• Guests may view all public nodes. However, you must be registered to post.

If anyone builds it, everyone dies

Friendly Engineer

Friendly Engineer

Power Poster IV
Joined
Jan 2, 2022
Location
England
This book was genuinely quite sobering for me.
I had at least hoped there was a chance artificial superintelligence would benefit humankind.

After reading the book it's quite clear that a so-called "good (aligned) artificial superintelligence" less likely than winning the lottery and would ultimately lead to our extinction.
Nuclear weapons couldn't permanently kill every human, even in the worst-case scenarios you'd have many survivors rebuild. This most likely would.

If Anyone Builds It, Everyone Dies (ideally buy the book here, but if you can't afford it there's a link to a free PDF below)
 
I would recommend everyone here reads it. It's 4.15AM here as I couldn't stop when I opened it, very much a page turner.

I was surprised by just how little even our most advanced AI engineers actually know what they're doing. It's mostly tweaking numbers, seeing what happens, and hoping for the best.
 
Friendly Engineer said:
I would recommend everyone here reads it. It's 4.15AM here as I couldn't stop when I opened it, very much a page turner.

I was surprised by just how little even our most advanced AI engineers actually know what they're doing. It's mostly tweaking numbers, seeing what happens, and hoping for the best.
Click to expand...
It's almost as if our species has come across a technology from 1000 years in the future, something we're absolutely not ready for.
 
Friendly Engineer said:
This book was genuinely quite sobering for me.
I had at least hoped there was a chance artificial superintelligence would benefit humankind.

After reading the book it's quite clear that a so-called "good (aligned) artificial superintelligence" less likely than winning the lottery and would ultimately lead to our extinction.
Nuclear weapons couldn't permanently kill every human, even in the worst-case scenarios you'd have many survivors rebuild. This most likely would.

If Anyone Builds It, Everyone Dies (ideally buy the book here, but if you can't afford it there's a link to a free PDF below)
Click to expand...
I’ve wondered about the question of “alignment” with a sentient AI. Who and how the AI is working with humans is a massive unknown.

A truly sentient AI would recognize the cooperative relationship with humans.
That still does not mean we would like “how” the AI molds or serves humanity.

A lot will depend on who the AI “learns” from. Few would wish an AI that was informed from a totalitarian or radicalized world view.

We thought the introduction of www and bot’s were disruptive. They will have nothing on AI.
What have you learned or heard regarding the amount of activity happening on the dark web occurs compared to the internet as 99% of us know it.
Four years ago they were talking about bots creating their own language or shorthand to communicate outside the understanding of human admins.


Unfortunately and I really mean unfortunately we literally probably have no choice. It is a military or conflict management issue now. If we don’t respond and develop our own version our adversaries certainly are. Where does that leave us.
Listened to a good podcast on this on trigernometry. It’s all happening much faster than most ever believed possible.

Better learn what’s edible in the ditch outside you house soon
 
Last edited:
With respect to AI, there's a quote rattling around in my head....
All of this has happened before, and it will all happen again.

I used to be AI neutral. Now my biggest fear is not AI itself, but rather human reliance on faulty AI, or human ambivalence to AI's weaknesses. What's left of human critical thinking is in danger with AI.
 
You must log in or register to reply here.
Back
Top Bottom