June 30, 2026
Hot takes, cold consciousness
Words Are a Byproduct of Consciousness. For LLMs, It's Backwards
Humans say ideas come first, but the comments are fighting over whether that’s even true
TLDR: The article argues that humans have thoughts before words, while chatbots build meaning from words alone, and that difference may matter more than speed or cost. In the comments, readers split between meditation gurus, philosophy history buffs, and skeptics saying the whole “consciousness comes first” idea is far from settled.
A spicy philosophy post about artificial intelligence has turned into a full-on comments-section cage match. The article’s big claim is simple enough for non-experts: humans usually feel like we have an idea first and then find words for it, while chatbots work the other way around, predicting one word after another and only seeming meaningful. The author says that one reversal explains why machines can sound smart without actually having the inner life humans do.
But the community did what communities do best: immediately argued about everything. One camp leaned deeply spiritual, with one commenter basically saying, go meditate for 10 minutes and you’ll realize you are not your thoughts. Another went full ancient-philosopher mode, dragging in St. Augustine and arguing that humans learn language through desire, body language, and messy real-world experience. Meanwhile the skeptics were not buying the article’s opening premise at all. One pointed to Helen Keller’s own writing as evidence that language and consciousness may be more tangled together than the author admits. Another delivered the thread’s sharpest reality check: saying “words are a byproduct of consciousness” is a huge claim when people can literally dream in words.
The result? Less “wow, AI is amazing” and more “hold on, what even is consciousness?” It’s the kind of debate that starts with chatbots and ends with everyone accidentally reenacting freshman philosophy, but with better one-liners.
Key Points
- •The article argues that humans typically form ideas before expressing them in words, while LLMs generate language by predicting the next word from previous text.
- •It presents a historical timeline from speech and writing to printing, computing, the internet, and smartphones to frame the current AI moment as another technological inflection point.
- •The article identifies the 2017 introduction of the Transformer by a Google team as the key development that enabled modern LLMs.
- •LLMs are described in the article as systems built from large amounts of text that use mathematical methods to predict subsequent words.
- •The article states that although LLMs are currently costly and power-intensive, they are likely to become more efficient over time, similar to earlier generations of computers.