🔥 This robot is scaring us ...

AI-HOI AInauts,

Welcome to the Monday issue! Ready for the week?

To get in the mood, we tested tools and looked beyond the horizon to see what's going on in the field of video and robotics. It's super exciting what's coming up. Most of it, anyway …

That's what we have in store for you today:

• 🤗 The emotional AI avatars are among us

• 🤹‍♂️ Images with text - The long road to perfection at Ideogram

• 🐶 Thermonator, the fire-breathing Robodog

• 😈 Videos: What is AI, and where is it going?
  (PLUS: Vidu, the Chinese Sora alternative)

Ready? Let's go!

🤗 The emotional AI avatars are among us

Many videos on Instagram, TikTok, YouTube, and other social media platforms are already AI-powered. However, due to a lack of emotion, you can often still see that the AI is working in the background. However, this is changing rapidly, as fake avatars are becoming increasingly realistic … last week we saw VASA, now this.

Synthesia introduces the latest generation of emotional avatars

We've been using Synthesia for “years” … or let's say: since the average Joe like us was given access to generate videos. Lately, we've often ended up using HeyGen - because it's cheaper and easier to access. But now there's another big update from Synthesia …

The brand new expressive AI avatars from Synthesia can now laugh, cry, shake their heads and so much more - all based on your input!

You can try it yourself here - we also did a test in German, you can find our video here. Important to know: Synthesia focuses primarily on corporate customers, less on hobbyist consumers.

What else is important?

• They are the first avatars generated entirely by artificial intelligence that can display human emotions with facial movements and micro-expressions in real time.

• Unlike other avatars, they “understand” the script and can proactively adapt their expressions accordingly.

• Simply enter text and the virtual companions act like real people, combining speech and facial expressions in a way that is virtually indistinguishable from real people.

• So far, around 200,000 people have created more than 18 million videos in 130+ languages with the 225 Synthesia avatars to date.

• You can see the evolution of the avatars in the video here, more details here.

• Companies also have the option of going to the studio and having their own avatar created.

The big problem with Synthesia videos - and what to do about it

Synthesia is also aware of the risk of misuse and has implemented strict measures to prevent this. Although, they prefer to talk about “synthetic media” instead of “deepfakes”.

The C2PA membership promotes the transparency of AI-generated content with content certificates and watermarks.

Nevertheless, there have already been incidents in which the technology has been misused for propaganda in Venezuela and false reports on pro-Chinese social media accounts. The guidelines have therefore been tightened, and the security team expanded to filter out malicious actors.

The will is there, and Synthesia seems capable of learning and genuinely intent on leaving as little to burn as possible.

🤹‍♂️ Images with text - The long road to perfection: An honest test of the new Ideogram features

The image generation platform Ideogram recently received a major update. Reason enough for us to put it to the test.

Our take: We are not there yet! But soon …

Ok, we think Ideogram is really cool. Unfortunately, the new features are only available to premium customers. But you can still generate cool images even in the free version.

If you look at the statistics above, you can see that half of the words still have errors. But we can report that it works better at English, than German.

But our attempt to generate a cool AInauten comic cover with correct text in German failed - it always had text fragments on it that didn't fit …

The bottom line is that this is a classic case of “We're 80% there!”

But we are certain that the text generation will be error-free by the end of the year!

Until then: it's best to generate a pure image and insert the text with the image program of your choice (we use Canva.com and experiment with Microsoft Designer).

🐶 Thermonator, the flame-throwing Robodog

May we introduce: The Thermonator is the first flame-throwing robot dog for the home. Or alternatively, the wasp drone for the wall of flames from the sky …

And if you're now wondering what you need it for … the provider's FAQ shows a few uses that aren't quite as dystopian as you might think at first glance:

• Soil cleaning or controlled agricultural burns

• Melting or removing ice and snow

• Burning of weeds and other unwanted ground plants

• Controlled fires to prevent forest fires

• Pyrotechnics & film production (… and even Burger King commercials)

If you also want to clear your driveway of snow and ice - or weeds - with the new robot dog, you can do so from $9,420 (and they also deliver outside the USA). Just don't burn down the house …

Our take: Something the world doesn't need - but it can be done differently!

Well, the flame-throwing Robodog certainly created a buzz. But it is more of a spectacle than a product that mankind has really been waiting for. And it stirs up fears, because you don't need much imagination to see negative horror scenarios in your mind's eye.

In other areas of robotics, there are some interesting announcements. We are always looking for assistance in the household … because there are now many prototypes waiting for the reality check.

Elon Musk has also announced that Tesla is now a robotics company. Its Optimus robot could even be on the market by the end of 2025 for under $30,000. We will see - Tesla is notorious for not delivering on its promises.

P.S.: And now it's also official that not only AI drones have a place on the battlefield, but that an AI-controlled F-16 fighter jet competed against a human pilot under the supervision of the American DARPA. The winner was not announced …

😈 Videos: What is AI, and where is it going? (PLUS: the Chinese Sora alternative)

We may be AI fanboys, but we are well aware that everything has its downsides. Therefore, we would like to conclude with two videos that we can highly recommend. The sentiment is not only positive, but the risks and dangers are also addressed in clear terms.

From AI's potential to emulate human intelligence to the ethical issues shaping its development, these two videos provide deep insights into this “brave new world” of current and future artificial intelligence.

And as a little tidbit at the end, here's the new TED promo video - proudly presented by OpenAIs Sora & Paul Trillo.

By the way, Chinese tech company ShengShu-AI and Tsinghua University have just launched the presented the new AI model Vidu - the answer to OpenAI's Sora.

Vidu creates 1080P video clips up to 16 seconds long using a combination of diffusion and transformer technologies. At first, there were only a few short 3 to 5-second clips, and now longer clips have also appeared. It's certainly impressive, but not yet Sora level. Right?

We made it! The AInauts will be back soon, with new content for you.

Reto & Fabian from the AInauts

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.