When I was a kid some family friends used WD40 on their joints - arthritic knees and such. Church friends, actually, which I mention only because stuff like that probably helped me reject the religion as readily as I did.
A web search for "WD40 arthritis" shows that there are still people doing this.
There isn't one. I guess the actual reason I mentioned it was I felt weird calling them family friends when they weren't, but I also felt weird calling them church friends for no stated reason, so I added a little personal anecdote about it, which now I think I shouldn't have.
But since I did, lemme clarify, it was a pretty out-there fundamentalist church that I'm glad to have escaped early, and my comment is just that seeing people there do stuff that I couldn't make sense of, even totally unrelated stuff like this, probably helped undermine any sense of authority they had in my mind.
I agree but find it's fairly easy noise to ignore.
I wouldn't replace human review with LLM-review but it is a good complement that can be run less frequently than human review.
Maybe that's why I find it easy to ignore the noise, I have it to a huge review task after a lot of changes have happened. It'll find 10 or so things, and the top 3 or 4 are likely good ones to look deeper into.
Using an LLM to code isn't the same as vibe coding. Vibe coding, as originally coined, is not caring at all about the code or looking at the code. It was coined specifically to differentiate it from the type of AI-assisted coding you're talking about.
It's used more broadly now, but still to refer to the opposite end of the spectrum of AI-assisted coding to what you described.
I was vibe coding in November 2024, before the term was coined. I think that is about as early as anyone was doing it, so 1.25 years ago. Cursor added its "agentic" mode around then, I think, but before that there was just "accept all" without looking at changes repeatedly.
I shipped a small game that way (https://love-15.com/) -- one that I've wished to make for a long time but wouldn't have been worth building other wise. It's tiny, really, but very niche -- despite being tiny, I hit brick walls multiple times vibing it, and had to take a few brief breaks from vibing to get it unstuck.
Claude Code was a step change after that, along with model upgrades, about 9 months ago. That size project has been doable as a vibe coded project since then without hitting brick walls.
All this to say I really doubt most claims about having been vibe coding for more than 9-15 months.
Because it's not? I use these things very extensively to great effect, and the idea that you'd think of it as "smart" is alien to me, and seems like it would hurt your ability to get much out of them.
Like, they're superhuman at breadth and speed and some other properties, but they don't make good decisions.
For small games I work on I make sure claude (well, codex cli) can produce screenshots of whatever screen it's working on and evaluate them. It has some instructions on using codex exec (claude -p) to use a clean instance for evaluation, so it can pass a screenshot and description of expectation and get a pass/fail and description of the failure. The main agent can also just view the image but for things with a clear pass/fail I prefer it invoke a clean context.
My personal distaste for typical narrative presentations of interesting information is how often the first interesting details come 4-5 paragraphs in and then are slowly peppered from there. Really doesn't seem at odds with the advice here which can easily be applied to the opening sentence or paragraph, and title.
reply