🔥 Weekly AI news: Did you miss it?!

👨‍🚀 The most important AI updates at a glance

AI-HOI, AInauts!

Maybe you didn't catch all the news, tools, and hacks about AI last week, or maybe you've only recently joined us. Either way, here's our recap with all the headlines from the newsletter - just one click away!

Click the links to jump right to the article - or read our picks below.

• 👷‍♂️ How to implement your business idea quickly with AI

• 📰 AI-News-Quickie: The HAI-lights

• 📖 Book: The future is closer (... when we merge with AI) by Ray Kurzweil

• 🤖 Become a ChatGPT pro in 20 minutes
  (… and learn tricks you're not using yet)

• 📹 The problem with AI video - and why you should still wait

• 👨‍🎨 3 cool AI image tools that you don't know yet

→ Selection of the top posts of the last week ←

👷‍♂️ How to implement your business idea quickly with AI

The new AI tools are the do-it-yourself kit to get new business ideas off the ground - or to adapt successful business models for a specific niche. As the saying goes: "The riches are in the niches!"

How can you build a niche business?

Let's take a look on how to approach this. As an example, we take the popular book summary sites getAbstract or Blinkist and create our own niche version by using the latest AI technology.

Step 1: Validation of the business idea

Of course, you should first have an understanding of the market, customers, competitors, etc. before you embark on the adventure. You can use a GPT for this, such as the Validate Your Business Idea-GPT (our result here - just say "Create a graphic for me").

Step 2: Designing the website

To provide input for our AI partner, we take a closer look at the websites. A screenshot here, an HTML download there (file → save page → download webpage (single file - mhtml)) - and the first data for the AI is collected.

For the design, we then use Claude with the new artifact feature (activate it under "Feature Preview"). First, let Claude make a suggestion for a website/landing page based on your idea (or the input from the analysis of the first step). If needed, provide the screenshots etc. to guide Claude.

After a few ping-pong interactions, you tell Claude to make the website 5x better - and you've got your first draft!

Claude can also walk you through on how to get this page on the web. Alternatively, you can use a free AI website generator or a builder like Carrd.co to build a site yourself.

And for payment, there are plenty of options like Digistore24, Gumroad, Sendowl, etc. (we use ThriveCart).

Step 3: Google Gemini processes huge amounts of data

Now that we have the virtual shopping window, we still need the goods - in our case, the summary of relevant AI books.

This is where Google Gemini Studio comes into play, with a context window of over 1 million tokens. That's where we upload our first book, with detailed instructions on what the output should be.

We can get the right prompt for this from the Claude's Prompt Generator.

"Create a comprehensive, concise English summary (you form) with bullet points from this PDF - min. 500 words per chapter" becomes a detailed instruction, which we pack into Google AI Studio.

And lo and behold, Google Gemini conjures up a concise summary that will make even the laziest reader happy.

Step 4: Marketing à la ChatGPT

Now that the product can be put on the virtual shelf, we ask ChatGPT for a marketing plan.

To do this, you can tag the Marketing GPT and simply add it to the conversation from step 1 with @MARKETING - and have a detailed plan drawn up.

A little outreach and SEO here, a few social media tricks there - and we're ready to conquer the digital masses.

(Ok, it's not quite that trivial, but you get what we mean 😎.).

Our take: From copy-paste to Cre-AI-te

In the world of digitalization, anything is possible - and with the right tools, a clone can become a king!

But the real key lies in intelligent inspiration. With AI at our side, we can not only copy, but also reinvent and drive innovation.

It's about being inspired by existing ideas to create something unique. Be the trendsetter in a niche that not only follows, but leads. And have fun experimenting!

P.S.: You can find the summary of the book at the end of this newsletter.

📰 AI-News-Quickie: The HAI-lights

There is something going on, so we've put together the most important news and updates from the AIniverse. Here we go!

• Anthropic with Claude 3.5 Sonnet has blown us away for the past 10 days, especially with the new Artifacts feature- it steals the show from GPT-4. Pro users can now also create projects, which include uploaded files and custom instructions. These Projects are the GPTs of Anthropic, so to speak - but unfortunately only shareable if you have a Team-account.

• OpenAI has delayed the release of their highly anticipated Voice Assistant (which can now also generate sound effects!), while the newly trained model CriticGPT challenges the code generated by GPT-4 and thus ensures better results. The content from TIME has been licensed by OpenAI, and the Mac desktop app is now available to all.

