Published On Apr 25, 2024
AGI or artificial general intelligence will be transformative for society. However, it is powerful enough that indiscriminate access could result in dangerous outcomes. One of the biggest risks of developing advanced AI is that human operators will use the power for malicious ends. When individuals possess powerful technologies such as weapons, the state has to ban them when they get powerful enough.
Societies themselves, i.e. nation states, will also have to adapt to the advent of AGI. The balance of power between countries used to rest on military strength, now more recently on mutually assured destruction (nuclear weapons). But the availability of cyberspace attacks, especially with AI added to the mix, may destabilize the status quo.
We discuss several solutions to these two problems, including having a singular world government or having humans merge with machines. Although AGI will bring massive good to the world, its introduction has to be navigated carefully to avoid bad consequences for humanity as a whole.
Discord: / discord
Patreon: / drwaku
#ai #aisafety #geopolitics #agi
Open Foundation Models: Implications of Contemporary Artificial Intelligence
https://cset.georgetown.edu/article/o...
Alternatives to mutual assured destruction
https://www.britannica.com/topic/nucl...
Why did the "Anglo-Saxon" society develop to be so individualistic?
https://www.quora.com/Why-did-the-Ang...
WTF is Will To Power?
/ wtf-is-will-to-power
0:00 Intro
0:15 Contents
0:25 Part 1: Can individuals handle AGI?
1:00 Effects of AI at individual level
1:13 Each person responsible for their own safety
1:36 Example: duels in England
1:57 Example: Samurai in Japan
2:27 Is it acceptable to carry weapons?
3:07 How AGI can be weaponized
3:46 Example: spammer making phone calls
4:06 Can society afford to give AGI to everyone?
4:15 Option 1: align models to make them safe
4:24 Option 2: funnel all actions through authority
4:34 Option 3: possible fallout is acceptable
4:44 Option 4: restrict public access
4:54 Tie AI actions to human moral agents
5:18 Part 2: The role of the state
5:47 History of colonial powers
6:31 The nuclear era
7:04 AI similarities with nuclear weapons
7:38 Missile defense project
7:55 Evolution from small to large scale defense
8:21 The cyber era
8:55 Cyber attacks occur all the time
9:26 Can we tie AI to hard assets?
9:57 AI will be used frequently by military
10:43 Part 3: New paradigms for society
11:20 Can we keep powerful AI out of people's hands?
11:46 Balance of power between nation states
12:30 Proposed solutions
12:38 Resolution 1: advances in AI safety
13:05 Resolution 2: totalitarian control over models
13:39 Resolution 3: singular world government
14:20 Resolution 4: humans merge with machines
14:50 Resolution 5: welcome overlords
15:00 Additional thoughts
15:15 Conclusion
16:32 Join discord for voice calls
16:42 Outro