(Fine Tuned DNA)
I came to the conclusion
that thinking really BIG is not about
changing the world, but ruling the universe.
One year after I had the idea of TranscenDeNtAl (transcendental-DNA) I
decided to put together (the edited versions of) what I previously wrote on my
blogs and different (online discussion) sites – such as the one HERE -- in one place and rename the concept as Fine Tuned DNA (FiT-DNA).
I believe that FiT-DNA is (a) the most intriguing (and, most
likely, the most important) research idea in the history of DNA-based
intelligence and thus (b) the only serious challenge to the highly popularized
and heavily financed Singularity concept that Google decided to pursue.
Briefly about the 2 concepts:
(I) FiT-DNA:
(it starts with what I wrote one year ago)
After watching the glorious idea of the song for π (which could go by the name 'the pi in piano') it hit me: how about the DNA
of (i.e. generated by the sequence of numbers in) π, e and,
eventually, other famous transcendental
numbers*?
I know, the DNA plays with 4 letters and 10 is not a
multiple of 4. Of course we can be creative and find plenty of ways around
that. Quite handy, for instance, would be to use the Base 4. To each digit (out
of 0 – 1 – 2 and 3) we assign a letter. There will be 4! = 24 possible
combinations.
Now I’m not saying that we’ll be able to create some
impossible entities such as, say, an honest politician, but could any (even
only partially) meaningful DNA sequencing pop up? Of course that a combination of the
meaningful DNA sequences of different transcendental numbers is also to be
considered (remember, the intriguing e^π is
a combination of
two transcendental numbers).
This reverse engineering research, if intelligently done,
can’t but spectacularly advance our knowledge. There’s an infinite set of
numbers, DNA-wise, that make sense, and this is not some hypothesis or theory,
but a FACT: just take into account the trillion of trillion of … trillion of
life forms/living entities in the history of this planet, AND the ones that
come to life every second. And counting …
Common sense also tells us that, at least within our
universe, there’s an even larger infinite of
numbers that would lead nowhere, life/intelligence-wise.
So where do we start? (A) Choose the lottery numbers for the
last 30 years and see what, if anything, comes up? (B) Starting from a dubious
(to say the least) premise** like in this story (and, hey, their research got funded!)? Or
(C) Looking for some particularly intriguing AND FUNDAMENTAL
numbers? Remember, we’re trying to create life, the more intelligent, the
better. The Fibonacci series, so much associated with life, should also
be taken into account.
And speaking about life it’s finally a good time now to wise
up and remember the fine-tuning of the universe issue: Plenty
of glorious fundamental constants to play with.
It’s obvious that if we could create an intelligent
machine/entity able to create a smarter machine/entity we’ll be on our
(asymptotically) way towards absolute knowledge. From that point on we just
have to relax and enjoy the (snowballing) show, we won’t need to ever invent or
create anything else. So how about we start with π, just to warm up …
… Although if I had the funding, the team and a state of the
art lab I would check “n” fundamental numbers in parallel and constantly
compare the results. I’d also check – within the resources and common sense --
if any mixing would lead somewhere. Remember: There are infinite combinations of
numbers that DO lead somewhere, but we have to intelligently design a research
method to improve the probability of hitting this infinite, rather than
the larger infinite leading nowhere.
**I’m sure that my proposal is based on a by far more solid premise than
the Moore’s Law: The fundamental constants/numbers are here, to rule our
universe, for more than one reason (or, should I say, for ALL the reasons that,
so far, we’ve been too blind, or inept, to figure out?)
From now on this is what I’ve
added on March 30, 2014:
What FiT-DNA is doing is to bring at the same table the 2 reasons life
(AND intelligent life), as we know them, developed on our planet/in our known
universe: the fine tuning of the universe and the DNA. They can be seen as
being apples and oranges OR as making perfect sense together (life-wise); I bet
on the latter horse.
The idea of the FiT-DNA is that in our fine tuned universe
(that just happened, probably among other infinite number of universes) life
started in a very primitive way, with a scrambled DNA. Think of it: if the DNA
was so “perfect” (from the very beginning), as the fundamental
numbers/constants, then it would have been no need for life to evolve ( is π .. evolving by changing
its digits?), through, mainly, DNA “accidents.” There were bad, neutral and
good “accidents”; I call the good ones the unscrambling process, which, so far,
only happened by chance. The educated unscrambling
process I’m proposing means editing the DNA to the point where it rhymes with
the fundamental numbers that rule our universe AND made life possible. We can
wait 2,000,000 years for some lucky DNA accident to happen, or we can design
“the accident” ourselves.
