I feel like snowflakes are memory structures -- the progression of shape starting from the center would reflect humidity/air pressure/temperature over time.
Does that make sense to anyone? Why couldn't you encode memory in this?
It would certainly make for an interesting sci-fi story. Imagine if the large snowfall on the east coast were a higher being providing us instructions on how to fix our planet, if only we knew how to reassemble the information. And the planet's glaciers contain massive amounts of information on how to create a wormhole to visit other civilizations, but now they're melting and the information is being lost forever...
Most "junk" DNA is not actually junk. There are several different parts of DNA.
The most easily understandable parts are the gene coding bits, we know exactly how that works (in isolation), the DNA gets transcribed to RNA which then gets translated into an amino-acid chain that becomes a protein. If you label everything that's not a protein encoding region of DNA as "junk" then you really don't understand DNA at all.
Equally important are the regulation parts of DNA. These are sections of DNA before and after genes which are there to help cells control the expression of different genes.
Additionally, there are big chunks of DNA which exist solely to ensure the DNA has the right chemical properties and is easily stored in a cell.
There's still plenty of "useless" parts of DNA once you account for all that, but much less "junk" than previously believed.
My previous comment was pretty tongue in cheek, but thank you for the explanation, that's really quite interesting. I wonder if it is possible that the 'useless' or 'junk' DNA is mostly stuff that becomes important under different environmental conditions and becomes active based on certain triggers with epigenetics?
I used to agree with you but I think nearly all scientists agree that the long repetitive regions (like LTRs) that compose ~75% of the non-protein-coding genome are actually and truly non-functional garbage that just replicates opportunistically and evolves neutrally.
There was a big argument about this a few years ago but ENCODE finally backed off on their public statements about how much of the genome was "functional".
Although the experiment would be roughly impossible, you could in principle replace all the LTR regions with random sequence and there would be no positive or negative effect.
Note that all of the above may be wrong, but it's really hard to argue that the viral replication stuff in our genome has a functional effect.
> Just get a bunch of monkeys pounding on keyboards instead. They'll produce the same information in due time.
Have you seen the Library of Babel[1]?
It contains every book which has ever been written, and every book which ever could be written, for it contains every combination of 1,312,000 characters that are possible.
It's a bit difficult to find any of that in all the noise, though…
ha, unsure why you are getting downvoted clearly some people deny their imagination
any deterministic process can 'encode memory', just look at compression,
so if you have an initial state that you know will produce a final result, like this technology is showing can be done with snowflakes then you can encode the final result as a piece of memory and 'store' that value as the initial state and processing requirements
but you are subject to working backwards, determining the memory from the process
so, could you use this snowflake process to encode arbitrary memory as an initial state and process? it is unknown because this article fails to speak to whether the tech can design the snowflake to an acceptable precision
Does that make sense to anyone? Why couldn't you encode memory in this?