AI is of course intelligence — “artificial” intelligence

It’s not human intelligence yet, of course

It exceeds us at some domains — and is utterly helpless in others

 

(it exceeds either when:

-there are clear rules — it can then “play with itself” and figure out optimal strategy (e.g. chess)

-or when there’s mountains of data — it can then become the most educated entity in history — using that knowledge to intuit answers (LLMs)

)

 

So AI’s superintelligent aspects already boost our intelligence profoundly

Most of us just didn’t learn how to use it yet

Most of us don’t even have notion of what intelligence is, we’re so stupid

We barely ever use intelligence of our own — let alone of others (books, knowledge, good advise, experts, coaching, etc)…

…or artificial

We just learn some WAY of doing things, way of thinking and believing things — so that we don’t ever have to think again

So yeah

If you were a mindless, opinionated moron before — AI is not going to help you. You never used intelligence either way: artificial, or any other

 

Now this text is not about it

It’s about the indirect impact of LLM’s of human cognitive ability, intelligence

 

I will not say it’s greater benefit than AI itself. It’s like saying that just because some men have now achieved unparalleled feats of strength — we shouldn’t use machines to help us move heavy things around

We should still automate as much intelligence as possible — and leave ourselves just to that which can not yet be automated…

= creativity, NEW knowledge creation

 

What I will say however, I think this is the greatest revolution in human intelligence since the internet

Which itself has been the greatest revolution since print

 

What do I mean?

Did internet not make us stupid?

Yes it did

…but it STILL made us smarter,

And some of us — much, much, much smarter

Just like print

The rise in witch trails and hysteria after the print was invented speaks to how those revolutions take time to adapt to

 

Of course it all hinges on knowledge

However not merely acquisition of knowledge, obviously

Collecting knowledge has little to do with intelligence. If anything — it’s unintelligence (when you know so much that you refuse to ever think yourself)

It’s the loop of:

-Thinking oneself

-Acquiring knowledge from others

-Thinking through it

-Testing in real life

-Then thinking yet more, acquiring yet more knowledge, and testing again

 

See? Knowledge is key component here

But you still must think

You still must do something with that knowledge, else it’s… essentially useless noise, religion, myth, nonsense

But knowledge is vital

 

What did print do? Most simply disseminated knowledge

What did internet do? EXPLODED that process

(+other media in between those two)

 

What’s absolutely vital here however is to observe that it’s not merely about the QUANTITY of knowledge, quantity of data

In BIG DATA there’s the idea of 5 Vs: Volume, velocity, variety, veracity, value

No point in explaining all that, but please notice how 2 Vs pertain to QUALITY and PERTINENCE of knowledge/data: VALUE, VERACITY

 

Arguably a monk in the medieval monastery was already oversaturated with knowledge

But not all of that knowledge was equally useful

Once we got print — obviously we became oversaturated with books to read

But not all of those books are worth reading

Internet just utterly oversaturated us with even more content, more ideas, more nonsense

I don’t have to convince nobody how it’s probably not a good idea to expose oneself anything above 1% of the internet

 

So the problem of QUANTITY, VOLUME, has been solved for quite some time,

It’s about curation

And we were solving the problem of curation in parallel to volume

We were writing books which better explained given subject

Or we would print an anthology of poems, so that we could just read the best of them

Etc.

And internet of course took it to another level

There emerged different layers of access to the same knowledge

+I believe that formulation of an idea is an idea in and of itself

 

The final layer is of course LLM,

The ULTIMATE interface

The access to the expert professor,

And the superior tutor,

The integration of every layer of synthesising and conveying given knowledge, all in one

 

Why is this curation and integration so vital?

On the most simple level — it makes access to the necessary information, insight, so much faster

That’s why one would rather have one encyclopedia in the house, vs all the expert literature on every subject under the sun

But what this really means, in practice, is that the feedback loop becomes compressed into the most optimal units

 

What does it mean?