How do we accelerate the unscrambling of the DNA? We look
for sequences (of at least a certain length, to be meaningful) that are the
scrambled digits of a fundamental number and then edit that sequence, i.e. we
just rewrite that sequence to be like the one in the fundamental number. Then
see if anything positive happens and learn from our experiments. Just a note:
the ‘junk DNA’ gives us plenty of room to play, BUT I’m NOT talking ‘junk DNA’
ONLY!
The FiT-DNA research should look for finding the MML – the
Minimum Meaningful Length (of a DNA sequence) that, when unscrambled, will lead
to a significant improvement in the subject tested-- and can probably be done
in 4 steps:
(1) The Mighty Potato phase (any appropriate plant would
qualify):
When the potato can sing like Pavarotti while moonwalking
like Michael we’ve got the plants’ MML;
(2) The bitchy Parrot:
Parrots can talk, so it’ll be really easy for the
researchers to figure out when the bird will actually make sense (i.e., reason).
For instance if when asked, “Who was the 43rd Prez of the US?" the
parrot will answer, “Mr. Nukular?” we’re ready for step 3:
(How about we just
skip the mammals? Hey, life is short ..)
(3) Primates (there are probably laws that won’t allow us to
start with the US Congress directly …);
(4) Humans.
It’s plain common sense, it’s how life works. Look below at
the same plant in different climate conditions and you can clearly picture the
transition human-transhuman-SuEn, by replacing “different climate conditions”
with ‘different levels of edited DNA.’
Of course that I’m not talking here about we, humans, becoming
bigger (like the plant above), but way smarter. We should be (at least) 100X smaller (how about 314,159X smaller? -- of course that we'll figure out the optimum size) and
consuming (at least) 200X less of the traditional resources, if any at all, and planet
Earth should be our unpolluted natural reservation.
(II) The Singularity
It’s too much online – you can start by reading THIS -- about it to waste your time with my version about the AI outsmarting
humans in 15 years. I’m absolutely fascinated by the AI’s potential but I just
believe that FiT-DNA will get results faster.
Just a few more things, in no particular order:
Trying to mechanically replicate life resulted in too much
fun: humans can’t fly like birds (remember THIS?), and birds (in
their birdbrain wisdom) don’t fly like an F-16, rocket or copter. Same story
underwater: subs don’t ‘swim’ like the marine life. On land Joe doesn’t run
like cheetahs and Bill doesn’t jump like grasshoppers. So why are humans – as of
today being at the ridiculous level of walking bags of bacteria and excrement -- still looking down and trying to make electro-mechanical copies of inferior
entities? Look at this Big Dog here (and there are countless other examples): these toys are a
good hobby for high schoolers, not for the top scientists paid with top money
to design, in expensive state of the art labs, some ‘pets’ with 1,000X less
abilities (like in running, jumping, changing direction, vision, hearing, etc)
than a 2 weeks old puppy. Misdirected resources!
Why FiT-DNA is way better suited to hit the Jackpot* way
before the AI
*“The
Jackpot” I’m talking about means creating a Supreme Entity (SuEn), an
entity/super-human asymptotically approaching the Absolute Knowledge.
Creativity-wise, thus far there has never been any creative
AI, while for tens of thousands of years the DNA-based-Intelligence (let’s call
it DNAI) has been, from the first tools and paintings back in the stone age to
the Opportunity roaming the Red Planet and the latest fancy app for a smart
phone.
The most logical way to design a SuEn is to first create a human with a probably 10-15% (this has to be determined) higher IQ than the top
genius IQ that we are aware of; let’s call this dude a transhuman, so I won’t
have to keep inventing new names. Now this transhuman, being sufficiently
smarter than us, should be – it HAS to be, that’s why we designed him! – able
to design an entity able to create a smarter entity (sounds like an
intellectual perpetuum mobile? Hell of a challenge, I’m all for it!). And
that’s it, we don’t have to create/invent anything else from that moment on. We’ll just
be gods the very next minute, laying back between two supermodels and watching
the Singularity dudes scratching their heads.
Why challenge Google? Remember the famous rivalries, Dostoevsky
vs. Tolstoy, Tesla vs. Edison, Galileo vs. the church, Einstein vs. Bohr and so
on? They were just glorious and, unless all you care about is boobs and the
Monday night game, you HAVE to appreciate a hell of a challenge.
And when it comes to ‘assembling gods, one letter at a time’ what could
possibly be more entertaining than this intellectual challenge: FiT-DNA vs.
Singularity?
No comments:
Post a Comment