There’s a computational theory of mind
Maybe it’s right maybe it’s wrong
It’s cute, philosophically
But isn’t it profound practically?
Your results are downstream from your actions,
which are downstream from your mind
Isn’t your mind quite like a computer?
We speculated about our nature since time immemorial
But aren’t computers some of the best MODELS of human mind that we have?
There’s hardware,
There’s software,
There’s input and output
Hardware is both CPU and memory
Memory itself is both CPU’s cache, random access memory, and then the disk, which is supposed to retain information
Software is a set of instructions
But also a base of assumptions, facts, necessary for the realisation of those instructions
Which are all, in turn, stored in memory
The simplest and very useful analogy is that of software-hardware combinations
Hardware without software is useless
Hardware without inputs resulting in outputs is useless
But software without hardware is useless too
And then the various less extreme combinations:
weak hardware and strong software can be quite powerful
weak software but strong hardware can also be powerful… maybe
Now, to the practical part,
I believe everyone, obviously, should seek to optimise their minds,
Which means optimising software, AND hardware, and all the components
More than that, one should always look at it in detail, not vaguely
A given problem likely has some optimal solution
So what SOFTWARE do you need to solve that problem?
What DATA do you need?
Do you have to put in your RAM, or your disk, or both? How fast do you need to access it, to perform? How long should it hold in your cache?
Is your hardware even capable of handling this level of computation? Know yourself, know your talents
Can your hardware be compensated by software?
And of course trash in trash out
Curate what you put in that fucking head
And of course not every solution is as elegant,
Not every solution is as efficient
You want to save your computational resources, by using more efficient and effective algorithms
Then there’s the AI
Much of the current AI boom is powered by neural systems performing statistical pattern completion
which is quite analogous to human intuition, actually
Intuition developed through vast experience (data),
It thus intuits the answer, having seen it so many times, or having seen something similar so many times
And therefore it sometimes hallucinates, confabulates, bullshits
But deterministic, robotic, algorithmic — it CALCULATES, based on fixed rules
The predicates are not shifting
For the same input you always receive the same output, unless you messed up, the fallible human
Know when you need to develop intuition, and when you need to develop RIGOUR
Know when you move from one mode to another
Know how to blend the two
(BTW: this is one thing AI can’t yet do: fluidly move between the two paradigms, in a creative and consistent way)
Thus
CLEARLY we’re QUITE like computers
So we should more often approach ourselves like computers,
And ask the meta-question of HOW to attack the main question
Not HOW to do it — but HOW to HOW-to
Indeed we underestimate the software part
I believe IQ differences can be to a meaningful degree explained by different software run. IQ itself reflects performance on a certain set of information-processing problems, not just the hardware on which that performance runs.
And I believe a lot of superiority of certain thinkers is not merely their superior, faster hardware — but the RESULTING SOFTWARE they developed
i.e. if you could copy their software, their mindset and rules and ideas — often you’d be better off cognitively, you’d be a more effective problem solver
I of course believe many intelligent people are rendered utterly cognitively debilitated, through the corruption of their software
I believe software can compensate for a lot of hardware defects, including natural neurological deterioration… but of COURSE there’s a threshold where nothing will help. Tip: Try not to damage your brain