Comments by "Michael Nurse" (@michaelnurse9089) on "Meta's LLAMA 3 SHOCKS the Industry | OpenAI Killer? Better than GPT-4, Claude 3 and Gemini Pro" video.

  1. I think its cool. In the early nineties this was how we did it.
    1
  2. Doubtful. Data has x amount of signal in it. The arch can either extract x or not. In the case of TF models they took the internet data from 0 to X. To get to 2x you will need another dataset on par with the internet.
    1