👨‍🚀 Good News From Europe?

Hello AInauts,

Welcome to the new issue of your favorite newsletter! Today with a short recap of the Rise of AI conference in Berlin.

We were extremely happy to meet so many AINAUTS in person - thank you, we are incredibly grateful for our community here.

So today a little more news, less practical tips. There will be more to try out again on Monday. Here's what we have in store for you:

• 🇪🇺 Good news from Europe?

• 🩺 Game changer: AI for our health

• 🤔 Conclusion: Europe really left behind?

Here we go!

🇪🇺 Good News from Europe?

The motto of this year's Rise of AI: Good News from Europe?
The question mark was palpable.

The opening panel started with Aleph Alpha and Silo AI, and it was clear: the mood is cautious.

Europe has no flagship models to show, many companies are disillusioned - the promised leaps in efficiency have not (yet) materialized, projects are stalling, expectations are being scaled back.

But that doesn't mean that nothing is working.

What we took away from our discussions with employees from SMEs, corporations, consultancies and panel guests: The opportunity lies not in building our own world-class foundation models - but instead in transforming our strong industries with smart AI! Selectively, pragmatically, productively.

A good example: the AI ecosystem around Paris.

It works here: talent, capital, political backing - and the awareness that it is also about infrastructure (e.g. affordable energy).

Mistral was born out of this. And it shows what is possible when research, the market and courage come together.

In line with this, there are two news updates from Mistral that are particularly relevant for enterprise customers.

Mistral Medium 3 is here

The latest and most powerful model from Mistral is here. The most important thing about it in one sentence: It performs almost identically to the state-of-the-art models from Anthropic and OpenAI, but costs up to 8x less.

The kicker for companies: It can be self-hosted.

Updates to Mistrals own chatbot Le Chat

At the same time, Mistral is launching a completely revised enterprise version of its AI assistant "Le Chat" - powered by Medium 3.

What's inside? Everything a company needs for real AI productivity:

• Enterprise search using your own data

• Agent builder for individual workflows

• Custom connectors for tools & data sources

• Document libraries & auto-summaries

• Own models & hybrid deployments (on-prem, cloud, Mistral hosted)

• Data protection compliant & 100% private by design

Their focus lied clearly on flexibility and control. You want to host Le Chat yourself? Sure thing. Build your own integrations? Do it. Integrate your own models? Absolutely.

Particularly exciting: teams can build their own agents without code, which talk directly to internal tools and take over tasks automatically. Writing content, analyzing data, organizing meetings - all via one platform.

Mistral shows that they are serious about enterprise AI. Open, affordable, flexible - but still powerful.

Of course, this is currently particularly exciting for companies. But all the features are also being rolled out for individual users.

We think the path Mistral is taking is fantastic. The quality holds up, and the focus is clear. We already wrote about Le Chat the other day and are trying to use it more and more in our daily tasks.

The German company Aleph Alpha, on the other hand, has focused on industry and governments and is going its own way.

This may also work, but we find Mistral's approach much more exciting.

🩺 AI in the healthcare sector

Another exciting topic at the conference was AI & healthcare, pharma and biotech.

AI is also increasingly finding its way into these areas, which in turn is also good for Germany because we have very strong players here.

And there are more and more language models that are specifically designed for this market.

At the moment, the creators recommend to not rely 100% on AI. But if you want to try out BioMistral, a biomedical model, you can do so here on Huggingface.

The latest news from OpenAI also fits in with this: with HealthBench, OpenAI has set a new benchmark for medical language models: 5,000 simulated doctor-patient conversations, 48,000 evaluation criteria, developed with 262 doctors.

The aim: to find out which models are really suitable for everyday clinical practice.

Surprise: GPT-4.1 nano beats GPT-4o - and alsoe comes with 25 times lower costs. A strong signal of where things are heading.

It is open source - if you want to test it, here are the GitHub evals.

A lot is happening

With AlphaFold, Google has already proven how LLMs can revolutionize protein research. Millions of complex molecular structures in hours instead of years - and the models are getting better.

The AI Scientist is also super exciting:

The AI Scientist project is an open-source system in which foundation models can conduct independent scientific research - from hypothesis setup to paper generation.

The results? Exciting and sometimes astonishing. AI writes entire scientific papers, tests ideas automatically, recognizes patterns and suggests experiments.

Of course, this should all be taken with a pinch of salt - but if you're wondering where the journey is heading - that's exactly where it's headed!

🤔 Conclusion: Europe left behind?

It was really exciting to talk to so many people from different sectors, and we’ve got an impression and spotted some patterns (which of course has no statistical relevance whatsoever).

But we want to share this because we know that many AINAUTS are asking themselves: Is my company doing the right thing in AI? What are the others doing? Where do I start? Are we missing the boat?

It feels like we are on the downward curve of the AI hype cycle. We are learning that AI does not work immediately as we want it to. That it doesn't make everything easier. And that transformation is difficult.

But honestly? That's perfectly okay for us.

Now is the time to keep at it. To build. We will definitely use AI even more! And we would like to conclude with two slides from the presentation by Adrian Locher, a successful entrepreneur and the founder of Merantix-Capital.

(They have built an ingenious AI ecosystem, and we will write more about Merantix Capital soon).

We have everything we need in Europe. Super research, great talent, a strong industry and regulation is moving in the right direction again after the excesses of recent years.

We just need to join forces and keep at it.

It is up to all of us to drive AI forward in Europe. Even if we are not technical folks. We are more enthusiastic than ever, and hopefully you are too!

With this in mind, thanks again to everyone we met in person. Feel free to share with us how your company deals with AI and what your view is topic. We are always happy to receive your feedback.

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 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.