I am focusing on the wrong things

This is a rant, but hear me out. Instead of working on how to utilize AI in my own field, I am:

• spending a felt 30 minutes per day to authenticate and re-authenticate to all the different digital properties of Helsinki University that I use.
• finding creative solutions for problems that have been a Helpdesk request since October 2025 (without any action, but a massive amount of emails).
• evaluating Bachelor’s theses (see also the P.S. below)
• giving lectures to which only two or three students show up. Actually, I just broke my no-show record with the whopping number of 1 (one) student. I can still break this record once!
• making a work plan for the academic year 2026/2027. Ok, that didn’t take long, but it is emotionally extremely draining for me to have the university task dashboard (“SAP Fiori”) always full of work tasks that are absolutely not contributing to scientific progress or student education, but which keep sending me phone reminders while I am trying to focus on things that matter. I feel like Honorius tending to chickens and pigeons while the Germanic tribes were dismantling the Western Roman Empire.

Don’t get me wrong: I genuinely love the lab work and the interaction with students! And some Bachelor’s theses are amazing reads (see the P.S. below). But almost everyone at our University is asleep when it comes to using AI. Hardly anybody has the resources to do meaningful work. I am not talking about researchers who research AI. I am talking about researchers in other fields that could benefit massively from AI, such as life science researchers like me.

The university-provided systems run out of tokens in about 20-30 minutes of use, and there is no programmatic access, which is mandatory for the type of work I would need to do (e.g., integrating AI into cloning workflows). To keep up with the rapid development at least somewhat, I recently bought myself a €2000+ computer with a somewhat decent graphics card to run local LLM models. Local is the keyword here. In one of our AI-themed educational events, our lawyers have stated that once data leaves the local machine, the risk increases manyfold, and sensitive data must never leave the local machine at all. When I interpret this correctly, that also prevents me from using most of CSC’s computing environment. But I fear that our lawyers don’t understand what they want to regulate. Nobody really understands AI. We share this handicap with most governments around the world.

Run-of-the-mill university computers are sadly only slightly enhanced typewriters, totally unsuitable for any creative and innovative scientific work that requires serious data crunching. Even though I have only time in the evenings to play around with AI, the results are quite impressive: Simple coding work that would take me 2-3 days to complete is done by AI in 20-30 minutes. And I have scientific coding projects that I never started, knowing they would take me years to complete. While I might be able to complete them now in a few weeks, I still need to find these few weeks. As every spring, I have high hopes for the summer!

P.S. I love to give feedback as long as it is formative. But evaluating theses is rewarding for both me and the student only for PhD theses, because they are the only ones that get modified and can improve based on my feedback. However, some BSc theses are really good reads! One of our BSc theses, which discussed the emergence of next-generation doping technologies, was developed into a full-fledged review article, and it was just last week accepted for publication in the IF10 JUFO3 journal Sports Medicine  !