@agentic.james: Saying LLMs are just fancy autocomplete is like saying humans are just fancy baby makers. Both are technically true and both completely miss the point. In the pursuit of making babies humans set foot on the moon, built the standard model of particle physics, and made weapons that can vaporize the planet. In the pursuit of predicting the next word LLMs have solved 100 year old math problems, mapped the structure of nearly every protein known to man, and found exploits in every major OS and browser we use. The simple output metric is not the story. The algorithm optimizing for it is, and that algorithm gets unfathomably complex and keeps evolving over time. So if you are fine calling humans fancy baby makers, go ahead and keep calling LLMs fancy autocomplete. Otherwise you might be underestimating what is actually happening here.

963

2026-06-05 14:42:09

To see more videos from user @agentic.james, please go to the Tikwm homepage.