👨‍🚀 26 words for better AI videos & photos

In partnership with

Hi AInauts,

Welcome to the latest edition of your favorite newsletter! The news is rather dull at the moment, so today we have some things for you to try out and implement yourself.

Here's what we have in store for you:

• 🇪🇺 Lumo AI: Chatbot with data protection from Europe

• 🎥 26 words for better AI videos & photos

• 💭 Google Gemini now builds presentations

Let's go!

Free email without sacrificing your privacy

Gmail is free, but you pay with your data. Proton Mail is different.

We don’t scan your messages. We don’t sell your behavior. We don’t follow you across the internet.

Proton Mail gives you full-featured, private email without surveillance or creepy profiling. It’s email that respects your time, your attention, and your boundaries.

Email doesn’t have to cost your privacy.

Ditch the Gmail data grab

🇪🇺 Lumo AI: Chatbot with data protection from Europe

In a mastermind group with other entrepreneurs, the discussion about data protection and data security came up again recently—also in relation to the big US tech companies, on which we are all dependent.

Emails, calendars, cloud storage, and new chatbots such as ChatGPT and others were discussed.

The name Proton from Switzerland quickly came up. We had long been familiar with Proton in relation to secure emails, calendars, VPNs, and cloud storage.

More on Proton below... But after creating an account there, we came across Lumo AI.

Proton's own AI chatbot, which focuses on privacy.

Of course, we think this is a great feature and have tested the bot for you.

First, a few nerdy basics:

• Lumo AI does not have one language model in the background, but several.

• All are open source

• Mistral from France is mainly used.

• But also OpenHands or NVIDIA

• The bot always decides which model to use based on the request.

The functions are super simple. Simple chatbot. You can load data into the chat as knowledge. Unfortunately, only text, CSV, etc. No images.

There is also a web search for current content. This is needed, as the last training update appears to be from April 2024...

The context window is between 8k and 128k, depending on the model. Which, of course, is not that huge.

Conclusion on Lumo AI

It's a super simple chatbot. Simple, without a lot of bells and whistles. The sole focus here is on privacy.

It is fast, and we were satisfied with the results during our test conversations.

Of course, it can't compete with the big players when it comes to very complex topics. It has significantly fewer features. But that's not the goal.

However, Lumo AI can be extremely useful, especially for sensitive queries where you use sensitive data about yourself. Proton generally makes a lot of sense.

Why you should take a look at Proton

As mentioned at the beginning, we have now become Proton customers. But not just because of Lumo AI.

Proton was founded back in 2014 by a few scientists who met at CERN in Switzerland. (Yes, exactly, the thing from Illuminati)

From the outset, the team had one goal: privacy, data security, and digital freedom. They now have over 100 million users.

It started with email, then calendars, and now it's a whole suite of tools:

• Email

• Calendar

• Drive (cloud storage)

• VPN server (especially important for us AI nerds 😉)

• Password manager

• and now the new Lumo AI chatbot

Everything encrypted. Privacy first. Open source. Based in Switzerland.

So many services in one tool.

As you may have noticed, we are big fans. If you would like to take a look for yourself:

🎥 26 words for better AI videos & photos

Christmas is just around the corner. And once again, everyone is talking about this year's new Coca-Cola Christmas ad.

Every year, Coca-Cola attempts to create an advertisement using AI as much as possible.

There's no question that we're sure there are still a lot of professionals involved here, and that production is probably more expensive than the annual turnover of many startups.

Nevertheless, it's really exciting to see how stark the difference is even between 2024 and 2025. This year, especially in terms of length.

The models have improved in consistency and adhere to last & first frame, making it easier to create longer videos.

This is reason enough for us to revisit the topic of video and photo prompting.

We came across a very cool thread on X that shows why it makes sense to learn certain words or some film speech in order to get really good results.

If you know a few technical terms related to camera angles, lighting behavior, exposure, and composition, you can create superb creative content.

1) Let's start with camera angles (We still believe that English prompts work better):

Define the camera angle in your prompts. Wide, low, over the shoulder (OTS)...

2) Light behavior

Now it gets a little more exotic: Here are a few more examples you can use in your prompts.

3) Light in general

Describing the lighting is a super important element. Here are a few simple words that should significantly improve your pictures and videos:

Here is everything together in an example prompt:

cinematic video of a lonely astronaut walking through a snowy forest, low angle shot, over the shoulder, soft camera motion, light falloff, dramatic rim light on the helmet, snow particles glowing in the light, natural daylight, soft shadows, shallow depth of field, high detail, 8k, ultra realistic, volumetric lighting

We are well aware that you can always have all prompts written for you. But if you really want to influence the results, and especially if you want them to be consistent, it makes a lot of sense to know a few basic terms.

(Or simply save this newsletter somewhere 😉)

💭 Google Gemini now builds presentations

Finally, a quick topic for anyone who already uses Google Gemini or might want to try it out.

In general, it's about time Google finally released the new Gemini 3 model, because the old models seem a bit outdated compared to ChatGPT and Claude.

But especially if you work a lot with Google and its various tools such as Docs, Slides, Sheets, etc., Gemini is truly one of our favorite tools.

The features are getting better and better here!

Now you can activate the canvas function in Gemini and create a Google Slides presentation directly.

All you have to do is activate the Canvas function and write in your prompt: "Please build me a presentation."

If you are satisfied with the result, you can open the presentation directly in Google Slides and continue editing it.

You can also talk to the AI directly in slides and even add your pre-trained assistants (Gems).

Google is doing a wonderful job here. All tools and services are interconnected, and work can be done across multiple tools with the help of AI.

Here is one of our workflows:

• We use an assistant (Gem) trained for our brand.

• Develop a concept for a workshop together

• Activate Canvas and let's build a presentation

• Open this presentation in Google Slides to finalize it.

Done.

There are many more great possibilities. Maybe we'll do our own Gemini training. If you're interested, give us your feedback below and write a comment.

You made it to the end—thanks for reading! We’ll be back soon with even more updates.

Reto & Fabian from the AInauts

P.S.: Follow us on social media - it motivates us to keep going 😁!
X, LinkedIn, Facebook, Insta, YouTube, TikTok

Your feedback means the world to us. We read every comment and message—just hit reply and tell us what you think!