DEV Community: Hyunseung Lee

⌈Computing Machinery and Intelligence⌋, a summary for those who don't understand like me

Hyunseung Lee — Tue, 19 Dec 2023 05:22:47 +0000

...written by a 15 year old

Not so brief summary and important notes

The subtitles of this part matches the subtitles of Turing's paper.

(1) The imitation game

Alan Turing starts by introducing a new game, the imitation game. This game is meant to help answer the question, 'Can machines think?'

He then explains another form of the question, but less ambiguous, to help the reader understand the imitation game.

(A) is a man, (B) is a woman, and (C) is the interrogator.

(C) is to try guess which person either (A) or (B) is of either sex man or woman

(B) is to help the interrogator guess correctly

Another important thing here is that the interrogator should be communicating only using a typewriter. Using a typewriting can hide the person's voice, pitch, handwriting and more.

But the interrogator won't know whether (B) is telling the truth or not.

Now, replace (A) with a machine and the interrogator will try to find out whose the machine. In short, Alan Turing's opinion is that if the interrogator is wrong on 70% of the time, the question is meaningless to answer because it already suggests machines are smart enough to fool another human into thinking that it is a human.

(2) Critique of the New Problem

The game is to hard on the machine. The interrogator could simply ask to perform a hard arithmetic problem and the individual to answer correctly will be the machine. So, can't machines do things the best to their ability rather than act like a human? Turing ignored this problem, saying that it could go away if we set the machine to play the imitation game satisfactorily. This suggests that Turing emphasizes the importance of AI acting like a human rather than acting rationally.

	Rational	Human-like
Think	Think Rationally	Think like a human
Act	Act Rationally	Act like a human

In today's world, acting rationally is emphasized and generally considered to be the definition of AI (though arguable).

Another interesting text to point out is Turing thinks ating like a human is not the best strategy for the machine. Not sure why Turing thinks so and how the machine would win if they weren't acting like a human.

(3) The Machines concerned in the Game

Turing gives 3 conditions for machines to ideally have.

We should allow every kind of engineering technique to be used
We should allow the possibility of experimental machine (so experimental, in fact, that the engineers don't even know what is going on), which is exactly what AIs are like today
The machines shouldn't be born naturally, but made artificially

(4) Digital Computers

Turing introduces one of the most oldest planned machine, the Analytical Engine. This engine is planned out by Charles Babbage, a mathematician in Cambridge. Turing notes that unlike in modern machines, Babbage planned on using purely mechanical parts as storage. He planned on using wheels and cards for storage instead of electricity.

Turing writes the following:

The fact that Babbage's Analytical Engine was to be entirely machanical will help us to rid ourselves of a supersition. Importance is often attached to the fact that modern digital computers are electrical, and that the nervous system also is electrical. Since Babbage's machine was not electrical, and since all digital computers are in a sense equivalent, we see that this use of electricity cannot be of theoretical importance ...(shortened)... The feature of using electricity is thus seen to be only a very superficial similarity. If we wish to find such similarities we should look rather for mathematical analogies of function.

Summarizing, Turing thinks that machines using electricity is only a superficial similarity and that it is not of theoretical importance. Not sure what he means, but if he means that mechanical machines can be smart, he seems to be wrong on this one as 100% of AI today runs on and saves data on electricity.

(5) Universality of Digital Computers

He then introduces 'discrete state machines'. Basically, discrete state machines are machines that are programmed to work for specific sets of scenario.

He points out an important property of this machine. This kind of machine is predictable and has the ability to mimic other machines. So, what if the smart machine plays one role and the mimicing machine plays the other role? The interrogator won't be able to notice the difference between these two.

He points out these kind of machines, if programmed to mimic the smart machine, can play the imitation game, too. He did point out that this machine would need adequate amount of storage and fast speed. But Turing, again, seems to be wrong on this one. There is almost no way to program every scenario as it would take more storage than ever imagined and the machine wouldn't be able to take action in unprogrammed scenario.

Here, the 'universality of digital computers' means the property to copy other discrete state machines.

(6) Contrary Views on the Main Question

Turing's Prediction

He gives a prediction in 50 years (2000):

He predicts that there would be machines programmable in a storage of 1GB. Back then, 10KB was widely used (history of data storage)
He predicts that with 1GB, the machine would be able to play the imitation game, tricking the interrogator more than 70%.

He says if this is achieved, 'Can machines think?' question would be meaningless to ask.

But, boy was he wrong on this prediction.

IBM achieved 50GB.
The first chatbot to pass the Turing Test, without any workarounds, was Google’s LaMDA around 2022.

Opposition Views

(1) The Theological Objection

Summary of Objection: God has given soul to only man. God did not give soul to animals or machines. Hence, animals or machines cannot think.

