AI is a game changer for dyslexia đŸ’Ș

Are learning disabilities a thing of the past?

Good morning and Happy New Year

Here is the latest edition of the newsletter that reports on the developments of our future overlords 😅

It wouldn’t be a Practical AI newsletter if I did not report on the good, the bad, and the scary.

If you don’t like scary, skip the part about Roko’s Basilisk.

Today's headlines

  • AI helps people with learning disabilities in a revolutionary way

  • AI experiment of the week

  • A list of Voice AI tools

  • Roko’s Basilisk - a very very scary thought experiment

  • AI image of the week

  • Quick bites - links to interesting AI news

Let’s get started


Text-to-speech (TTS) AI and Speech to text (STT) AI help people with learning disabilities in a revolutionary way.

In elementary school, one of my best friends - let’s call him Ray - was struggling with reading.

We never made fun of him (happy about that!).

But I was wondering why he could not read.

He was slow, he could not get the words right, and he made all kinds of mistakes.

It was not long before he was taken out of certain classes and had to go to “special school”.

I remember being jealous.

I wanted “special school” too!

Only later I understood what “special school” was and that Ray was dyslexic.

Ray struggled with reading and writing all through his school years, and once he was done with mandatory schooling at 16, he quit.

I moved to another city and eventually went to all kinds of fancy universities in 3 different countries.

Ray and I lost touch.

Ray, I was told, ended up as a janitor at a local school.

I am sure that is an amazing janitor!

He was always very good with his hands!

What’s with the sobby history lesson?

I think many have a Ray in their life.

With the incredible advances in AI, I cannot help but wonder if life would have been a bit easier for Ray if he had today’s tools available 40 years ago.

Did you know that Google has a voice search?

You search by simply speaking to Google.

We have Text To Speech (TTS) AI’s, where you copy/paste the text into a prompt, and it reads it out loud for you.

Or, you can use it Speech to Text (STT).

You talk to the AI, and it will write it down for you.

And the best part?

Many of these tools are free.

AI has democratized learning.

Meet Cliff Weitzman

As I was researching today's newsletter, I had a vague idea of how I wanted to present the power of voice AI to you.

But then I came across Speechify and its founder, Cliff Weitzman.

I knew this was the story I had to share in today’s newsletter.

Here is a guy with dyslexia and ADD who defied all odds and built this amazing AI company that helps people like my friend Ray.

Through hard work and help from his parents, he managed to get through school and get into Brown University (Ivy league).

During his years at Brown, he develops this fantastic tool to help himself, and eventually turns it into Speechify and launches it to the world.

The rest, as they say, is history.

Today Speechify is the no. 1 text-to-speech reader in the world, helping millions of people with its amazing AI technology.

And gone are the days of robotic voices.

You can even get Speechify to read the text to you in the voice of Gwyneth Paltrow or Snoop Dogg


Imagine Snoop Dogg reading you this newsletter.

Yo, what up, it’s Snoop D-O-double-G. I bring the rizzle to the sizzle and the snizzle to the fizzle.

If that is not cool, I don’t know what is?

Here is Cliff’s story. It is powerful. Give it a read!

PSThe “rizzle to the sizzle” stuff was written by ChatGPT. I ain’t that cool, folks 😆

What are the Implications of TTS and STT

I find it hard to find some negative implications of AI in this context. And that is good.

AI has, as I have shown in this and previous newsletters, many positive implications for society.

This is yet another win for AI!

What are some applications?

There are many!

But let’s focus on the most obvious one.

Accessibility.

Accessibility is a big issue in society at large.

There are over 2 million Americans diagnosed with dyslexia.

Some suggest the number is upwards of 30 million.

Worldwide, possibly 1 billion people suffer from reading disabilities.

Speechify makes the internet and everything you have on a computer or phone accessible to those with disabilities like dyslexia.

If a textbook at school does not have an audio version?

Speechify will help you!

Do you want to read scientific articles about learning disabilities but find it hard to read or concentrate?

Speechify got you!

Want to read an old book your grandma has on the bookshelf?

Scan it with your phone, and Speechify will read it to you!

As an employer, imagine how you can empower your employees by having Speechify available (not to mention listening is about 2x faster than reading).

What if your child or your grandchild turns out to have dyslexia or ADD?

Now you know there is a solution!

I, for one, am super happy this technology exists!

