Thank you very much for the invitation

the very kind words.

First and foremost, it's a great pleasure for me to be here.

I named the company EMI

which exists since approximately a year after EMI-Nöte.

So it's a special honor to be here in Erlangen

as you all know.

It's true

I now have a double affiliation.

My heart beats for what can be done in engineering with the help of AI from a

let's say

realistic

point of view.

And I will try to give a realistic

positive future in trying to blend what I consider

is flabbergasting and exaggeration.

And this is how we will introduce the word reference models

which is

I think

a future

in engineering which is realistic

which is doable

and which is also in the sense of

a new tool helping numerics and improving the way we engineer.

So the way or the main messages I want to convey today is I want to quickly touch up

or I want to touch up about the topic of AI for engineering.

So everyday processes which are data-driven

where you use simulation

where you use your

numerics and how this can be enhanced and is going to and is already enhanced by AI

which is a very different world than LLMs.

And I will show why I think that reference models are the way forward.

And I will also show you a few of these models we are already building.

This is very

very opinionated

very much what I believe in

very much how we got also

funded, our company.

It's my personal opinion.

Opinions change over time.

Don't take me too serious.

Build your own opinion based on this talk.

So if we go to the current situation in the AI world

we see basically a pattern.

And if I have to predict what's the next breakthroughs in five years, I don't know.

But what I will surely know is that these breakthroughs are built on transformers.

These breakthroughs run on NVIDIA GPUs

which are built at TSMC

which are built with ASML

machines using Zeiss technology and Trump laser techniques.