Trump says AI companies must 'self-regulate' | BBC News

From Meat to Meaning: The Forgotten History and Hidden Irony of the AI Boom<br><br>The modern tech industry speaks of artificial intelligence in hushed, almost mystical tones. We are told that by stacking enough computing power, data, and processing speed, a machine will inevitably cross a threshold into &quot;sentience&quot; or &quot;true understanding.&quot; The tech world has convinced itself that it is building a mind.<br><br>But this philosophy suffers from a profound epistemological blindness. It assumes that human language and meaning are merely software programs running on biological hardware—that if you mimic the structural output of the mind, you have successfully recreated the mechanism that generates it. In reality, the tech industry has spent billions of dollars accidentally proving the theories of a long-dead psychological framework, only to misinterpret its own success through a haze of 17th-century mysticism.<br><br>To understand the core illusion of modern AI, we have to look back to 1959 and the most explosive intellectual feud of the 20th century: the battle between B.F. Skinner and Noam Chomsky.<br><br>───<br><br>The Tunnel of Language: Skinner vs. Chomsky<br><br>In the mid-20th century, B.F. Skinner championed radical behaviorism. He argued that human language—what he called &quot;verbal behavior&quot;—is not a magical product of an unobservable, internal mind. Instead, language is shaped from the bottom up by an organism interacting with its environment, responding to stimuli, and accumulating a history of statistical reinforcement.<br><br>In 1959, a young linguist named Noam Chomsky published a scathing review of Skinner’s work. Chomsky argued that behaviorism could never explain the infinite creativity of human speech. He claimed that children hear too few examples of language to learn it purely through environmental exposure. Therefore, humans must possess an innate, hardwired Language Acquisition Device (LAD)—a universal, rule-bound grammar engine inside the head.<br><br>Chomsky’s top-down mentalist view won the intellectual argument. Behaviorism was declared dead, the &quot;Cognitive Revolution&quot; was born, and the first era of artificial intelligence spent decades trying to manually code explicit rules, symbolic logic gates, and universal grammars into brittle &quot;Expert Systems.&quot; <br><br>It was an approach that ultimately hit a wall and triggered a decades-long AI winter.<br><br>Decades later, the field finally abandoned top-down coding and pivoted toward massive, data-driven neural networks. In doing so, it picked up the exact shovel Skinner had dropped.<br><br>───<br><br>The Unexpected Convergence inside the LLM<br>Large Language Models (LLMs) like ChatGPT or DeepSeek represent the exact point where Chomsky and Skinner’s tunnels finally crashed into each other. The machine proves both of them right, yet exposes the limits of both.<br><br>LLMs do not possess a formal grammar chip, a syntax calculator, or a hardcoded concept of a noun or a verb. Instead, they ingest billions of pages of human-written text and calculate the statistical probabilities of words co-occurring. By scaling up data exposure to an extreme degree, they achieve human-level linguistic structure entirely through statistical patterns. This is Skinner’s ultimate vindication: structure can be completely extracted from environmental data.<br><br>Yet, once the machine absorbs this data, a strange thing happens. If you map the internal mathematical vector spaces of an LLM, you find that the model organically builds highly constrained, geometrically perfect parallel structures to represent concepts. This is Chomsky’s ghost: for language to function, it ultimately requires an internal structural logic.<br>Skinner and Chomsky were digging from opposite ends of the same tunnel. They finally met inside the neural network.<br><br>───<br><br>The Epistemological Blunder: Confusing the Echo for the Speaker<br><br>The tragedy of the modern AI boom is that the tech industry used a purely behaviorist methodology (statistical learning and reinforcement via human feedback) to build a machine, but then turned around and used Chomskyan, mentalist language to explain it.<br><br>When a model exhibits a complex capability, tech builders frequently use the word &quot;emergence.&quot; They treat &quot;reasoning&quot; or &quot;understanding&quot; as a mystical internal state that magically flipped on when the compute got big enough. They look at a massive spreadsheet of statistical probabilities and assume it has developed a soul.<br><br>From a radical behaviorist perspective, this is a catastrophic confusion of topography (the structure or form of a behavior) with function (the relationship between the behavior and its environmental consequences).<br><br>An LLM mimics the topography of human language flawlessly. It outputs text that looks exactly like a human reasoning. But it lacks the function. The machine’s output is not shaped by an organism interacting with a physical universe. It is a simulation of the byproduct of human behavior, completely divorced from a living history of reinforcement.<br><br>───<br><br>Why the Current Models are Hitting a Wall<br>This epistemological error is already breaking the current generation of AI through a crisis known as Model Collapse or &quot;data cannibalism.&quot;<br><br>As tech companies run out of human-written text, they have begun training new AI models on data generated by older AI models. When an LLM trains on human text, it is training on an artifact anchored to a real world of gravity, hunger, emotion, and physical stakes. But when an AI trains on AI text, it is training on a statistical average of an average.<br><br>Without a real-world environment to provide feedback, punish errors, or reward utility, the statistical weights begin to drift. The models progressively degrade, losing the rare, creative, and highly specific linguistic behaviors that humans use to solve real problems. They become functionally senile. It is the ultimate proof of a foundational behaviorist truth: if you sever a behavioral repertoire from a live, responsive environment, the behavior inevitably extinguishes.<br><br>───<br><br>Conclusion: The Missing Meat<br><br>No-one in Silicon Valley knows how meat turns to meaning.<br><br>Meaning does not arise because a sentence is syntactically perfect (Chomsky) or statistically probable (Skinner). Meaning arises because an organism needs to communicate to affect its environment, avoid danger, and survive.<br>By separating verbal behavior from the living organism, the tech industry has not figured out how physical matter gives rise to subjective consciousness. They have not built a mind; they have built a magnificent, empty library. It is a beautiful archaeological dig of the linguistic tunnel we left behind—a mirror that reflects our own intelligence back to us, while the actual meaning remains entirely on our side of the screen.

Yeah....none of them were British of course. The UK isn&#39;t even a Democracy anymore. A Socialist regime that has outlawed freedom of speech and speaking out against the Regime. Just like your buddy&#39;s in China and Russia and Iran. UK is now an Islamic state. Like France is and Spain and the Netherlands etc.