5 Cool AI Tools & The TLDR in AI This Week

There was a lot of news coming out of the world of large language models this week. Open Assistant, an open-source version of a ChatGPT-like chatbot, was released; Stability AI launched their own models; Elon Musk talked about creating his own LLM; and Google Bard finally added the ability to code inside their model.

We also got MiniGPT-4, a model that's not actually related to GPT-4, but that finally allows us to use both images and text to gain insights about the images. Meta also launched and open-sourced some cool computer vision models. And some lucky folks got to witness some of the most exciting advances in AI at TED, which is happening in Vancouver right now.

On the Future Tools front, I'm working on some really cool tutorials that I can't wait to share. I've slowed down on posting videos slightly (from 5 per week to 3 per week... for now). That's because some of the videos I'm working on are more in-depth and are taking a lot more time to record and edit, but I'm excited to share them with you. I'll likely ramp up the pace of videos again once I finally get a few of these projects that I've really been wanting to make completed and out into the world.

Thank you so much for the support here on the newsletter! This newsletter is about to cross 75,000 subscribers today. The YouTube channel is approaching 270,000 subscribers. My Twitter just crossed 30,000 followers this week... And the Future Tools website is getting 1,000,000 visitors per month now. This has all been so mind-blowing, and I can't thank you enough for all the support. I love the fact that my silly nerd-out habit is shared by so many others!

So let's dive into what I came across this week and continue to nerd out together!

Thank You To This Week’s Sponsor!

Augment AI

• Compose messages in your own tone & style, in the apps you already

  use (Slack, iMessage, GMail…)

• Automatically summarize video conference meetings.

• Stop copy-pasting into ChatGPT. Just reference anything you’ve come across.

• Boundless Possibilities: ask your Augment questions, have it take a stab writing that press release, rewrite entire paragraphs, you name it.

Experience the game-changing feeling of having your own AI model, learning from you and getting things done. We’re briefly re-opening our early access exclusively to Future Tools readers. Sign up for Augment AI.

🛠️ 5 Amazing Future Tools

1. MiniGPT-4: With MiniGPT-4, we're getting our first real taste of a multimodal chatbot. Upload an image, and then use chat to identify what's in the picture and learn more about it. Post a picture of a famous person and ask who it is. Post a picture of food and ask for a recipe to make something similar. It's incredible what this chatbot can deduce from just a single image.
(Free & Open Source)

2. Amazon CodeWhisperer: If you're familiar with GitHub Copilot, Amazon CodeWhisperer will be right up your alley. It's a coding assistant that helps autogenerate code for you, as well as assists in debugging existing code. It plugs directly into your IDE, like Visual Studio Code, and, unlike Copilot, Amazon's version is completely free.
(Free)

3. God Mode: Auto-GPT has been one of the most talked-about tools in the AI world recently. God Mode is my favorite user interface for Auto-GPT that I've come across. Plug in a goal you want to achieve, and the tool will automatically keep working until it accomplishes the objective. It's free to use, but you can also plug in your own OpenAI API code to use GPT-4 if you'd like.
(Free)

4. Animated Drawings: Here's a fun one from Meta. Upload a drawing, rig it up by telling the tool where the eyes, elbows, wrists, knees, and ankles are, and then watch as your drawing gets animated in all sorts of ways. If you have kids who love to draw, wow them by scanning their artwork and bringing it to life.
(Free & Open Source)

5. Open Assistant: Open Assistant is a completely open-source ChatGPT alternative. It's not amazing yet, but its training has been entirely gamified. If you log in and use the chat, providing feedback, it will improve and learn from your input. Since it's open-source, we can also expect forks and new iterations built on the back of Open Assistant.
(Free & Open Source)

🗞️ 3 Fascinating Articles / News

1. Stability AI Launches the First of its StableLM Suite of Language Models - Stability AI is the company behind some of the most popular Stable Diffusion image models. They've been working behind the scenes to develop their own large language model, similar to GPT-3 or GPT-4, to compete with OpenAI. This week, they finally released their first version in both 3 billion and 7 billion parameter versions. Initial tests show that it's not quite ready for everyday use, but it's great to see new tools that could eventually compete with OpenAI.

2. Elon Musk claims to be working on ‘TruthGPT’ - In a similar vein, Elon Musk also believes that AI control shouldn't be solely in the hands of a single company. He recently teased plans to build his own alternative to ChatGPT, called TruthGPT, with less bias. He also claims that he can develop it more safely, with fewer negative long-term implications for humanity. Elon's got a lot on his plate these days, so who knows how quickly we'll see something come from him.

3. Bard now helps you code - Last month, Bard was released for early testing and kind of fell on its face. Everyone wanted to compare it to ChatGPT or Bing Chat (with GPT-4), but it was nowhere near as capable. When asked which tool was better, Bard even admitted that ChatGPT was much more useful. However, today, they announced that Bard can now help you code. It will even write code and push it to a Google Colab so you can share and let others run the code. This is a huge improvement from Bard's previous capabilities. I'm sure we'll be hearing from lots of coders in the coming days about how it compares to the capabilities of GPT-3 and GPT-4.

You can also stay up to date with the latest AI news here.

📺️ Must-Watch Videos

💰️ A Money Idea To Try

Sell Patterns on Print on Demand Sites:

This is not an idea I came up with myself. In fact, I stumbled upon this YouTube video and thought it was an interesting idea.

The concept is that you can use MidJourney to create a pattern using their “tile” feature. Once you have a few patterns that you like, you can create clothing items, phone cases, pillows, blankets and more that use that pattern.

You can take the patterns that you generated to a print-on-demand site like Printful and, when someone purchases one of your items, they will dropship it directly to the customer for you.

Find more AI income ideas at the AI-Income Database here.

Announcements:

• Interested in sponsoring this newsletter? We just crossed 74,000 subscribers and people are opening and clicking! Details about sponsorships are available here. Slots are filling quick as I only have one sponsor per week.

• If you haven’t joined the Discord community yet, there’s always amazing conversations and fun competitions happening in there. We just passed 6,000 members! Join up right here.

Thanks for checking out this week's FutureTools newsletter!

Have a tool that you want to share? Submit it here.

And don't forget to check out all of the newest tools we've just added on FutureTools!

Finally, if you have any feedback for me or ideas for this newsletter, get in touch! Replying to these emails will get to me and I will read all replies. I want to improve this and continue to make it something that you're super excited about each week. So all ideas are welcome!

You rock! See ya next week. :)

Matt Wolfe (FutureTools.io)

P.S. This email was 95% written by a human. GPT checked my grammar for me. :)