Getting smarter or dumber using AI

[Bootstrap]

In another thread, I mentioned this principle: if you are using AI in ways that keep you engaged, creative, and doing the thinking you should be, it can help you become more intelligent and more informed in doing whatever you are doing. If you are using AI to get easy answers, with no idea how to know if the answer is true or how to find the answer yourself, you will gradually become dumber in that subject area.

Think calculators. Before calculators, most American adults were much better at mental arithmetic than we are now. Many cannot even compute simple percentages mentally, unless we had to help our children with schoolwork.

I can use AI like a calculator: "What is 18% of 24.80?"

Or I can ask it to be my tutor, giving me exercises to do, giving me problems like this:

Problem 2

What is 5% of 480?

Tell me your mental process. For example, you might think:

* “I already know 10%…”
* “Then I…”

Don’t worry about being fast. I’m interested in how your mind is doing the calculation.

Let me make an intentional mistake and see how it responds:

96. I divide by 10 (move the decimal point one place to the left), then multiply by 2.

Here's its response:

Here’s a concise version that captures the teaching point:

Tutor: Correct method, wrong percentage.

You computed:

* Divide by 10 → 10% = 48
* Double it → 20% = 96

So you found 20%, not 5%.

For 5%, take half of 10%:

* 10% = 48
* Half of 48 = 24

The important thing is that the strategy was right. You were thinking in terms of transformations (“divide by 10, then…”), which is exactly the habit we want. We just need to associate the correct transformation with each common percentage:

* 10% → divide by 10
* 5% → divide by 10, then halve
* 20% → divide by 10, then double
* 1% → divide by 100
* 2% → divide by 100, then double
* 25% → divide by 4
* 50% → halve

If your mental arithmetic is fading, this is a way to improve it. And you can easily measure whether your skills are improving or not.

I think the same principle of engaging your mind in the right things, making sure you are learning and in control of the subject matter works across many fields, including programming, writing, etc. For instance, instead of having AI rewrite your texts lazily, discuss what you wrote, push back when you are convinced your sentence was better or when the AI changed the meaning. When writing code, let the AI help you get an overview of larger codebases and their architecture, but also make very sure that you can verify that the AI is telling you the truth, looking at the code, designing tests, etc.

[Neto]

I think that outside of math, any information aids CAN tend to make a person dumber, IF the desire is to have that knowledge at your disposal without the info aid at hand. I think that at some point in our educational system 'knowledge' or 'smartness' turned from knowing the facts to knowing where to find them.

In experiences in living with a preliterate people group, when they prompted me to answer questions about something they had previously taught me (usually the name of a certain bird or tree), I too often had to say that "I don't know, the paper knows". (In other words, I had written it down in my notes, and knew where to find the answer, but I didn't possess the answer in my own mind.) In a language group like the Banawa, all knowledge is held collectively in the minds of the people. So in that sense, literacy has made us 'dumb', and I expect the same thing is true of knowing correct spelling, grammar, and historical facts as well. Someone will possibly say that the Banawa have much less collective knowledge than we do, as an industrial society. That is true, but they also have a great deal of ready knowledge about things for which we have literally no clue, even some things that are not available to us in all of our literature (and the internet, AI, etc.). (I was going to do a test by asking Google AI something about a certain fruit which I know in Banawa, but since even the local Brazilians have no name for it other than 'jungle fruit', it came up with an entirely different fruit. It DID, however, refer to a river town in the area, and one tribal group that is spread throughout the larger area, into the adjacent area along the same river system. I had given it the Portuguese name for the specific area, a certain segment of a certain river system.)

As for math, I'm not terribly good at complex computations, but I can figure out fairly complex multiplication and division problems in my mind, without paper & pencil. (Laying in bed at night, in fact.) In the example you gave, a calculator, and normal paper-based computations follow a different route, one not normally used for 'mind-solving'. There are people who can do extremely complex computations in their mind, and very quickly as well. I didn't attempt to look it up on-line, so I don't know how well science understands the processes those people use to come up with the correct answers.

[Soloist]

Wife: whenever there’s a Spanish word that I’m having a lot of trouble with remembering, I often look up memory aids/the root word. It seems OK at that, although some of the memory aids are really stupid. Like for example, last night, I was trying to memorize the words impatar and agarrar, which mean to tie/draw (like in soccer) and to catch/grab. Latin root of the first one is par (pair), and the other one had some stupid idea of a gorilla or an alligator grabbing something. turns out the root word, garra, comes from Iberian instead of Latin, and it means claw.

I’m not sure, but hypothetically, these memory aids are probably just from some web search where a human made those up. I suppose Google AI is probably low quality, but it seems to work for stuff like that.

[barnhart]

I remember reading a segment from a book by an ancient Greek in which he opines that cheap paper has degraded the mental capacity of the youth since they prize memorization less. I would guess this sort of thing eventually stabilizes an about in time for the next innovation in communication.

