On AI coding. AI certainly cannot replace serious devs, but rather it makes them more productive. But it can be immensely helpful to people who use development, or a particular kind of development sparingly, as just one tool in their toolbox. I am an ERP consultant, most of my work time is not development, but I do write smaller programs, really just customizations, not whole apps, in the AL for the Microsoft Business Central, https://learn.microsoft.com/en-us/dynamics365/business-central/dev-itpro/developer/devenv-programming-in-al which is very, very limited, it cannot really call external libraries and so on. So when I just want a Python script to transform a DBASE DBF file to CSV so that I can import it into BC with AL, Claude or even Bing AI can do the job well. Well enough. I still need to fix bugs, but the effort and learning invested is much smaller. Due  to the simplicity and heavy limitations of AL, AI can also help me develop in AL with much less  typing, in a much shorter time. I simply use the free web based Claude or the AI in the Bing browser. No subscription. 

I think AI is a great help to those for whom coding and rather simple coding is only a part of the job.

It has been noticed long ago, that expertise in a given field plus combining with basic scripting type of programming works well. When they started to develop diagnostic software in the 1980's, they found out programmers do not understand well the specs doctors write, because the doctors forgot to write down those things that seem basic and obvious to them. So a simple "medical BASIC" programming language was made, doctors were taught to use it, and it was very successful. ERP development, like AL, or say SAP ABAP is the same. 

Now AI pushes it further, it really helps the "major in something, minor in programming, automate your major with simple scripts" kind of people like me.

Thoughts?

[BEGINNER] The uses of Bayesianism, do I see it right?

1. Originally, Reverend Bayes was only talking about statistics.
2. Then Yudkowsky took this as the whole basis of rationality and knowledge.
3. More recently evidence was found that the human brain works on Bayesian principles, so Bayesianism is rooted in our biology, it is necessary for our self-understanding, and that made Bayesianism far more solid?

On the third, I Scott wrote about books like Surfing Uncertainty (I have bought it, have not read it yet) https://slatestarcodex.com/2017/09/05/book-review-surfing-uncertainty/ basically Karl Friston stuff, predictive processing, free energy and so on.

Do it seems today, that Bayesianism is not something one chooses to do, but we automatically run in a Bayesian mode, we just often do not do it well?

Thank you! "having another person reflect your situation back to you" sounds exactly like "paid friend", though the reality is that I do not actually have such real life friends so ultimately I might take it up. I could try AI, but so far AI is too much of an "ass-kisser", Claude AI just agrees with me about everything... I hear GPT-5 is "harsher" which might be good. 

One thing that is holding me back is that I am not really a big believer in the, how to put it, the power of thought in such matters. I think more like we are chemical machines, for example I heard stories like for some people a literally ONE time iron supplementation gave them the immediate energy to overcome certain obstacles. For this reason I tried some antidepressants, but the results were very underwhelming, yet, I think it is possible to fix our minds through our bodies, be that exercise, healthier eating, supplements etc.