• Google has made Gemini 1.5 Flash and Pro generally accessible via Google Cloud. It can process huge amounts of data with up to 2 million tokens. And a small nano-model will soon live directly in Chrome under window.ai.

• Google Vids was launched for Workspace Labs (plus: Gemini for GMail is rolling out now). With Vids, you can create storyboards, and a variety of templates help you put together scenes with stock videos, images and music.

• Suno and Udio are at the center of copyright lawsuits filed by major record companies such as Universal, Sony and Warner. They are demanding compensation of up to USD 150,000 per recording ...

• YouTube is taking a different approach, negotiating with the record labels to train their AI music tools legally. And soon, they may even offer virtual chatbots of your favorite creators!

• Character AI now has the option to speak to AI avatars via "phone". This feature enables interactive conversations with fictional characters. Just install the mobile app and get started - although some non-English languages, like German, are still quite bumpy ...

• Vidnoz.com makes it easy to create short AI avatar videos based on a photo or available avatars.

• WebSim.AI is a cool app that lets you create entire websites in real time. Simply enter what you want to see and it will be generated. Fascinating and visionary!

• Figma has introduced AI-powered design features that allow app designs to be generated from text prompts. These Beta-features promote creative design. The only thing missing is that the app itself is also being programmed by the AI.

• Factory AI's autonomous systems, the so-called. Droids, dwarf the competition. Machines build machines...

• Pliny the Prompter has exposed the weaknesses of all (!) great language models (our interview here, first 😎) - and now the Financial Times, Venture Beat, BBC and Co. also caught on and are talking about it.

📖 Book: The future is closer (... when we merge with AI) by Ray Kurzweil

Renowned futurist Ray Kurzweil has released the long-awaited sequel to his 2005 bestseller The Singularity is Near. The appropriate title: The Singularity is nearer: When we merge with AI

We picked up the book straight away. It's over 500 pages long (over 200 pages are filled with references and sources), with lots of graphs and data - here you can find our summary in English.

He paints a wild picture of our future, in which we will merge with AI by 2045 - a sci-fi mash-up of man and machine. Welcome to the Singularity, or in other words: "Just don't die - you might live forever!"

But it is also a fascinating look into what’s coming, highlighting both the opportunities and challenges of technological developments. And a call to prepare for the changes and use the opportunities to shape a better future.

The boundaries are dissolving as we race towards the singularity, driven by an exponential development of computing power.

AI is developing into hybrid super brains that are redefining the way we see ourselves. Soon, your smart assistant could know more about yourself than you do!

But don't panic: Kurzweil promises that all this technology will make our lives better. Think of a universal basic income, more time for what really matters, personalized medicine and nanobots that screw around in your body - the revolution is already just around the corner.

But where there is light, there is also shadow. But with clever development and a little bit of regulation, these can be brought under control. So he says ...

And if you'd rather smile than read the book summary, here you go!

🤖 Become a ChatGPT pro in 20 minutes
(… and learn tricks you're not using yet)

If you've been reading us for a while, you're probably already pretty familiar with ChatGPT and prompting. 🙂

Nevertheless, we would like to recommend a free e-book to you today, because even we’ve learned something new.

HubSpot has published a free ChatGPT Playbook on how to use ChatGPT at work, which is really helpful if you want to supercharge your tasks.

Here's what you can expect:

On the one hand, it gives a super quick, simple overview - whether you're a beginner or advanced. And above all, everything is very close to practice.

Sure, 100 prompt ideas are cool. But it also has some really exciting tips that are not very well known.

This is a comprehensive guide on how to enhance work productivity with generative AI! 😉 Definitely something for insiders, and very useful.

Get the free e-book right here:

📹 The problem with AI video - and why you should still wait

Let's start with the most important news of the last couple days:

The new video model of Runway Gen-3 is now available for everyone. We have already reported on it and have to admit that the videos look incredible.

Here are a few highlights:

The associated prompts seem downright simple when you look at the magnificent results.

Here you can find more videos and the prompts.

Gen-3 really is the most impressive AI video model since Sora from OpenAI.

But now to the problem with AI video ...

Despite the impressive videos, we would still advise against investing too much time and effort into AI video.

The problem is not the quality of the videos - but rather the price of generating them. The computing power required for such videos is high, and therefore everything is expensive (or if its free, it is very limited).

For just under 15 dollars a month, you currently get 62 seconds of Gen 3 video ... But by the time you have a scene that you really like, you've certainly used up your entire budget.

