Hey there, friend.
Good to see you here again.
There are quite a lot of things to cover on this subject. So, let's begin right away.
I started writing the first draft of this post on the 18th of this month.
After finishing it, I left it to review the next day with a different perspective.
I went through it to make sure everything was looking good and it was correct.
And I realized that I left out some things that should definitely be included. Sooo, I added them afterwards.
But before posting it I found about more "news" that came out and I thought "I should probably mention some of that even briefly".
Making those edits and adjustments made the entire draft like 14 min long to read.
I've never made an article that long in the past 3 years. So, I thought about having a different approach.
I also tweeted about that.
And later decided to rewrite the entire draft and cover this subject in a series. Meaning, shorter articles focused on one aspect instead of a mega, ultimate, too-long-to-ever-read one.
That's it. All the context you need to know.
What follows is the first part of the original draft.
The week that changed the world
What craziness we have experienced in the past week.
It started as a regular week, business as usual so to speak.
Everything was going normal until that Wednesday when the OpenAI folks decided to do that reveal and live demo of GPT-4.
Calling it mind-blowing would be an understatement.
In the following hours, people were tweeting the kinds of things they started using it.
A whole range of stuff like...
A fully functioning website (on the GPT-4 demo itself).
And so on.
But that was only on one day.
In the next day Microsoft did a keynote presentation where they introduced the "Microsoft 365 Copilot". Like AI Bing but for the entire Office Suite or better said, AI integrated in ALL their current products.
The stuff they showcased was seriously mind-blowing. Word, Excel, PowerPoint, Outlook, Teams, everything completely enhanced.
And clearly, Google couldn't be left behind. Shortly after they announced, "A new era for AI and Google Workspace".
Basically AI for all Google products. (Although to be honest, this one looks like a worse version of Microsoft's offering.)
And I wish it would been just that.
But the folks of Midjourney came out with the release of V5.
And if you've seen the kind of illustrations you can create with it. 🤯
(That's of course if you are using good prompts...)
It's now in a whole different category. I've used DALL-E 2 before but it doesn't even come close to what V5 can do.
And after that...
You guessed it...
There's more stuff that came out. Not as hyped up as the previous 3 but still very impressive and relevant for many folks.
I feel like those infomercial guys who use the typical line, "but wait, there's more!".
Because in this case, there's still more to talk about.
I have to mention here that after that week everything changed. The world will never be the same. Kind of like what happened 3 years ago in the middle of March...
Now, let's talk about some of those other tools.
There's one tool that has been called a "Midjourney alternative" for its particularly stunning images. From anime-style images, cartoons, fantasy landscapes, and even photorealistic illustrations.
This one is called Leonardo AI. I think it wants to help you be some kind of Da Vinci using it.
It's a free alternative for now and it has waitlist for which I already applied. (Update: at this time, I've already been accepted and created an account on the app. I'm yet to try it out and play around with it.)
You can search on Youtube the name and it will show you tons of videos explaining what it's and how it works.
This particular video is a good starting point.
Another tool that you will find interesting is called Mage
What's interesting about this one is that is not "just another image generator". This is more like a platform that has several different models fine-tuned for different kinds of image generation to suit your needs.
Need photorealistic images? Done. Need anime characters? Done. Need sci-fi scenery or landscapes? Done. Need 3D characters? Done. Need high-quality artwork? Done.
And I could keep going with this...
Most models are derived from Stable Diffusion although there are other different ones for different results.
Those 2 are the most interesting but I'm sure there are much more out there.
You know the deal by now. Having a chatbot that can answer pretty much all your questions in a wide range of topics and areas.
These ones do just that but with the added benefit that some of them have been trained with more recent information and even have access to the internet to give you additional resources to look up.
I could very well start talking about Bing AI with its 3 different modes of "creative", "balanced", and "precise". But it really doesn't need any further introduction.
Let's give the spotlight to other tools with less advertising budget.
First off, we have Character AI.
Like with ChatGPT, you can make it assume a personality or an specific role to play. You can even make it say something in the voice of a person or a character.
With this tool you don't need to do that pre work.
You can literally get started talking with a videogame character (like Mario the Italian plumber) or a famous person (like Elon Musk for instance).
But you can also use it to talk with your own virtual assistant, a pair programmer, a study buddy, or even a psychologist.
Next up, we have the search engine YOU
This is pretty much like Google. You can search anything you want here, with the exception that is not going to fill you with ads and put the first relevant like halfway down the page.
You can leverage your searches for finding images, videos, news, etc.
You can also use that search as a query for the You chat. Ask it questions and it will give relevant answers and with access to recent info and sources.
And here's where it gets interesting. You can also use it for text generation, image generation, and code generation... 😲 (yeah, I know)
Now, moving on...
Here we have the strongest contender by far.
It's called Perplexity.
This is pretty much a better version of ChatGPT that doesn't "hallucinate" responses because it has access to the internet, can do searches, can tell you about events of this year, can also answer questions in a wide range of topics, and can give you an expanded or summarized version of the answers.
You know, like It can give you the tl;dr in many subjects.
Add to that the fact that you don't necessarily have to go to the homepage and signup every time that you open the browser and want to ask it something.
It has also a browser extension for Chrome and Edge. You can literally ask questions about sites you're visiting or the content that's on text heavy pages.
It recently came up with an app for iPhone so if you want to be a lazy bum and ask it questions while laying on a sofa... Now you can do just that.
Pure computing power at the palm of your hand.
This last one is not a ChatGPT alternative but I felt the need to include it because of how cool it is.
It's called Bloom. And its tagline is, "The world's largest open multilingual language model".
That's a mouthful, I know.
In essence, it's a Large Language Model (LLM) that is open-sourced, has even more parameters that GPT-3, knows a ton of languages (both human ones & programming ones), and folks can play with a smaller version on the Hugging Face interface.
The whole fact that it knows so many languages and is super versatile is awesome.
Just try to translate something from English to Spanish (and even all the localized variations of it) while retaining the original meaning using traditional methods and you'll see what I'm talking about.
If that week where GPT-4 was released was a pretty eventful week for AI, we were not prepared for the next week.
The new releases started coming right from Monday.
And it seems that everyone who was working on AI projects chose to prepare their launches and showcases for this month after they saw that OpenAI reveal.
We'll be covering those in the next post.
I'll end this one here.
Stay tuned for the next one that's coming soon.
(I just need to do some editing on that remaining big chunk of text 😅)
If you have read till this point, thank you!
I appreciate it a lot.
See you in the next one!