Do you ever think that the language reasoning models use to think (CoT) makes a difference in their performance? I.e. would a model thinking in a more evolved language (eg Greek, Chinese) be better than models thinking in English?
Languages evolve like the species. So the assumption here is that an older language has gone through more generations and evolution than a relatively newer language.
The history of COVID taught us two things:
- Populist governments won't listen to scientists.
- Governments won't collaborate, but rather try to fool international agencies where possible.
https://t.co/mjVH1xGrGE
"And what I claim will happen is that people will start to act in unprecedentedly collaborative ways out of their own self-interest."
This TED talk was post-COVID right?
What could be said about healthy eating and exercise for 30-something year olds today?
I mean, either superintelligence will kill us all before aging will. Or it will produce foods and drugs that will make us super healthy in our 50s/60s?
That's because to do so, ASI would need to think and feel like a human. And while ASI can simulate human thinking, it may not be as good as stimulating human emotions.
You know how everyone says that "taste" will be the last thing ASI will acquire? And by that we mean choosing what problems to solve, what products to build, etc.
Once ASI knows what to build, it will do so better than humans. It will invent new materials if it has to.
Resistance training for your brain: avoid content that you know will make you feel good because you agree. Seek content you fundamentally disagree with. Do it regularly.