LLMs are bad at novelty, but that is our chance to Singularity
3 comments
> still one area where it is notoriously bad is creation of novel things
Is there any proof of this? My feeling is that people just assume that since something has been generated by AI, then it must not be original. And truly original thought or craft is exceedingly rare if it exists at all: everything comes from something else. Individual innovations are often small.
> Because that is the only loophole in LLMs that can not be fixed
This is magical thinking. Humans don't have a small device in their head capable of "true creativity" whatever that even is. At most, they experience so much random input that their outputs are less predictable. But nothing that LLMs won't be able to do soon; as for raw intelligence, they're already smarter than the large majority of the population.
Is there any proof of this? My feeling is that people just assume that since something has been generated by AI, then it must not be original. And truly original thought or craft is exceedingly rare if it exists at all: everything comes from something else. Individual innovations are often small.
> Because that is the only loophole in LLMs that can not be fixed
This is magical thinking. Humans don't have a small device in their head capable of "true creativity" whatever that even is. At most, they experience so much random input that their outputs are less predictable. But nothing that LLMs won't be able to do soon; as for raw intelligence, they're already smarter than the large majority of the population.
Is there any proof of this?
- ARC-AGI challenge
- ARC-AGI challenge
Is that specifically about originality? Or is it about other aspects of intelligence and cognition that LLMs can't yet reproduce at human levels?
> "In most books, putting out of account those that are thoroughly bad, the author, when their content is not altogether empirical, has certainly thought but not perceived; he has written from reflection, not from intuition, and it is this that makes them commonplace and tedious. For what the author has thought could always have been thought by the reader also, if he had taken the same trouble; indeed it consists simply of intelligent thought, full exposition of what is implicite contained in the theme. But no actually new knowledge comes in this way into the world; this is only created in the moment of perception, of direct comprehension of a new." - Arthur Schopenhauer
I can envision a future where people are freed from doing repeatable jobs and 99% are shifted into enterprises of creating something new.
- New life saving medicine - New life extension medicine - New art work - New novel - New product on the market - New engineering breakthrough - New scientific knowledge
Event though today only miniscule percentage of people are working in those kind of jobs it will very fast become dominant industry. Because that is the only loophole in LLMs that can not be fixed but is highly valuable and people with money will be ready to pay obscene amounts of money to extend their life. Who knows maybe there will be whole new big companies with sole purpose of trying to extend the life of Elon Musk or Jeff Bezos.
> I work for Musk foundation and my job is to extend the life of Elon Musk - anonymous employee
Jokes aside, imagine the gigantic potential we can unlock if we transfer billions of people on a quest to unlock novelty. For millions of years people were working on repeatable jobs and our brain is not able to comprehend that there can be any other job out there that does not contain "skill", "productivity", "seniority" which are just another words for someone who mastered the art of repeatability which is not something humans should brag about.
I am saying that massive creation of novelty can cause double digit rates of GDP growth or even unlimited rates which means singularity. But it will not come from AI, but from people on a mission to discover something new.