Comments by "" (@diadetediotedio6918) on "Microsoft LongNet: One BILLION Tokens LLM + OpenAI SuperAlignment [SINGULARITY APPROACHES]" video.

  1. You are confusing memory with reasoning, that's why you think you are looking at a "god-like intelligence".
    3
  2. 15:45 Any definition? This is very, very simple to win, I can only say that my definition of AGI encompasses the need for consciousness and real subjective experiences, the Chinese room thought experiment still remains undisproved and that implies that by that definition you would still be wrong.
    1
  3.  @marcusmoonstein242  Manipulate them in what sense? Whether the idea is merely to repeat the order of numbers or letters, or to reverse them and perform basic symbol manipulation digital computers without AI have been able to do this basically since they began to exist. IQ tests were made for <humans>, they don't consider the intrinsic limitations or underlying possibilities of other forms of intelligence (if any), so you might say that merely having a "working memory" (and there are many, many, many quotes in "working") would imply having greater intelligence than a human being you would be making a claim so weighty that you would need to write an entire book just to start demonstrating it.
    1