a mistake from the beginning to use language as the basis for tokens and embedded spaces between them to generate semantics. It wasn't thought out, it was a snowball trial and error that went out of control.
The effect is easy to grasp, we're moving away from Pleistocene communication technology, which is benchmarked in dated tools like reading, math, SAT/ACT scores. Human intelligence isn't going away, another form of expression that measures intelligence that isn't text or numbers based is arriving, we simply don't have the capability to measure that, though, here clearly are markets for that intelligence.