4 ms·

Implying that these people quite because they realized they weren't needed (no need for safety if AI isn't coming) rather than what seems like the common perspe

• zer00eyz 2y ago
  Rewind the clock a few years. The google engineer who was saying google had AGI. The MS paper saying GPT had hints of AGI. The MS deal "you can have everything up to AGI". The approach was one where a big enough model could be made to feed back in on it self (Q*). If we feed enough data into a big enough model then it will "tip over" and turn into AGI, or sentient or sapient... The problem is that this doesn't work, it never would have worked, the math does not support it. Pointing it back at it self doesn't get you more it gets you an ouroboros. Why do you sign a deal "up to AGI" if you dont think you're close to AGI? They did, they were high on their own supply, reality set in, people look foolish, and they moved on.
  • voiceblue 2y ago
    Seems to me like you’d be more likely to sign a deal “up to AGI” if you knew that was a pipe dream that would never happen. It’s easy to sign away your firstborn child if you’ll never have any children to begin with. Edit: Nevermind, looks like it's just a way to hype then!
    • zer00eyz 2y ago
      You got this backwards. M$ gets everything OpenAi does till they reach AGI. The deal is "give me everything till you have that baby"
    • kaast202 2y ago
      that does not make sense,the deal is msft gets everything except AGI. They do believe they ll reach AGI whether thats pipe dream or not is a different story because they believe in agi they are fine with giving up everything upto that point