>>10958415>>10958429Kurzweil still optimistically thinks that we can implement AGI with the use of hacks and brute force methods. (or maybe he's just pretending to be 100% optimisitic in order to save face)
> hacksMany prominent AI researchers think that the only way to implement AGI is to create artificial humans. You cannot hack it away with GOFAI, ANN, or mathematical logic.
It might be that there's no shortcut to this thing.
> brute force methodsTo create artificial humans, you need molecular engineering.
To simulate virtual physics, virtual earth, and virtual humans, you need nuclear fusion to supply an enormous amount of energy.
Many of our problems *requires* an AGI (e.g., story understanding, problems that require a conscious experience).
This is why the bot from the movie "Her", or a bot that can learn all information from the internet is still a pipe dream. It looks so simple but actually it's so fucking hard.
Many AI researchers are starting to lose faith (again) with the AGI idea, and now they cope with "muh emergence of modular devices" kind of intelligence, or stats and probability.
Moore's law is dead. Quantum computing is a meme. And artificial molecular engineering and nuclear fusion might be a failure of imagination.
It does not matter if "The Law of Accelerating Returns" is working if most it returns are garbage and duplicated data. (See the current state of academic papers)
Most of the new tech we produce are solutions waiting for a problem.
Earth is a limited resource. The average human is financially and psychologically broke.
Instant mass communication is making everyone dumber.
Capitalism and Communism have very bad symbiosis with modern technology.
And it's easier to be degenerate than disciplined with technology as time goes by.
Collapse is more likely to happen first than Singularity.
