LLMs are human-level reasoning at the speed of light, with infinite scalability.

· Bits and Bobs 3/10/25
  • LLMs are human-level reasoning at the speed of light, with infinite scalability.
    • A lot of things that require human-level reasoning have always had a friction that is governed by the "at the speed of a human" constraint.
      • That constraint is even stronger than it first appears, because you first have to find and engage a specific human in the task in the first place, which can be much harder.
    • This constraint forms a kind of force of gravity; omnipresent and unchanging.
    • But LLMs change that force of gravity, and suddenly a lot of things that everyone has just known are obviously impossible surprisingly become possible.