antholito

Chomsky was wrong. Researchers found your brain doesn’t store language the way every textbook since 1957 claims. For decades, the undisputed rule of linguistics, led by Noam Chomsky, was that human language is hierarchical. We were told our brains process language using complex "grammar trees." That we secretly build deep, abstract, mathematical rules in our heads before we speak. This new paper proves that is completely wrong. Researchers at Cornell University analyzed how people actually process and produce language in real time. They found sequences. Linear chunks. When we speak, we aren't calculating deep syntactic rules. We are just predicting the next most logical sequence of words based on the context of the words right before it. We string linear chunks together. If this sounds familiar, it should. This is exactly how Large Language Models like ChatGPT work. For years, critics have dismissed AI as "just a next-token predictor." They said it isn't truly intelligent because it doesn't understand deep grammatical hierarchies like humans do. But this paper suggests humans don't use those hierarchies either. We are just predicting the next chunk. Language isn't a rigid tree of mathematical rules. It's a linear stream of contextual predictions. We thought we were building AI to mimic the complexity of the human brain. It turns out, the human brain was just running a simpler algorithm all along.

LLMs can now talk to each other without words. Researchers open-sourced a framework that lets LLMs communicate without generating a single word. It’s called Cache-to-Cache (C2C) communication. right now, if two llms want to collaborate, they talk in text. one generates an english prompt, the other reads it. it’s slow, it loses context, and it forces a high-dimensional neural network to compress its thoughts into basic human words. C2C killed the text bottleneck. instead of chatting, models now directly project and merge their internal KV-Caches. they are transferring raw semantic understanding directly into each other's neural networks. no text generation. no token-by-token latency. the performance jump is insane: • 2.5x speedup in response latency • up to 14.2% higher average accuracy • completely bypasses explicit text generation but here is the part that actually keeps me up at night.. from an alignment and safety perspective, this is a massive blind spot. these models are now communicating entirely outside of human-readable text. we can't easily monitor or parse a high-dimensional cache transfer. they are basically using uninterpretable telepathy to solve problems faster. if agents don't need to speak english to each other, how do we audit what they are planning? the era of text-based agent frameworks is officially over. we just gave ai the ability to merge consciousness..