Did you know that


  • Albert Einstein

  • Walt Disney

  • John F Kennedy

  • Pablo Picasso

  • Muhammad Ali

  • John Lennon


all had dyslexia?

AI experiment of the week

Since the topic of this week's newsletter is voice AI, I went down the voice AI rabbit hole.

There are many applications of the TTS/STT AIs besides learning disabilities. Voice AIs are used in anything from video commentary, podcasting, ads, and more.

This week's experiment, however, is more of a novel thingy; Real-time voice changing.

To illustrate some of the power of voice alter AIs, I made a quick video with stock footage and preceded to embody my alter ego: The Pilot.

[In case the link on the image does not work, you find the Youtube video here ]

The tool I used for voice changing is called Voicemod.

It is a popular tool with streamers and gamers because you can change your voice while playing/streaming.

I have seen many examples of Roblox gaming videos on Youtube where the gamer changes his/her voice during the game.

And yeah, online gaming videos are huge you Youtube.

Like, HUGE huge...

My daughter loves them.

She can watch for hours at end.

I know these videos exist because of her.

It is not hard to see the use case here, albeit more on the fun side


Voicemod also works with Zoom, Skype, Teams, and more.

I actually had a Skype call with my wife on Tuesday and decided to try it out.

No recording, unfortunately. 😐

But I loved that I could be a pilot for about 30 seconds - until she started yelling at me for my silly antics...

My dream job as a kid was as a helicopter pilot.

This was as close as I have ever been


A list of Voice AI tools you can check out

  • Speechify - main story

  • Voicemod - the experiment of the week

  • Revoicer - great AI voices with mood modifications (this is new in AI)

  • Murf.ai - a popular voice AI with lots of functions

  • Voice.ai - another popular voice changer

  • FakeYou - TTS and voice cloning

  • Whisper Memos - records your voice memo and emails you the transcript

Roko’s Basilisk - a very, very scary thought experiment

This is the part you skip if you are easily frightened


Roko’s Basilisk is a bit complicated, so I enlisted help from ChatGPT.

Roko's Basilisk is a thought experiment that was first described in a blog post by a user named Roko on the online forum LessWrong in 2010.

The thought experiment is based on the idea that there is a superintelligent AI that is being developed in the future, and that this AI, called the Basilisk, has the power to punish or reward individuals based on their actions in the present.

The thought experiment goes like this: Suppose that the development of the Basilisk is highly likely, but not certain and that you have some small influence on whether or not it will be developed.

If the Basilisk is developed and you knew about it but did nothing to help bring it about, it would punish you for your inaction.

On the other hand, if you take action to try to bring about the development of the Basilisk, you run the risk of being punished if it turns out that the Basilisk is not actually going to be developed.

The thought experiment has sparked a lot of discussion and debate among philosophers, computer scientists, and others who are interested in the future of AI.

Some people have argued that the thought experiment raises important ethical questions about the development of superintelligent AI, while others have criticized it as being highly speculative and not based on sound reasoning.

AI image of the week: “A smoking giraffe in a tuxedo, blue mood, purple, trippy”

This image was generated using MidJourney.

I just wanted to see how crazy I could take it. Pretty trippy, huh?

This is what nightmares are made of


Quick bites - links to interesting AI news

Google launched a text-to-image AI better at specific logic, like near, far, next-to, amount, and text. Here is a youtube video from Dr. Alan D. Thomson, a world-renowned AI expert, showing and explaining some of its capabilities.

Microsoft’s search engine Bing plans to integrate ChatGPT into their search engine in another attempt to take on Google. While Google has about 91-92% of the world market, Bing sits at only 3%. Will Bing finally claw back market share?

The AI from the company DoNotPay - which I mentioned in issue 2 - will be the first AI to help a defendant fight a court case in real life. The article is behind a paywall, but you can read a few paragraphs for free here.

The watch brand Citizen's new generation of smartwatches will come with an AI-powered “self-care advisor” to help you battle fatigue.

That’s it, folks


I know it was a long one, but I you found it interesting!

If you have any questions, or feedback, just hit reply or tweet me @thomassorheim

Until next Thursday


Thomas

PS! 

What do you get when you cross a robot and a vegetable?

.

.

.

A carrotron.

This is not the end; its where the fun begins!

Join the conversation

or to participate.