The "Magic" Problem: There is a growing concern that AI may soon produce scientific breakthroughs (like fusion or new drugs) that are incomprehensible to human researchers, making the world feel like it's run by "magic"
Timeline
Post
Remote status
Context
6
(how about curing all diseases?)
The "Magic" Problem: There is a growing concern that AI may soon produce scientific breakthroughs (like fusion or new drugs) that are incomprehensible to human researchers, making the world feel like it's run by "magic"
The "Magic" Problem: There is a growing concern that AI may soon produce scientific breakthroughs (like fusion or new drugs) that are incomprehensible to human researchers, making the world feel like it's run by "magic"
Most major diseases we know of *should* be cured by now, but research is wildly inefficient due to "scientists" who one day have an idea and then make it their life's work to discredit anyone who disagrees with it - rather than working the evidence and trying to find better and better models.
When's the last time a major scientific consensus was just ... overturned, because someone discovered an experiment that showed it's wrong?
When's the last time they had to go back and re-write text books?
The inventor of freaking handwashing was thrown in a mental institution because MeDiCaL sCiEnTiStS thought rubbing your hands in water is stupid superstition and he needs to STUF. Nothing has really changed since then.
The ONE exception is computer science, and that's because everything is trivially reproduced by ... sending people your code ... so there's just no way that deniers can gain any traction when *everyone* can see what you see.
When's the last time they had to go back and re-write text books?
The inventor of freaking handwashing was thrown in a mental institution because MeDiCaL sCiEnTiStS thought rubbing your hands in water is stupid superstition and he needs to STUF. Nothing has really changed since then.
The ONE exception is computer science, and that's because everything is trivially reproduced by ... sending people your code ... so there's just no way that deniers can gain any traction when *everyone* can see what you see.
> happens all the time
> doesn't provide a single example
:slow_nod:
> doesn't provide a single example
:slow_nod:
Replies
1
Don't worry, it was wrong anyway.
Have a beer, happy Friday 🍻
Have a beer, happy Friday 🍻