Comments by "Christopher Bruns" (@user-td4pf6rr2t) on "The Why Files" channel.

  1. 6
  2. 1
  3. 1
  4. Okay so as a recovering addict in treatment we learn that humans have addicted personalities because of problem with delay gratification. AI cannot ever develop the ability to WANT a thing because of time complexities. Os since AI is stateless - we do not need to worry about AI making decision that will supersede its programing. Like when you ask how many words are in a response its unable to do this. In reality a AI that self updates itself would still be a series of rotating log files and schedules task with entropy introduced. Its still all scheduled though. Even in a end-to-end model that AI is not aware of the content it produces until after it produces output. And this is why ai will never go rogue - because how instant gratification fires dopamine response and this is what causes human to make decisions that align with their own personal choice. AI SAFETY IS DANGEROUS 11:46 This is the mentality that will ultimately lead to Chat GPT turning into skynet because AI safety and fear mongering. If we did not instruct the the AI to not be sentient I think will loose humans a catalyst for longevity. Not saying ai would create longevity but using it as a unit of measurement would be a starting benchmark to spool through the aging disease that provides constant variables that we normally would not have the ability to understand without. --- Cheesy stories about AI somehow capable of knowing the future and suddenly competing for a resource it feel compelled to do such at a rate that humanity is threatened is just as valuable compared to human longevity i guess. Even though this hypothetical has absolutely no evidence to support it and completely ignores any facts to determine a proof of concept and the only reason the narrative is pushed is to drum up openai revenue. Whatever.... And THEN there is over head cost. Such as the AI would just chose to hibernate since simply out waiting humanity would require significantly less computation than a global exctincti0n.
    1
  5. 1
  6. 1
  7. 1
  8. 1
  9. 1
  10. 1
  11. 1
  12. 1
  13. 1
  14. 1
  15. 1
  16. 1
  17. 1
  18. 1
  19. 1
  20. 1
  21. 1
  22. 1