20250719 - Different Intelligence...
Take a typical well accomplished doctor as a professional (in India as an example)...
People have to finish
~ 4-6 years of childhood prior to schooling
~ 13 years of schooling (K-12)
~ 5.5 years of M.B.B.S
~ 3 years of specialisation
~ 3+ years of super specialisation
And a few years of practice afterwards, they will be considered a well experienced and accomplished doctor.
Roughly speaking at least 30-35 years. And they probably have to have had a very good track record all through.
A Doctor was just an example and is on the higher side. Take any other profession and to be a good/ accomplished professional, it will take at least 25-30 years, after being born.
Point is, none of us (humans) are born professionals. Sure, some may have an edge because of their genetics of being a little more "intelligent" or having a little more "memory capacity", etc. However, everyone is born with these capacities with a "clean canvas".
None of the learning and experience/ accomplishments percolates to our DNA and trickles down to our offsprings. They are born with their "clean canvas" and have to start drawing/ painting from scratch.
That's the unfortunate reality of "Natural intelligence". None of the human accomplishments that have been made, have made it to the genetics/ instincts / subconscious and each person has to start from scratch.
It is one of the main things limiting and holding us back on every front. A 40 year accomplished specialist surgeon having a child doesn't mean that child can progress with the 40 years of knowledge and can further the field with ground breaking work. They still have to invest the first 40 years at least to get to the same point and afterwards - well, everyone has the same amount of productive time left in their life to further anything...
I have been using "Artificial intelligence" tools (LLMs) for some time now. Free versions mainly.
I recently got my hands on the Pro version of one of them. I am flooded with what it is capable of.
These tools are in an infant state broadly speaking (it has been only a decade and half of intense work that has gone into developing them)
They say AGI ("Artificial General Intelligence") is 5-10 years away...
Our human limitation I described above has already started limiting us/ holding us back which many have not realised.
E.g. #1 - Take a software professional (most immediately impacted & felt as well). AI has started replacing many "low level" software developer jobs. But not to worry, more senior/ specialist software designer/architect jobs will be there. But how do you fill them in some years? We need folks to start at the "low level" rungs of the ladder and work their way up to the specialist roles. but the junior jobs and folks will not exist anymore... So, how will this be met?
E.g. #2 - It has been proven now that AI can read/analyse an X-Ray much better and faster than humans. So no need for junior radiologists anymore.
But, some complications need a senior/ specialist Radiologist to have a look to ascertain the findings. So, not to worry, the senior/ specialist jobs will be around.
But how will we fill it in a few years? there won't be any junior radiologist to turn senior/ specialist...
It got me thinking, if an AGI controls/ powers/ runs a robot manufacturing facility, when it makes a new robot (a robot being born), it will load it with the latest "Artificial Intelligence" available at that time.
A new robot will not be born with a "Clean Canvas".. it will have all the intelligence of its predecessor (parent robot) and can/ will continue to make ground breaking progress from day-1
That's something we are not aware/ familiar with and can fathom.
So, I wonder - what will we (humans) be needed for by AGI?
Sure, we keep saying that we have emotions, feelings, consciousness, etc., etc. Qualities humans recognise and need from each other.
But really, what will AGI need those things for in its world ...?
What a perspective I am more worried for the future generations.
ReplyDeleteIt's true that AI capabilities are improving by the day, literally. However, my gut says that there are still two limits that they cannot cross: the Godel limit -(any deterministic system contains statements that it cannot confirm or deny the truth of), and "garbage in, garbage out". Any AI is still pattern matching, and while there's no reason to assume our own brains don't do the same, there is an inherent limit. And quantum computers will only speed things up, not make them more intelligent
ReplyDeleteSeparately, we've already been through multiple cycles where humans have had to reinvent themselves to be gainfully employeed. This is just the software guys turn. I don't know how, but I'm hopeful that we'll be ok
Yuval Noah Harari has explored this in his book Homo Deus. A good read
DeleteOh yaa, what a thought sir, this means that we may have only theoretical knowledge… and that too taught by AI systems and what for? Only to repair/invent more of these AI systems. Why would an AI teach a human to repair/create itself. Considering that eventually these AI systems learn to dominate, it may try and extinguish the human race feeling that humans are a threat. This might sound destructive but its The law of nature “survival of the fittest”, AI can and will learn this law. I would love to be proven wrong though
ReplyDelete