Hacker News Reader

Comment by btbuildem | Hacker News Reader

Given how many accounts online are bots or anonymous, will they ever be able to shake out which real people which personas belong to? Or will it be a case of "we said this is you, so that makes it you"?

View on HN

The causation chain here is "we don't like someone so we find dirt on him online and arrest them over thought crime"

The latter. Checking for identity won't matter to ICE. They already break into apartment buildings without warrants and arrest children and US citizens. There's 0 logical reason to believe ICE will do anything in good faith considering their past and current behavior as well as by the words of Trump himself.

https://chicago.suntimes.com/immigration/2025/10/01/massive-...

Yes, they’ll use bots to detect the bots. I guarantee it’s not hard to detect outliers (bots) via many different heuristics at this point. It’s going to turn into a perpetual game of whack a mole just like SEO and ad blocking.

I'm going to modify my keyboard to generate em-dash when I input "-"...

That works well until the following doesn't compile

int x = 1 — 2;

  #define — -

fixes that issue for C.

Option+shift+dash, on a default English Mac keyboard. Easy to remember—"modify [option] the dash [dash] to make the biggest common form of it [shift]", or else you can think of it as modifying the underscore (shift + dash) to sit higher on the line (with option).

There are already attribution technologies that are reliable and definitive. There's no real technical reason that they wouldn't have the capability to directly attribute accounts.

Now..

is any of that gonna stop John or Joan Q Public who has a running disagreement with their asian or hispanic neighbor from calling in a "tip"? Probably not.

Will that tip result in an automated digital proctology exam that may land them, or people they may have been communicating with (like grandma), in ICE custody?

I'm thinking probably so.

The danger here is not really technology. The danger here is people. They are essentially automating the "snitch-to-detainment" pipeline.

> There are already attribution technologies that are reliable and definitive.

Gonna stop you right there and ask you to back that up. It's not plausible. Especially not if people are actively trying to mess up the "attribution".

E.g., https://news.ycombinator.com/item?id=33755016 from 2022. Only N-gram frequency analysis, no modern ML techniques needed.

Probably time for somebody to revisit this topic with an actual model.

That's not even close to "reliable and definitive". I'll give you "suggestive". And it presumably has certain volume requirements.

What on Earth makes you think that the people who will be using this kind of tech will care how "reliable" or how "definitive" it is? False positives aren't a bug with ICE, they're a perk of the job!

Not to put too fine a point on it, but do you consider handwriting comparisons "reliable and definitive"?

Because juries regularly convict people in courts based on such evidence.

Maybe a better question is, what is the level of "reliable and definitive" that evidence would need to reach for you to believe it should be admissible in court?

I suspect that maybe that's the misunderstanding. The level for you is much higher than maybe it is in the justice system. Which is fine. But that doesn't change the justice system. In court, the currently accepted levels are still enough.

People will be free to try to refute the attributions during the proceedings if they think it will do any good. I mean the adversarial system is also part of the justice system.

Reply on news.ycombinator.com