As someone who regularly has to deal with code that has been broken needlessly into smaller functions so that I have to constantly jump around to figure out what is going on, this really resonates with me.

The latest case was someone who took something that really only needed to be a single function and instead turned it into a class with a dozen tiny methods.

I was responding to the following paragraph in the article:

We used to get proof-of-thought for free because producing a patch took real effort. Now that writing code is cheap, verification becomes the real proof-of-work. I mean proof of work in the original sense: effort that leaves a trail: careful reviews, assumption checks, simulations, stress tests, design notes, postmortems. That trail is hard to fake. [emphasis mine] In a world where AI says anything with confidence and the tone never changes, skepticism becomes the scarce resource.

I am a bit wary that the trail of verification will continue to be so “hard to fake”.

Unfortunately, the next step is for AI to get better and better at generating verification trails that look correct but are not.

Yeah, the transition from the third frame to the fourth frame just doesn’t make sense as currently depicted; why would people playing video games suddenly start to revolt?

However, if the third frame were to depict rich men getting richer as you suggest, then the revolt in the fourth frame would make more sense.

I’ve been thinking about trying it, but it wasn’t obviously “magical” to me when I last took a look. What makes you say that?

What makes it that bad?

Hey, now, just because I am an overly paranoid person does not mean that you have to be as well!

Thanks, your comment is an antidote to my paranoia that it is impossible to do anything to address all threats. 😀

Given that your advice is very sound, I have a question: would I gain much by using OpenBSD? The conventional wisdom when I last checked is that it is the most secure unix-like operating system on the planet.

I admit nothing.

Everything that you mention is sensible, but it seems like it would take so much time not only to set up but to perform the ongoing maintenance you described that it just is not worth the trouble to self-host, which is a significant factor in why I have not taken a shot at it.

The researchers in the academic field of machine learning who came up with LLMs are certainly aware of their limitations and are exploring other possibilities, but unfortunately what happened in industry is that people noticed that one particular approach was good enough to look impressive and then everyone jumped on that bandwagon.