@@ -0,0 +1,39 @@
+# Workshop Feedback (2026-06-11)
+
+1. I learned what professional alternatives there are to ChatGPT, and that they can all be used on a unified, easy-to-use platform.
+
+2. Important information sources. I learned how many things AI can do.
+
+3. For code writing. New info about different programs.
+
+4. Getting to know the OpenCode interface, application possibilities.
+
+5. Making it easier to work with many databases, learning about different elements, using things through APIs.
+
+6. The whole OpenCode infrastructure and the possibilities it holds. It will be a big step forward compared to the GPT-based code I've been using.
+
+7. Now at least I have some idea of what I *don't* know about modern AI tools.
+
+8. The intro was really good, the stack layer part.
+
+9. The fewer vs. more skills distinctions were great, and that cheap models can work just as well. But in practice I saw that there are dangers in letting the agent loose on your machine.
+
+10. How should you logically build up a research workflow — how to incorporate checkpoints, pre-planning?
+
+11. OpenCode can run and quickly format charts well.
+
+12. I can get started.
+
+13. Basic concepts, overview — I understand better when people talk about this stuff. I like the course, good course structure.
+
+14. I understand better what an Agent is and what it's for. For quick analysis or a pilot program, for instance, to see if there's anything there — I'd use it more boldly now. I can see I could use AI for the crappier parts of research. Especially short writing, READMEs.
+
+15. Using a proxy is a super idea. Using OpenCode cautiously.
+
+## Post-workshop email feedback
+
+16. The course was very useful for me too, thank you so much! On Friday I was still experimenting out of curiosity to see what it can do. I had it make a chart of the most frequent topics based on presentation titles from the last conference I attended. It made the basic chart surprisingly well. Editing it afterwards was more cumbersome, though. The main problem was that the session probably expired in the meantime, and all the earlier information was lost. Related to this, I'd like to ask: is there a way to keep a session alive for longer? Also, you mentioned that the system can't be used with Claude. I have a Claude subscription, so I'd be curious whether there's some creative solution or workaround that would still make it possible. I also haven't been able to solve the Dropbox problem yet. The files show up in the file manager, but I can't open them from within the program. It's not a big deal, but if you have any ideas or experience with this, I'd be happy to hear them. I know this is a lot of questions at once, but maybe that also shows how useful I found the course. This week I'm already going to try using the system in my actual, "live" work; until now I was mostly just experimenting and getting to know what it can do. Thank you again so much!
+
+17. I just had it generate data citations for an article that uses a ton of databases — it neatly looked up the BibTeX citation on every website and wrote the LaTeX code. Saved me a huge amount of time.
+
+18. I just had to make a README file for a paper's code. It's pretty tedious work, so on Miklós's suggestion I went at it with OpenCode. The result was far from perfect, but it still saved me a lot of time (and hassle). So thanks a lot!