re: OFFICIAL: Sam Altman to return as Open AI CEO w/New board

quote:

---

The Ilya side wants to decelerate and do more "studies" to make sure its all "safe". To make sure it doesn't hurt feelings and mirrors Ilya's sense of morality.

Ilya would hold back world changing technologies that could save thousands of lives just to make AI a little "less racist" or less anti-LGBTQ.

One sees AI as a tool to benefit the world, the others want some sort of god-like being.

---

Let me preface this by saying I don’t know enough about the personalities involved to address your portrayal of these specific individuals.

That being said, I’m a little surprised to see the entire concept of “AI safety” being mocked in this thread. AI safety goes way beyond SJW concerns. Once you hit AGI, ASI becomes a near certainty. And when you’re dealing with a computer system that’s way smarter than any human, you better make damn sure you programmed its goals correctly.

It just feels a little disingenuous to lump all of the serious researchers working on AI safety (who, BTW, are few and far between compared to the people working on making more advanced AIs) in with folks who are just worried about offending people.

quote:

---

That being said, I’m a little surprised to see the entire concept of “AI safety” being mocked in this thread. AI safety goes way beyond SJW concerns. Once you hit AGI, ASI becomes a near certainty. And when you’re dealing with a computer system that’s way smarter than any human, you better make damn sure you programmed its goals correctly.

---

Literally ZERO evidence to support this is a realistic outcome. Comes off as someone who watched the terminator one too many times. It's pure scifi doomerism and this doomer thought is enabling bad actors to seize control of AI progress under the veil of "ethics concerns". Never-mind there's absolutely nothing about them that makes them more ethical other than a perceived superior ethics understanding based on their academic background.

Its another inside-outside game where activists on the inside are causing trouble to steer the power of the technology into the hands of centralizing bureaucrats who think they have to dictate how the power is used. Same people that constantly "fear" the end of democracy.

quote:

---

Literally ZERO evidence to support this is a realistic outcome. Comes off as someone who watched the terminator one too many times.

---

Zero evidence to support what as a realistic outcome? ASI? Because that’s the only actual outcome I suggested.

Saying AGI will almost certainly lead to ASI makes me sound like someone who has watched Terminator one too many times? Come on, man.

quote:

---

That being said, I’m a little surprised to see the entire concept of “AI safety” being mocked in this thread. AI safety goes way beyond SJW concerns. Once you hit AGI, ASI becomes a near certainty. And when you’re dealing with a computer system that’s way smarter than any human, you better make damn sure you programmed its goals correctly.

---

I personally think it's a very legitimate concern. Not necessarily that AI will eventually become fully sentient, but that it almost certainly will be used at some point for nefarious purposes if some safeguards are not considered. Infosec researchers are already saying that phishing e-mails could get very difficult to identify if the usual hallmarks (grammar mistakes, misspellings, language translation semantics, etc) are made irrelevant by generative AI.