I guess you aren't kidding with your comments. Mentioning what something is, where that thing isn't that well know is not egregious. Also, don't the people who make SuperDuperer! want people to consider buying it? Telling them it's not a floor wax or desert topping might sell a few more copies.
The site looks like every other Claudesite, the docs are Claudemade, and I imagine Claude did the v4 rewrite. There's already been one bugfix release, probably more to come if the super-smart LLM took shortcuts around arcane MacOS esoterica in the name of "optimization".
HN detectives, hard at work. I wish there was a way to automatically hide all the comments saying "but I think LLM was involved in this software". These days it seems like most of the content here on HN.
This is especially galling when we discuss software that has been around for >20 years and that people (actual users, not HN haters) universally love. See other comments around here.
So, I put the notebook with that design in it on a shelf.
For nearly 20 years.
Today, I’m happy to say that I’ve finally been able to carve out enough time to grab that “old” redesign, flesh it out further, modernize it, and make that dream a reality.
Hmm, what could have allowed the sudden burst of productivity? /s
That said, we shouldn't assume that if he used an LLM we know what quality the finished product is.
For all we know, he used an LLM to get a working rough version, and then rewrote every method, bit by bit. For this sort of product - an decades-old app that can result in data loss and so the wrath of furious users - it is more likely the developer was prudent.
Eh. Sooner or later (likely sooner), every project is going to be at least, informed, if not completely written, by an LLM. There’s just too much upside.
It’s also very likely that most copy will be at least edited, if not completely written, by an LLM. There’s just too much upside. Some people may pass every comment they make, on every venue, through an LLM.
Some folks are likely to write agents to be their entire public persona, but that’s just silly (however, I guarantee it’s already happening -people are silly).
For myself, I’m in the middle. I use an LLM to assist in my development, but without resorting to a full agentic “write me a Facebook” methodology.
I will also often use an LLM to edit copy that I use in my apps, or in things like the App Store page for my apps. In some cases, the edit may be a full rewrite, and I’ll choose the rewrite, over my original copy. Usually, I will just use suggested minor edits, and grammar corrections.
I’m not [yet] confident in using LLMs for main translation, but they have been useful for very small “patch” translations.
Overall, LLMs have been a huge help, and I’ve been coding for over 40 years. I’ve been writing, for much longer. I’m not going to use a quill pen, just because it’s fashionable. I’m retired, and only have so much time left, to get a lot done. I appreciate these new tools.
If the bar is that all code and copy must be hand-written, artisanal prose, then it’s unlikely most modern, production-oriented venues will be satisfactory. I assume that new venues will appear, catering to that demographic. I doubt HN will be that.
Then the article isn't for you. That's okay!
Not every piece of writing has to be delicately crafted to the sensibilities of The Hacker News Reader.
Just because someone parachutes it onto the frontpage one day doesn't change that.
But if you're still curious and you can't find the easily reachable homepage then that's on you.