Who is going to be cast in the inevitable Netflix series on OpenAI?

- Simplified AI -
Your gateway to discovering the world of Artificial Intelligence!

Turkey Day is over and now we have Christmas, Hanukkah, Kwanza and "Answer the telephone like Buddy the elf day" right around the corner. I just learned that's a thing, so if you call me on December 18th, I WILL be asking what your favorite color is and explaining the 4 major food groups to you - candy, candy canes, candy corn and syrup.

A sneak peek of what’s below:

• AI Training

• Drama at OpenAI, the creators of ChatGPT

• Stop wasting time managing your email inbox

• Get creative with superpowers

How does AI get trained ?

Turning AI topics into a walk in the park, not a mind-boggling maze

Supervised vs. Unsupervised learning:

Despite the popular belief of AI being "all-powerful" or "all-knowing", it still has to learn things just like humans do - it can just do it at a much faster pace.

Supervised learning is one category of AI learning where a computer model learns from labeled examples, like having a cheat sheet so the AI knows the answers to the test. The idea is to show the model enough examples of a certain item or concept, allowing it to use that information to recognize that same thing in different variations.

To you and I, it's pretty straightforward to look at any type of apple and say "yup, that's an apple alright!" But to an AI model, the concept of what constitutes an apple isn't innate, so it has to be taught. By exposing it to apples of varying colors and sizes, and apples in different context (on a desk vs. a painting), the AI model learns to recognize an apple. Then it will be able to identify all apples it encounters in the future and appreciate that it's something that keeps the doctor away if you eat it every day!

You know those CAPTCHA tests on websites where you prove you're not a robot by selecting all the images of stoplights? Well, I have news for you - you're essentially providing supervised learning for AI models by labeling images they can learn from. This supervised learning plays a vital role in enabling autonomous vehicles to "read" the road!

On the flip side, we have unsupervised learning, which, hopefully, you find easier to grasp than picturing a class of 3rd graders jumping out of their seat the moment a teacher leaves the classroom.

Unsupervised learning is when a computer model independently explores and uncovers patterns within data without guidance or explicitly being told what something is. The idea revolves around enabling the model to autonomously identify inherent relationships or hidden features to gain insights and make sense of complex information.

Imagine working on a 5000 piece puzzle, but there isn't a box cover or a picture to reveal the end result. You enlist AI to help, so it begins by putting pieces that appear similar next to each other. It searches for color patterns, identifies the number of edges, bumps and slots on each piece and eventually, it assembles the puzzle and understands what the image shows. Through this process, the AI learns about the puzzle's components and the method that led to its solution.

Why is this important?

Have you ever entered something into an AI tool and the output didn't match your expectations? It could have something to do with how the AI model was initially trained.

Supervised learning, much like our own learning experiences, is extremely important because it tells the AI model what is considered "correct". However, when training these models, what one person views as "correct" can be subjective. There may be underlying biases, both conscious and unconscious, present in the training data that ultimately influence how the AI model makes decisions in the future.

On the other hand, unsupervised learning is adept at data exploration, but poses the challenge of defining a clear measure of success. Additionally, AI companies are currently facing legal challenges related to "stealing" data online from popular forum websites like Reddit and Twitter, sparking debates about data usage.

AI News Flash

Catch the latest AI news that is making all of the headlines this week!

OpenAI Reality Show Drama

This past week almost saw one of the biggest shakeups at a defining company of our generation - OpenAI, the creators of ChatGPT.

A seismic change occurred late last Friday, the 17th, when OpenAI's board of directors abruptly fired the company's Founder and CEO, Sam Altman. While the public remains unaware of the full details, the board cited Altman's lack of consistent candor in his communication. However, there are rumors that suggest the real issue might be connected to advancements in AGI (Artificial General Intelligence - more on this in a future newsletter!)

Altman was subsequently hired by Microsoft, one of the leading investors in OpenAI, within 2 days. However, he returned to his CEO position at OpenAI a few days later, following a strong stance from 735 out of 770 employees at the company who threatened to quit if the original firing was upheld.

Why does this matter?

Given the company's critical role in the AI boom, there was plenty of concern and discussion this past week regarding control over AI models, ethical considerations, and the balance between profit and non-profit approaches in the commercialization of AI tools.

Huge shakeups like this within a company that exerts substantial influence in the AI world should prompt industry leaders, new outlets and the community to monitor large decisions more closely and scrutinize the business choices made by AI companies.

Other Stories We're Following:

• Siri is getting replaced for ChatGPT voice (if you have an iPhone15)

• Sports continue to find new ways to integrate AI as Formula 1 trials AI to combat track violations

AI Company Spotlight

Explore trailblazing companies that are shaping our digital future

SaneBox

If you're anything like me, you've probably thought about hiring an assistant just to manage your email inbox. I juggle work email, a personal email, a second personal email dedicated to signups and spam, and of course [email protected] - where you're all welcome to send me your great ideas about what you want to learn!

While there are a few impressive people out there that can keep their email inbox pristine, for the majority of us, it becomes a daily hassle that often gets pushed aside, leading to a virtual pile-up.

What if you could use AI to help you with this challenge?

Sanebox looks to do exactly what their name says - keep us all sane - by using AI algorithms to learn from your email habits and provide long term solutions. If you already use folders to sort your inbox, then you're already a step ahead at teaching this tool how you operate.

For as low as $7 a month, they can help you manage the clutter. That's the price of one (overpriced) latte a month, for potentially hours back of your time. The effort you put into this tool will determine the outcome. This isn't an "overnight" solution to email clutter…you'll have to spend the time teaching the AI system how you interact with your inbox.

Why does this matter?

At its core, AI is supposed to help humans be more efficient with our lives. By teaching a tool like SaneBox how we operate, we train it to streamline an aspect of our life and free up time for individuals to focus on the things that matter most.

If this tool saves you just 30 mins a day, that's 180+ hours a year opened up that you can spend with your family, reading Simplified AI, or learning a new skill!

We're testing out the service right now with a free trial and we'll share our thoughts on Twitter.

If you can't afford their pricing plans, connect with the company and they said they will work with users on appropriate pricing.

Get Creative with AI

Practice using AI and unleash creativity with your new collaborator

Copy the prompts below and paste them into ChatGPT, BingAI, Google Bard, or your generative AI tool of choice and let AI work its magic. Share your results with us…and take it a step beyond the initial prompts and see what you can create!

Text-based

• Imagine you wake up tomorrow with the superpower of invisibility, but there's a catch – it only lasts for 24 hours! What thrilling adventures would you embark on, and how would you use your newfound power to make the world a better (or more mischievous) place?

Visual-based*

• Create a world where giant, colorful birds act as the primary mode of transportation. In this vibrant landscape, people of all ages, dressed in bright, eclectic outfits, are seen riding atop these majestic birds, soaring through a sky filled with fluffy, pastel-colored clouds. Below, the terrain is dotted with lush, oversized flowers and winding, emerald green rivers. Small, quirky villages with houses shaped like mushrooms and teapots are nestled among the hills.

Visit us on Instagram and Twitter to see how our examples turned out!

*Visual-based examples are for users of Dall-E, BingAI, or premium users of ChatGPT - may require paid subscription.