I can’t take that AI bullshit anymore. At a recent conference a “worldwide recognized AI leader” told the audience: “Machine learning is just for insights extraction—it’s not that relevant anymore.”
I wanted to shout out: “Quit the shit!” But I kept quiet. I decided to do better: as program director of the Machine Learning Week Europe conference.
Yes, it’s still called “Machine Learning Week” and not “AI Week”.
After all, at the core of every AI application is a machine learning model (MLM). To illustrate the difference, I used a simple comparison during a meeting with a CEO of one of our clients (and yes, perhaps too simple—but every analogy has its limits, and its purpose is to make a point):
🚗 AI is the vehicle. A vehicle is more than just an engine. It needs tires, a steering wheel, and a hundred other things to fulfill its purpose. An AI application requires more than just a MLM: a user interface, data pipelines, etc.
⚙️ The machine learning model is the engine. Although the engine needs many components: cooling, oil, etc. The same applies to MLM: it needs model performance indicators, MLOps, cloud computing, etc.
⛽ The data is the fuel. There are different types and qualities of fuel. A vehicle needs the right fuel. The same applies to AI and data. Data drives AI. (It’s no coincidence that I named my company “Datentreiber” 12 years ago.)
And here we’re just talking about a simple vehicle. Complex vehicles—aircraft, ships etc.—have multiple and different engines. The same applies to complex AI applications: multiple and different Machine Learning Models.
Also, not every engine is equally suitable for every vehicle. An engine for a Formula 1 race car is completely different from one for a Fendt tractor.
Not every AI application needs an LLM. Some perform better with a much more simpler MLM.
The engine must fit the vehicle, and the vehicle must fit the intended use.
The MLM must fit the AI application, and the AI application must fit the use(r) case.
Exactly—that’s our job: figuring out when what fits best. That’s design and engineering.
And that’s exactly what Machine Learning Week Europe 2026 is all about: an exchange of experience and expertise. When do I use which tool and which technique for which use case?
📈 Have you figured this out for a specific case in your industry or application domain? Then present your solution in a case study.
🤿 Do you know of a tool and/or technique that can help others? Then do a deep dive.
👩⚕️ Would you like to help others with your knowledge? Then apply for a clinic.
With these and other formats, you can apply for a slot at Machine Learning Week Europe 2026 until July 15th:
👉 https://machinelearningweek.eu/call-for-speakers/
Not sure yet what topic to choose or how to present it? Then let’s discuss it together:
👉 https://calendly.com/martin-szugat/
Just one rule: No bullshit.
