As far as I know, this is an extrapolation from the number of possible synapse connections in connection with superimposed usage possibilities, which result from the fact that a single synapse does not only represent binary 0/1, but can probably assume a whole range of different states. (e.g. https://elifesciences.org/articles/10778).

How did you arrive at your estimate?

Gilbert Brands.

Thank you for your reply and for the reference to the elife article. For others, the full cite is:

Thomas M Bartol Jr, Cailey Bromer, Justin Kinney, Michael A Chirillo, Jennifer N Bourne, Kristen M Harris, Terrence J Sejnowski (2015) Nanoconnectomic upper bound on the variability of synaptic plasticity eLife 4:e10778

The Bartol et al 2015 article is discussed at:

https://www.scientificamerican.com/article/new-estimate-boosts-the-human-brain-s-memory-capacity-10-fold/

The Bartol et al article does not give a number. But based on Bartol et al, supposing 100 billion neurons, each with 1000 synapses, each with capacity 4.7 bits, that would give 54,715 GB, which is about 54 TB.

An earlier reference is, based on an answer to a Scientifc American question in its May 2010 issue:

What is the memory capacity of the human brain? Is there a physical limit to the amount of information it can store? 2010-05-01 Scientific American Mind , May/June 2010

Paul Reber gave memory capacity as 2.5 petabytes.

2.5 petabytes is about 2.5 million GB.

Maybe I have a math error? The 2015 Bartol et al based estimate is much smaller than the 2010 Reber 2.5 petabyte estimate.

Both approaches are based on brain anatomy. My approach is much different, but before proceeding with it, I am trying to figure out the 2.5 petabyte estimate. What is it based on?

I think the sheer capacity is quite overrated. Ultimately, what is stored is what has proven to be useful in evolution.

The associative access techniques to the information are probably much more important. When thinking about complex processes, helpful associations from completely different fields often come to mind. I don't think that something like that is even being considered in AI techniques at the moment.

Gilbert Brands.

Thank you for your observation earlier today (December 22), very helpful, since even the Bartol estimate seems high to me, let alone the 2.5 petabyte estimate.

The 2010 article itself observes "2.5 petabytes would be enough to hold three million hours of TV shows. You would have to leave the TV running continuously for more than 300 years to use up all that storage".

While presented as an illustration of how big 2.5 petabytes is, it seems more of an argument contrary to the 2.5 petabyte estimate. A person living 100 years could not physically absorb, much less retain, information that takes 300 years, without interrruption (no sleeping), to transmit.

Another heuristic is to compare the information stored in a computer's HDD or SSD, and how much of it might be stored available for ready retrieval in a person's brain.

Regards

Maybe you should look for studies on people with photographic memories and how far back their memories go unimpaired.

Gilbert Brands.

Thank you for your kind help on this question.

Today (Dec 24), I posted an initial attempt at quantifying human knowledge capacity on RG:

Preprint Knowledge, in GB, per average human brain

In case you might be interested.

Regards

The article posted yesterday (Dec 24) , Preprint Knowledge, in GB, per average human brain

on brain bit capacity might leave a door slightly ajar to new problems, following up on this question.

As an example of the kinds of problems that might be raised:

Suppose a person can add to their existing knowledge at the rate of 5.6% per thousand years, at the rate of 10 bits per second. In effect, a human can fill up their 581 GB capacity at too slow a rate to ever fully utilize it.

If the cumulative capacity of 1,000 years of bit processing is 581 GB, then a human would never come close to the 581 GB theoretical 1,000 year capacity in a human's lifespan, which is much shorter than 1,000 years .

How is human progress possible?

Conjoining bits per second processing and percentage increases in knowledge may raise questions relevant to computation, AI, epistemology, brain studies, theories of intelligence, philosophy generally and so on.

"How is human progress possible?"

To be a little satirical: considering what both politicians and their voters are doing, wouldn't it be more appropriate to estimate brain capacity to be on the order of old 8" floppy disks?

Gilbert Brands,

Yes, it is a bit satirical. But underneath the satire, my 2009 paper ( Article A theory of intelligence: networked problem solving in anima...

“There was more difference between the problem solving capacity multiplier of English speaking society in 1989, considered as a single mind, and the problem solving capacity multiplier of a primitive society of 150 people with only 100 words, than there is between the brain of an average human being and the nervous system of the worm, C. elegans. This third set of data is consistent with what we know about human nature and human capacities. ...”

Humans have enormously more problem solving capacity collectively than they do individually. Which seems to justify your light-hearted remark.

That's probably true.

I remember a thermodynamics lecture in which the lecturer conclusively demonstrated that no useful work can be expected from a committee (a lot of enthalpy, but even more entropy).

My own investigations showed the same thing: https://www.researchgate.net/profile/Gilbert-Brands/publications (4 books on the topic "Mathematics of Opinions")