What is Artificial Intelligence?

Well, according to John McCarthy, who is often considered the godfather of AI, it’s “the science and engineering of making intelligent machines, especially intelligent computer programs.”

But the history of AI goes way back before McCarthy – to the 1950s when Alan Turing, the legendary “father of computer science”, came up with the “Turing Test.” This test, which is still debated today, was designed to see if a computer could trick humans into thinking it’s a real person. Imagine your computer pulling off a prank like that! 😲

Fast forward a bit and we have Stuart Russell and Peter Norvig, who wrote the AI bible (figuratively speaking), “Artificial Intelligence: A Modern Approach.” They came up with four potential goals or definitions of AI, which basically boil down to computer systems that can think or act like humans, or do so rationally. I guess we could say they’re the “Four Commandments” of AI! 😉

Nowadays, AI is everywhere – it’s like the seasoning in our technology soup. 🍲 From machine learning to deep learning, AI algorithms work tirelessly to create expert systems that can predict and classify information. It’s like having a psychic computer buddy! 🔮

But, like any new trend, there’s a lot of hype around AI, and it’s at the peak of its “inflated expectations.” (Think of it like the bell-bottom jeans of the tech world.) 🕺 As ethical discussions around AI emerge, we’re starting to see the first signs of the “trough of disillusionment.” Sounds like a scene from a fantasy novel, right? 🧙‍♂️

AI comes in different flavors: weak AI, strong AI, and everything in between. Weak AI (or Narrow AI) is like a dedicated personal assistant – great at specific tasks like Apple’s Siri, Amazon’s Alexa, or IBM Watson. Strong AI, on the other hand, is a mix of Artificial General Intelligence (AGI) and Artificial Super Intelligence (ASI). AGI would be equal to human intelligence, while ASI would surpass it, making HAL from 2001: A Space Odyssey look like a party trick. 🤯  Some people think GPT-4 on ChatGPT is a strong AI.

Deep learning and machine learning are often used interchangeably, but there’s a subtle difference. Deep learning is a sub-field of machine learning and uses neural networks to learn. Imagine a computer that can teach itself how to learn – like a self-taught prodigy! 🌟

AI is all around us and has a bunch of cool applications. From speech recognition that helps you talk to your phone (Hello, Siri!) 📱 to customer service AI chatbots that assist you with your online shopping 🛍️, computer vision that helps cars see and drive themselves 🚗, recommendation engines that know what you want before you do, and even automated stock trading that’s faster than a speeding bullet! 💨

AI’s history is filled with exciting milestones and breakthroughs, from Turing’s groundbreaking paper to IBM’s Watson winning at Jeopardy! 🏆 and Google’s DeepMind beating the world champion in the ancient game of Go. It’s been a wild ride, and we’re just getting started!

So, there you have it – a whirlwind tour of AI that’s sure to leave you craving more. And who knows, maybe one day we’ll have AI that can write

AI Models

Large Language Models (LLM)

The most commonly known AI model, ChatGPT (GPT-3.5 & GPT-4), is an LLM.

Essentially, these are known as LLM’s because they use vast quantities of text to train the model…

…to the point that people have suggested that, perhaps, OpenAI ran out of text data to train the model on after GPT-4.

AI Image Models

The most commonly known AI models are DALL-E (OpenAI), Stable Diffusion, and Midjourney.

AI image models are at an amazing point where, at least in the case of Midjourney, it can create photorealistic images and other images that look to be created by a real human artist.

Amazing!

AI Sound Models

There are a variety of AI sound models out there.

There are 2 main categories that I see right now: (1) Text to Voice & (2) Music.

Both have wide ranging applications.

AI Video Models

While a few AI videos have gone viral, overall, as of April 2023, AI video models are not really there yet…

…but, it seems like just a matter of time before they get to being able to create realistic looking videos.

Narrow AI Models

Narrow AI models is where there is virtually limitless untapped potential (in our opinion).

AI models that are created for a specific purpose (vs all the general models above).

We, at 42 Robots AI, feel confident that over the coming years, narrow AI models will play a huge role in changing the world.

By creating an AI model for a specific use case or specific target market opens up a world of opportunities.  Essentially, you can the same quality results as a huge, $100 mil+ cost, AI model, but for a fraction of the price.  Flip it around and, still for less of a cost, you should be able to get a BETTER result than any general model that costs 100x to build and train.

OpenAI

OpenAI is leading the way in the Artificial Intelligence world.

Below is a quick summary of their AI Models.

ChatGPT

We should start the discussion about OpenAI with ChatGPT, the fastest growing subscription software ever.

It took the world by storm, impressing almost everyone with its amazing capabilities that many thought wasn’t possible in our lifetimes.

ChatGPT was originally built with GPT-3.5.  Now, if you have a Plus account ($20/m), you should have access to GPT-4, a major improvement over GPT-3.5.

Whisper

Whisper is OpenAI’s speech to text AI Model.

They recently released the API for Whisper.

This is one of their less…unique models as there are quite a few other speech to text AI models.

As of April 2023, it does not current do text to speech.

GPT-2

OpenAI’s first AI model released.

GPT-2 was trained with 1.5 Billion parameters.  That might sound like a lot, but it pales in comparison to GPT-3 and (presumably) GPT-4.

GPT-2 had some good capabilities, but didn’t exactly take the world by storm.

GPT-3

Trained with 175 Billion parameters, GPT-3 was a HUGE improvement over GPT-2.

The quality of GPT-3’s responses showed great promise for OpenAI’s large language model strategy of aggressively increasing the size of the models.

GPT-3.5

ChatGPT’s primary model that was built with bot chat in mind.

This is the first AI model that many people encountered and had their minds blown.

There are/were still many issues with it.  Most prominently, it would “hallucinate” confidently incorrect facts.

GPT-4

GPT-4 is a MAJOR improvement over GPT-3 & GPT-3.5.

It’s ability to use reason and logic is incredible.

This was the first time someone could honestly say it had signs of an AGI (Artificial General Intelligence).  It definitely isn’t all the way to an AGI, but, wow, does it feel quite close in a lot of ways.

GPT-4 is incredibly useful in a wide variety of use cases.

If you have not tried GPT-4 yet, I highly recommend getting ChatGPT Plus and taking it for a spin!

GPT-4 is supposed to be able to read images, but, as of April 2023, they have not released this capability yet.

DALL-E

DALL-E is OpenAI’s text to image AI Model.

By comparison to other text to image models, it is mediocre.

We recommend Midjourney over DALL-E.

Prompt Engineering

Prompt engineering is a new discipline that didn’t exist just a few years ago.

It refers to working on the prompts (inputs) to AI models to get increasingly better results from the outputs.

Don’t sleep on this skillset as it is likely to be useful to virtually everyone in the coming years.

AI Safety

Artificial Intelligence safety is of great importance because there are some serious downsides to AI.

The most obvious is an evil AGI killing or enslaving all or most humans.

We think a more likely and imminently dangerous situation is the reality splitting nature of AI.  Specifically, around misinformation and trusting good information.

Bad actors have their misinformation tools enhanced.  We are already seeing this in play.

Additionally, the massive increase in fake pictures and misinformation will make it harder for people to trust the good information.

AI Tools

Click here to check out a huge database of AI tools.

AI Detection Tools

AI Chat Tools

AI Copywriting Tools

AI Marketing Tools

AI Text to Speech Tools

AI Speech to Text Tools

AI Programming Tools

AI Image Editing Tools

Chat Bots

Types of Chat Bots

Customer Service Chat Bots

Shopify Customer Service Chat Bots

Sales Chat Bots

All Purpose Chat Bots