Public speaking

I did some public speaking lately. Namely on Data Festival organized by University of Economics in Prague and HotSpot AI conference organized by Profinit company.

At Data Festival I spoke about our data product for marketing optimization. Mainly from point of view of in-house development, iterative discovery of requirements and need to look at all of this as holistic stuff. Not AI model and the rest. I like, that EU AI Act introduces “AI system” term, that is more suitable for industry setting than “AI model.” Slides

At HotSpot AI I did a 30 minute about R&D/AI department treating its work as an assembly line. Mentioning Google’s hybrid approach to research and some of our use cases. Slides

Blogs are not dead

Nor is reading. By some coincidence I stumbled into Read more books with great further links such as: Read. or Writes and Write-nots. I have a feeling (now reinforced by aforementioned posts), that by reading long form I am learning more and getting more advantage over my former self. My highlights:

The average American spends 16 minutes per day reading. Only 46% of Americans read a book in the past year, and only 18% have read ten or more. Meanwhile, we spend nearly three hours watching TV and two hours and fourteen minutes on social media (often while watching TV).

If you’re thinking without writing, you only think you’re thinking.

And somehow controversial “Learning Styles” are Largely Bogus:

There’s this theory that everyone has a different way to learn; some people need demonstrations, some people need a lecture, some need a discussion, some need a book. If we just matched everyone to the right learning style, anyone could learn anything! Maybe you learned differential equations by reading the book and doing a ton of exercises, but maybe I could learn the same thing with watercolors. It’s just not really true. Successful people converge on three ways to learn: lots of reading time, some exercises and projects, and some conversations with people who are slightly ahead of them. I simply can’t name a single person who became an expert on a topic by watching lots of videos. Even if the topic is video! Brian Stelter routinely puts his foot in his mouth when he talks about TV, even though his job appears to consist of watching it for 18 hours a day. In fact, you can use videos to perform a sort of error-archaeology. The people who learned about the crisis from watching Too Big To Fail are wrong in a different way from the ones who saw The Big Short instead.

And for the blogs. Thorsten Ball has a newsletter providing a lot of high value blogposts, weekly. How does he do that? The #13 touches on so many interesting topics I am rereading most of them.

I thought blogging is basically dead and everyone has only X and publishes there.. Boy I was wrong.

# Insight into Python 3.13

Software engineering daily podcast had python developer as a guest and he shared a great amount of insight on small vs big bets. GIL and JIT, how they see it and approach it. Also on importance of predictability of release cycle.