(2) The 'Heads in the Sand' Objection

(3) THe Matematical Objection
(4) The Argument from Consciousness
(5) Arguments from Various Disabilities
(6) Lady Lovelace's Objection
(7) Argument from Continuity in the Nervous System
(8) The Argument from Informality of Behaviour
(9) The Argument from Extra-Sensory Perception

TWA Apps - How to hide the URL Bar (Browser Bar) and Display app as full screen

Hyunseung Lee — Wed, 01 Mar 2023 06:24:13 +0000

Intro

I have been making Google Play Store apps in the last 3 days with TWA using Bubblewrap

It took me 1 day to prepare the app for publishing and 2 days to figure out how to remove the URL Bar and show my app as full screen.

The Problem

The picture below is an image from Bubblewrap Repo's Issue.

The app is shown like it is opened from the browser.

However, if you want to remove them, there are few steps you have to take.

The solution

In a nutshell, you have to paste the following code to .well-known/assetlinks.json file of your PWA website.



[
  {
    "relation": [
      "delegate_permission/common.handle_all_urls"
    ],
    "target": {
      "namespace": "android_app",
      "package_name": "app.web.finalscalcu.twa",
      "sha256_cert_fingerprints": [
        "AA:E5:9C:AB:40:56:31:CD:2F:AF:98:FB:67:F0:0B:68:AC:F1:9A:2C:3A:77:0E:61:42:38:37:66:43:AB:55:4D",
        "39:77:6A:AE:F1:88:86:E9:81:EE:16:CE:05:ED:37:13:7B:61:F6:04:44:D5:AC:CB:0A:A8:79:08:0D:DD:27:B3"
      ]
    }
  },
  {
    "relation": [
      "check_validation"
    ],
    "target": {
      "namespace": "android_app",
      "package_name": "app.web.finalscalcu.twa"
    }
  }
]

You have to replace the array in sha256_cert_fingerprints with your own fingerprints

Getting the fingerprints, app - Google Play Console

If you are publishing to Google Play Store, you are probably using Google Play Console.

Steps

Go to Google Play Console and Select your app.
Select App Integrity

Go to App signing

Copy the fingerprints (2)

Paste them in your json file
Go to Dashboard

Check your app

Make sure the information given to you matches the one you put in your assetlinks.json

After uploading the JSON file, check whether you succeeded in making the URL bar disappear by going to Deep Links.

If it looks similar like the image below, you succeeded.

If not, click the button below and expand the fingerprints section to copy the code Google Play Store gives you

Help reducing Firebase Realtime Database Download

Hyunseung Lee — Wed, 01 Feb 2023 00:08:11 +0000

Hello.

Does anyone have any specific tips about reducing Firebase Realtime Database Download?

I couldn't find any code-related specific tips about reducing download in the internet.

I am having some problem with the download limit.

Is this a scam or legit?

Hyunseung Lee — Fri, 27 Jan 2023 14:17:25 +0000

Story

The other day, I got an email from someone named 'samuel essel'.

The only time I revealed my email was through GitHub, so when he said he wanted to talk to me to work with something, I thought this guy was asking about programming or coding.

Email

After replying what he wanted to do, he replied back several days later.

My thoughts

Well, I don't have many experience with these emails.

My first thought was that it was a scam. I mean who contacts a random person from the internet because they have the same surname?

My second thought was that it may even be legit. Though it was my first time receiving emails like this, I was pretty sure there were better ways to spam people than simply making up stories.

Make a Finals Calculator

Hyunseung Lee — Thu, 26 Jan 2023 05:23:50 +0000

Before you read

This article is to take notes and to use them in the future when writing about my life to go to college and high school.

If you are interested about a boring teenager's life, read on.

Or you could just leave.

Inspiration and App

For about a few pages, I will be talking about my app's inspiration and how it came to be.

Web App: https://finalscalcu.web.app
Play Store App: https://play.google.com/store/apps/details?id=app.web.finalscalcu.twa

🤔 Inspiration - Problem

A few months back, in December, it was the Finals term and a lot of my Middle School friends were struggling.

Not everybody is perfect, so many of the students get a few points taken away from their projects as well as their midterms.

Few of my friends were very curious how much they needed to get in their finals to get at least A or sometimes B.

💡 Inspiration - Ideas

They asked people very good at math, including me, to calculate their scores.

One of my classmates and I calculated all of my friends score, but while doing so, I realized the process could be automated.

So, I decided to make an app in Svelte the following day to help my friends.

👍 Reactions

A lot of my friends complimented on my app.

They said it was useful, less stressful to study, fast, and easy to use.

👤 Users

This app is actually my most successful one so far (by users) though it is far the simplest to create.

During finals term, there were over 200 people per day.

Though the app is used once a year because the finals happen once a year.

Theory

My theory, before making the app was the following:

The way you calculate the lowest score to get A is by making sure the overall score was higher than or equal to 89.5%
The app will ask for inputs of midterm, the percentage of midterm and finals respectively, and the projects score
Then, it will display a panel that shows how much score is need to get at least A, B, C, e.t.c.

Analytics

As seen above, the most popular time was the Korean Middle School Finals Exam Week.

The little surge after the Exam Week was after posting this Dev.to post and making the play store app.

Links

My GitHub
Repo