🔥 Weekly AI news: Did you miss it?!

👨‍🚀 The most important AI updates at a glance

Author

Die AInauten
September 15, 2024

This issue is brought to you by:

🔥 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.

→ Selection of the top posts of the last week ←

🌀 Text-to-video for free? MiniMax shows how it's done - and it's EPIC!

And yet another new text-to-video model! The Chinese AI company MiniMax is currently making waves. Why? Because their new model can generate damn good videos.

You can even use it for free without registering if you know the direct link. Here it is: https://hailuoai.com/video

Whether futuristic cities, ninja fights in a bamboo forest or beautiful nature shots from a bird's eye view, the tool spits out hyper realistic scenes - and all that without a Hollywood budget!

A few tips for use

If the page is displayed in Chinese, right-click in the browser and select "Translate to English" in the context menu (or simply enter your prompt and click on the 😎button).

English prompts are understood. Take inspiration from the existing prompts on the site, or use our new AI video prompt generator here.

It takes a few minutes to create the video. The output is 6 seconds long and in a solid resolution of 1280x720 pixels at 25 frames per second. With a little effort, you can do some really, really cool things!

Our take: This thing makes you want more!

As with every new AI tool, this is the worst version we will ever have. And the results for short clips are already extremely strong.

Longer video lengths and more creative control options are still missing - but are already in the works (e.g. picture-to-video, as we know it from the competitors Runway, KLING, LumaLabs and Co.).

We'll definitely be hearing more about MiniMax - use it now and take advantage of the free honeymoon period.

📰 AI news quickie: The HAI highlights from the industry

Finally, a few tidbits - here are the most exciting news. But first: check out this really well-made AI clip to get you in the mood!

Industry

Anthropic

  • Anthropic has launched Claude for Enterprise - with 500k context windows (almost 4x more than ChatGPT Enterprise!), native GitHub integration and full control over your company data.

  • And you can now experiment with AI at lightning speed - Anthropic releases starter kits for popular programming languages.

Funding

Research

  • Loopy shows how you can create a realistic video with lip-sync output using just one image and audio - currently only as a research demo ...

  • DeepMind has developed AlphaProteo, an AI that develops novel proteins and thus revolutionizes drug development and disease understanding.

  • A robot controlled by a fungus? Yes, that's possible!

  • Scientists have developed a Minecraft civilization with 1000 autonomous AI agents that communicate and cooperate with each other to solve complex tasks. Fascinating!

🎧 Podcasts: Some food for thought

🇪🇺 How marketing teams are successfully using AI

Let's start this issue with a very exciting study conducted by LinkedIn and our partner HubSpot.

The study examined the extent to which European marketing teams are already using AI and in which areas.

With very interesting results.

Many companies are already using AI to achieve greater customer satisfaction and have even been able to increase their returns.

How do they do it? Through the targeted use of AI in marketing! The report breaks this down even further.

The regional profiles are also interesting. There are clear differences between the Nordic countries and Germany, for example.

Compared to other countries, German companies do not yet see a general increase in returns. However, the outlook for the coming year is much more positive.

As is so often the case, strategic backing is sometimes lacking at management level.

Finally, the report also looks at specific measures, for example how best to use AI in marketing and what is needed to really roll it out profitably.

But we don't want to give too much away. Just get the whole 46-page study, completely free of charge, here:

And yes, we know the report is in German and you might not be able to read it. But don’t we have AI that can just translated the whole thing? Yes, we do - and the file upload feature from Deepl.com (a German company, btw) makes it a breeze.

We find it super exciting to get a look behind the scenes at how other European companies are tackling the issue and how successful they are with it.

🤳 How to combine several trained image AIs

We received a lot of feedback on our Deep Dive on the topic of "Train a image AI of yourself".

Thanks for that, and congratulations to the many AInauts who successfully trained their own model (LoRA)!

We have also received some questions about how to combine several models.

That's why we're going into more detail today on how to find LoRAs. (If you have no clue what we're talking about, please read the first deep dive here).

So, the cool thing is that you can train the image model FLUX not only on people, but also on many other things. Animals, styles, objects, products and even fonts are possible.

The process is always the same.

Here, for example, is a LoRA that was trained to create half real, half illustration images:

Or here is one that has been trained in the style of tarot cards:

Or how about pictures of a rainy Tokyo at night?

The possibilities are almost endless! So let's be clear: you can train LoRAs for all kinds of things.

Combining different LoRAs

Now we come to the exciting topic. You can combine several LoRAs and create corresponding images from them.

Thanks to Replicate and the Flux-Dev-Multi-LoRA model, the whole thing is super easy. Let's take our AInauten influencer AIlexa as an example:

For example, we can now combine the AIlexa LoRA with one of the LoRAs from above - for example with the Tarot LoRA.

To do this, first store the two Hugging Face paths here:

And now comes the most important step: you must address both LoRAs in your prompt with the corresponding trigger words.

For AIlexa, the trigger word we have defined is "ailexa" and for the tarot model you can see the trigger word in the example images - in this case "a trtcrd, tarot style".

If we combine these two trigger words, a possible prompt could then look like this:

The result? Here's our AIlexa, giving a talk in tarot card style:

Sure, with a style like this, it's no longer so easy to recognize. But woman, brown hair etc. This seems right.

Here is another example when we combine AIlexa with the half illustration LoRA:

Cool, isn't it? The whole thing works super simply, quickly, inexpensively and without a subscription.

Here is the prompt for the picture above: 

In the style of TOK, a photo editorial avant-garde dramatic action pose of ailexa wearing 90s round wacky sunglasses pulling glasses down looking forward, in Tokyo with large marble structures and bonsai trees at sunset with a vibrant illustrated jacket surrounded by illustrations of flowers, smoke, flames, ice cream, sparkles, rock and roll

How do I find other LoRAs?

Finally, a few quick tips on how to find other LoRAs. You don't want to have to train everything yourself, do you? There are essentially three good sources for this:

1) Hugging Face

Simply go to Models, Text-to-Image and enter "Flux" in the search bar. We need LoRAs that have been trained with the Flux model.

And lo and behold, you will find a long list of over 1000 exciting LoRAs, which you can sort by trending, etc.

2) Replicate

Of course, there are also many LoRAs on Replicate. This is practical because you can use them directly, and also via the API interface if you want to use them in automation.

The procedure is the same again: Explore - Search Flux - Find LoRAs.

3) Civit AI

We have saved the most extensive collection until last. You can find incredible LorRAs on Civit AI.

It is also important to filter here:

Simply select LoRA under Model Types and Flux 1.0 under Models. You can then also use these LoRAs on Replicate (you can also find examples there in the "Examples" tab).

That's all for today. In the next part, we'll take a look at how you can use LoRAs in your automations.

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