[R7ehr]

I think it does make us dumber in the sense that we may be less able to instantly recall certain knowledge.
I wonder if it has increased any of our other knowledge. I wonder if it has helped make us better aware of our own limited knowledge, and has shown us that although we may have a pretty good handle on a topic, there are aspects that we hadn’t thought about. Maybe it has filled out some corners of certain topics and abilities. I don’t know - I’ll have to run this thought through AI and I can get back to you.

[Szdfan]

I’m not sure “dumber” or “smarter” is the right framework. I’m not sure that we’re dumber because mental math is harder because of calculators — rather there’s a different skill set that’s being prioritized.

[Bootstrap]

I’m not sure “dumber” or “smarter” is the right framework. I’m not sure that we’re dumber because mental math is harder because of calculators — rather there’s a different skill set that’s being prioritized.

I think we can be smarter or dumber at a particular skill, such as mental math. And that's the meaning I intended in this thread.

We can also decide where we want to focus our learning and attention.

[Bootstrap]

I think that outside of math, any information aids CAN tend to make a person dumber, IF the desire is to have that knowledge at your disposal without the info aid at hand. I think that at some point in our educational system 'knowledge' or 'smartness' turned from knowing the facts to knowing where to find them.

In experiences in living with a preliterate people group, when they prompted me to answer questions about something they had previously taught me (usually the name of a certain bird or tree), I too often had to say that "I don't know, the paper knows". (In other words, I had written it down in my notes, and knew where to find the answer, but I didn't possess the answer in my own mind.) In a language group like the Banawa, all knowledge is held collectively in the minds of the people. So in that sense, literacy has made us 'dumb', and I expect the same thing is true of knowing correct spelling, grammar, and historical facts as well.

A very good point. Oral cultures have real skills that we do not in our culture. And what is a crutch vs. effective thinking also depends on your perspective.

As for math, I'm not terribly good at complex computations, but I can figure out fairly complex multiplication and division problems in my mind, without paper & pencil. (Laying in bed at night, in fact.) In the example you gave, a calculator, and normal paper-based computations follow a different route, one not normally used for 'mind-solving'. There are people who can do extremely complex computations in their mind, and very quickly as well. I didn't attempt to look it up on-line, so I don't know how well science understands the processes those people use to come up with the correct answers.

In many primitive cultures, the kind of mental math we learn is simply beyond them. Some do not even have numbers past 10 or 20 in their languages, using a word like "many" beyond that point.

I originally learned mental math on a Japanese abacus - the operations turn into finger movements, I can visualize my fingers calculating it, it's muscle memory. I discovered, to my surprise, that I'm now much worse at it than I thought. I just ordered a Japanese abacus to brush up ... but is the abacus a crutch or a way of doing mental math?

Regardless, I just ordered a new abacus. I haven't had one since 1988, when I originally moved to Germany.

[Neto]

I think that outside of math, any information aids CAN tend to make a person dumber, IF the desire is to have that knowledge at your disposal without the info aid at hand. I think that at some point in our educational system 'knowledge' or 'smartness' turned from knowing the facts to knowing where to find them.

In experiences in living with a preliterate people group, when they prompted me to answer questions about something they had previously taught me (usually the name of a certain bird or tree), I too often had to say that "I don't know, the paper knows". (In other words, I had written it down in my notes, and knew where to find the answer, but I didn't possess the answer in my own mind.) ....

A very good point. Oral cultures have real skills that we do not in our culture. And what is a crutch vs. effective thinking also depends on your perspective.

As for math, I'm not terribly good at complex computations, but I can figure out fairly complex multiplication and division problems in my mind, without paper & pencil. (Laying in bed at night, in fact.) In the example you gave, a calculator, and normal paper-based computations follow a different route, one not normally used for 'mind-solving'. ....

In many primitive cultures, the kind of mental math we learn is simply beyond them. Some do not even have numbers past 10 or 20 in their languages, using a word like "many" beyond that point.
....

Banawa is such a language. The only numerals that have one-word forms are one and two. (1 has two different words, depending on the emphasis to be expressed. One word just means 'one', and the other means 'just one', for which English does not have a word.) Three is two and one more, four is two and two more, five is one hand, etc, until ten, which is two hands. They can also use the toes to go up to 20, but Portuguese number words are increasingly replacing the larger numbers. Then of course they have 'many' and 'lots of many'.

But in their traditional way of life they had no need for large quantity references. That's not 'primitive'; if it is, then WE are 'primitive' in other ways in which they are 'advanced'. Westerners have a lot technical language which is known by only a small sector of society. In oral cultures, nearly everyone knows all of it.

[barnhart]

Very interesting Neto. I'm curious about other aspects of their culture, for example do they make or hold sacred any images?