MSnap just settled a lawsuit over whether Snapchat was designed to be addictive — days before opening statements.
CSettled. So we never get to see the evidence.
MTwo teenagers sued, saying the app contributed to their mental health struggles. Snap settled during jury selection in Los Angeles.
CAnd here's the part that gets me: those same plaintiffs already settled with Meta, TikTok, and YouTube.
MRight. So this was the last one standing. And Snap folded too.
CI want to see one of these actually go to trial. Just one.
MWhy? The companies clearly don't want a jury deciding what 'addictive by design' means.
CExactly! That's the whole problem. We keep settling, so we never get a legal definition. We never get precedent.
MBut think about what a trial would actually look like. You'd have engineers explaining why streaks exist, why notifications fire at certain times.
CGood! Let them explain it under oath. If it's not addictive, prove it.
MYou know they'd argue it's just good product design. Retention isn't addiction.
CAnd that's a real argument! But we never hear it because everyone settles. The public learns nothing.
MHere's my problem with the lawsuit framing. Two teenagers had mental health struggles. That's real. But proving Snapchat caused it? That's a huge leap.
CIt's not about causation for one kid. It's about whether the design itself is predatory.
MPredatory? Snapchat? The app where messages disappear?
CStreaks, Matilda. The streak feature. Some kids say they feel physically sick if they break a 500-day streak. That's not an accident.
MOkay, streaks are manipulative. I'll give you that. But settling doesn't mean Snap admitted anything.
CNo, it means they paid to make it go away. And the next lawsuit starts from zero.
MSo what's your fix? Ban streaks?
CI'd start with real transparency. Show users exactly how the app is engineered to keep them coming back.
MThat's not a bad idea. But you know what happens? People read it, shrug, and keep scrolling.
CMaybe. But at least we'd stop pretending these apps are neutral tools.
MFair. And look, the bigger story here is that the whole industry is paying to avoid the courtroom.
CWhich tells you they're all afraid of the same thing: a jury actually looking at the code.
MOr they're just doing the math. Settle for a few million, avoid a decade of discovery.
CThat's the cynical read. And it's probably right.
MSpeaking of things that never go to trial — there's a new open-source tool on Hacker News that's actually kind of clever.
CWhat is it?
MIt's called Gauge. A command-line tool that recursively improves your documentation for coding agents.
CRecursively improves docs? How does that work?
MThe idea is you set up tasks that a coding agent would need your docs for — like setting up CI checks. The tool runs the tasks, sees where the docs fail, and fixes them.
CSo it's like a self-healing README.
MBasically. And it runs locally, so you don't need an account to try it.
CI like that. No cloud dependency, no signup wall. Just a CLI that does one thing.
MThe top comment on Hacker News is someone correcting the author's word choice. They said 'recursively' but the commenter says it should be 'iteratively.'
COf course that's the top comment. It's Hacker News.
MThe author's README even calls it iteration. So the pedant has a point.
CI'll take a useful tool with a slightly wrong adverb over another crypto signal any day.
MDon't get me started. Half the posts in the feed right now are token shills and engagement bait.
CWhich is why something like Gauge stands out. It's just a person building a thing and sharing it.
MAnd the local execution backend is smart. You're not sending your codebase to some third-party service.
CFor teams that care about that, it's a real selling point.
MThe bigger trend here is that documentation is becoming a first-class problem again, because AI coding agents are only as good as the docs they read.
CGarbage docs in, garbage code out. We've been saying that about humans for years.
MNow the agents are the ones complaining. And tools like this are the answer.
CI'd try it. Worst case, my README gets slightly better and a Hacker News pedant corrects my grammar.
MThat's the spirit.