Our brains can only retain so much information at once, in the short term memory

If you’re tackling a very complex problem — you will need to break it down into subcomponents — otherwise you won’t understand what is going on

Alas those subcomponents are themselves complex

If you’re wise you will become aware of the limitation of your own brain, and take your time to deeply penetrate each subcomponent, until it is well in your grasp

However that is also costly, it takes time

And it is difficult

And it is not entirely practical. We’re not all academics having all the time in the world to penetrate some subject deeply. We have real problems to solve in real world — and it’s a good thing.

But what this means is that for many subjects we thus develop very poor, makeshift understanding

We skim over the details

To put it most simply — we just learn ineffectively and poorly

 

However with this superior synthesis of knowledge, superior curation, superior targeting — you can cut straight to the essence

You don’t have to burn time you don’t have on details you’ll STILL not going to fully grasp

You don’t have to FAKE understanding, out of time and cognitive constraint,

 

Here’s a good metaphor:

Let’s say you want to get through a dark forest. You don’t know the way. You’ll wander and eventually get through.

In your wandering, you’ll have learned various parts of the forest. It was necessary, to figure out the way,

Granted you’ll still forget some of those parts, as one tree is not so different to another, and you have been wandering for days, and you can’t really tell which tree is more important and what is it’s significance in the context of the forst

Now, imagine someone gave you the MAP,

Not only are all regions and landmarks clear, the important trees, mushrooms and rocks

But the way through is revealed too

You still have to learn to follow the map

And you’ll still learn the features of the forest as you make your way along

But you’ll learn what is important, for the task at hand

AND you’ll actually retain it — as that which is practical is far more easily retained

AND by definition it is TESTED — in the very act of you exercising this knowledge

And finally, at the end, who do you think actually knows the forest better?

The one who derived insights from a map — and then applied them?

Or the one who lost his way, barely got out, and had seen a lot of fucking trees?

 

This is the power of curation of knowledge

Power of knowledge about knowledge

 

But again to the point of optimal loops,

The knowledge loop, the feedback loop — cannot be too slow, and cannot be too large at once

Otherwise the details perish from your memory BEFORE you have received the feedback

It’s like playing tennis — where you strike the ball — but have to wait a week to see if it made it over the net

That’s what happens when we’re dealing with too much knowledge which we fail to properly synthesise,

It’s just very ineffectual

 

AI of course largely solves this problem,

If you actually want to learn something — you can absolutely drill it down

You can find JUST THE RIGHT understanding of the problem for your purposes — that’s also NOT untrue

Likewise your ideas — you can just ITERATE them and test them against entirety of human knowledge

When else was that possible?

You’d have to have all the top professors and erudites in the world at the tip of your fingers… and shamelessness to badger them with every silly little idea that went through your head

 

See how this accelerates the loop of ACQUIRING and TESTING and IDEATING with KNOWLEDGE?

It’s completely not unlike someone having drills 10000 topspin shots in tennis — vs someone who only tries to pull it off every now and then in a game

We all know of someone failing to progress at something at all — vs someone who has had smart coaching and DELIBERATE PRACTICE — resulting in them learning very rapidly

 

AI does it — but to everything

 

And ultimately — it accelerates HUMAN INTELLIGENCE itself

Intelligence is quite plastic

And extremely reliant on efficient algorithms

Poor or rigid or useless knowledge and mental models — are weak, suboptimal algorithms which held you back

But there are plenty algorithms which make you think clearly and effectively and fast

And you best hone them by, again:

-DOING THINGS

-acquiring new necessary knowledge

-thinking it through, interiorising it, coming up with new ideas

–repeat

Now this loop has been drastically accelerated

You can be so much smarter so much faster

You can think on so much higher level

You can pick up mental models and algorithms that are so much superior — and ACTUALLY prove it by testing them in this loop

Knowledge is no longer the bottleneck, nor are the conclusions from that knowledge

 

This is why AI is such a revolution, culturally,

YES, it will be a new industrial revolution, and that’s already historical

But I think there will also be a cultural revolution, pertaining to knowledge and understanding and HUMAN intelligence

I’m arguably more exhilarated about the latter

Humans will always work — being alive is to work. You can’t NOT do things. It’s just goals which change

But understanding, that is a truly inspiring venture