More than the tool it’s the principles outlined in this article that are interesting if you think about what are the parts/systems/rules that would make a harness successful.

“[People] simply drop it into where it needs to go (or not), and somehow, magically, it seems to work. And I am not talking about not remembering flags for Linux cli tooling (tar, looking at you), I am talking about people copy pasting entire piped xargs commands into their terminal and then praying that it doesn’t ruin your day somehow. No thinking, no learning, just slop everywhere. Yes, people did this before AI using Stackoverflow snippets, but at least you had to fit the puzzle piece into your architecture, and think about the code in some way or form. Not anymore.”

A contrarian opinion piece that draws parallels between some Mao’s campaigns and the rush to AI. Lots of great views that give perspective and help keep things in check.

An account of using AI to do a project that was just a bit out of reach previously. There’s some really good insights, like comparing AI coding to slot machines.

An in-depth article about working with LLMs from a senior engineering perspective.

Really good quote material in this article that introduces the notion of cognitive debt induced by AI coding without time to ingest the architecture and edge cases.

• https://antropia.studio/blog/to-ai-or-not-to-ai/

Insightful notes on the use of AI in coding, in a nutshell, own the code and look for (high gradient) opportunities.

Why is it that we craft our work so AI can understand but we do not do so for people?

It’s like we have an expectation people should put forth the effort to understand, but AI has hard limitations, so we have to take extra steps for AI but can be lazy for others.

But what they cannot do is maintain clear mental models.

LLMs get endlessly confused: they assume the code they wrote actually works; when test fail, they are left guessing as to whether to fix the code or the tests; and when it gets frustrating, they just delete the whole lot and start over.

This is exactly the opposite of what I am looking for.

I believe that language models aren’t world models. It’s a weak claim — I’m not saying they’re useless, or that we’re done milking them. It’s also a fuzzy-sounding claim — with its trillion weights, who can prove that there’s something an LLM isn’t a model of?

A great write-up on the “My AI Skeptic Friends Are All Nuts” article, with some really solid thinking on the linguistic & logical shenanigans in AI hype talk.

On the very likely outcomes of the impacts LLMs are having on software engineers and their craft.

Arguments against the use LLMs for homework.