Thank you for highlighting this research!
At first glance it's interesting that sigmoid functions re-emerge as more useful using the approaches evaluated in that article.
It's really strange that the prompt templates are being applied and called AI and no reference is made to what LLM is used to generate the output based on the prompt template and transcript provided.
A cursory look at rhe source didn't make it clear what LLM was used either.
It's also weird that the prompt uses the term 'hidden message" as an antonymn for "overt message".
I actually like this a lot. Why not make computer things easy for people using them? In environments where people send forms as non-editable .pdf files, this is great!
The good news is that the author has a third edition coming out with updated material. But it won't be ready until August 2024. I'm assuming you won't want to wait for that, so I suggest using the current edition to get used to the language and then going through the list of updates and differences in the blog post and "figuring out" how to do some of them on your own.
Also, the author has a book for people who finished Automate the Boring Stuff and want more guidance on good, effective practices with less focus on people with no prior knowledge of programming, Beyond the Basic Stuff with Python, is also available to read for free online.
I should mention that if you have the means, I encourage you to pay for the eBook version of these books on the publisher's website since this allows the author to continue creating updated and high quality content that's free for those who feel that $35 is a strain on their budget. If not, no worries, that's why I paid for his first edition a while back. Al is pretty active online and in the fediverse (@AlSweigart@mastodon.social) and seems like a good dude. He deserves some love, so please give him a thank you at a minimum.
This was a weirdly long reply, but I hope other people subscribed to the Learn Programming community on Programming.dev see it and find some value in it too.
The most pragmatic approach to that is Automate The Boring Stuff with Python which is free to read online. But it might be a little dated and you said you want to learn some basics which I take to understand as underling fundamental theory of programming which it doesn't provide.
Think Python, 3rd edition, which is also free to read online and was just updated to use Jupyter Notebooks, is a great introduction to the fundamentals of programming theory, but it is lengthier and will take more time to get to practical projects.
If you really want to get into an introduction to Computer Science theory, it would be hard to find a better introduction than A Data-Centric
Introduction to Computing which is used at Brown University as well as others in their introductory Computer Science courses. It's can also be read online for free.
There was a merge into the main branch of the source material to eliminate references to JuliaBox as that service is no longer available. However, that change didn't make its way to the version published online (and new print versions haven't been published either). I'll reach out to Ben Lauwens about this.
This is huge. Unfortunately, as you indicated, there's no standard tool for this and new ones are being added to the mix. Many in the science feilds are pushed towards Conda but I'm not sure it's the best option. However, Conda will be infinitely better than not using anything to manage environments and dependencies.
Thank you for highlighting this research! At first glance it's interesting that sigmoid functions re-emerge as more useful using the approaches evaluated in that article.