And you can only generate individual clips that are 5-10 seconds long. In other words, longer videos can quickly become costly and require a lot of trial and error. That's why think that the AI video trend comes with some drawdowns for most of us.

What's more, these videos are also generated without sound. Of course, there are AI-developments on the horizon, such as the audio-for-video feature just presented by Google DeepMind. This allows the AI to generate the appropriate audio track for any video, even with dialog. Very impressive - but we can't use it yet ...

It is therefore worth waiting a little longer before experimenting with it more intensively - until prices fall, and the technology becomes more affordable.

In the meantime, we can simply enjoy the impressive inspirations (they are really awesome):

👨‍🎨 3 cool AI image tools that you don't know yet

Last but not least, we have some exciting tools when it comes to image editing and generation. Best of all: you can try them for free!

Magnific.ai - How to change the background and lighting

First and foremost, Magnific.ai has presented a new feature: Relight! This allows you to dynamically change the lighting and background of images. What does that mean? LUMOS MAXIMA, but see for yourself!

Yes, a few pictures really do say more than a thousand words.

What is Relight suitable for?

Whether landscapes, characters or backgrounds - this tool enables realistic and diverse scenes to inspire your creativity. Even for non-experts, it is intuitive to navigate, and anyone can create stunning images or edit existing ones.

There are some use cases that are particularly well suited:

1. Advertising: Products can be easily inserted into different scenarios, making it easier to adapt to different markets or campaigns.

2. Portraits: Relight can dynamically adjust portraits and create different lighting moods, ideal for fashion photography or creative projects.

3. Creative projects and new interpretations: From artists to designers, Relight offers a playground for creative experimentation to create unique visual experiences. You can also revive old images and give them a fresh perspective

Change the images with a simple prompt, a reference image or a light map on which you can draw the lighting yourself.

What is possible, where are the limits?

The Relight function is impressive and allows dynamic lighting adjustments that really change the image. The clarity and depth of field of images is excellent, especially when using the "Creativity" slider. However, it is possible that the original image is sometimes altered slightly.

Magnific AI + Freepik = ❤️🌏

Magnific AI was founded by Spanish entrepreneurs Javi López and Emilio Nicolás. In May 2024, the company was acquired by Freepik to expand its suite of generative AI tools. This acquisition unites innovative forces and strengthens Freepik's position in the global market.

And that brings us straight to the next tool ...

Freepik.ai - an all-rounder with a large library

To be honest, Freepik.ai wasn't even on our radar.

But Freepik has a large library of content and tools

You can download an extensive content library with high-quality photos, videos, vectors, PSD, AI images, icons, etc. - and you can also do a lot with their AI tools.

Freepik's most important AI tools are an image generator, an extensive Mockup Generator, a Reimagine tool to freshen up old files, Sketch to Image, Background Remover, an Upscaler and a Retouch tool - a free alternative to Photoshop's "Generative Fill", so to speak, with which image areas can be filled with AI-generated data.

Leonardo.ai - Generate images with text, edit them precisely

Also, Leonardo.ai recently introduced its new "Phoenix" model, which creates really interesting images - especially if you want text on them (Hello, Ideogram and DALL-E, you’ve got some competition!). And: Leonardo also has an API-interface that you can use.

What's new about the Phoenix model?

• Proper text in images: Perfect for banners and logos, as it generates clear and precise text - something other tools often struggle with.

• Precise prompt conversion: New standards in the accuracy of image generation, detail and image quality - ideal for detailed visual content.

• Prompt extensions: This function converts simple prompts into detailed, hyper-realistic descriptions.

• Edit with AI: Edit the images interactively to make specific changes.

• Advanced creative control: With features like the Prompt extension, you can easily customize and refine generated images.

For more details, visit the Leonardo YouTube channel, or watch the video here.

🪄 Our take: The choice is yours!

Each tool has its strengths and is suitable for different use cases in the creative process.

• Select Magnific.ai if you want to transform existing images and adjust the lighting.

• Use Freepik.ai for a wide range of design tasks and when you need access to an extensive resource library.

• Opt for Leonardo.ai if you want to generate detailed images with text and need precise control over the result.

P.S.: And if you're a die-hard Midjourney user: the website srefhunter.top has collected hundreds of Midjourney sref codes. Simply use the parameter "--sref" followed by the number of the desired look, and you'll have images in the corresponding style in no time.

Your AInauts, Fabian & Reto

Follow us on Twitter & LinkedIn!

Your feedback is essential for us. We read EVERY comment and feedback, just respond to this email. Tell us what was (not) good and what is interesting for YOU.