[drewbug01]

> The chances of a citizen being targeted by ICE is low.

You can’t start with this premise, though. Recent rulings allow stops based on “probable cause” such as a combination of “speaks Spanish”, “is brown”, and “is in a place where we think illegal immigrants might be”.

So like: any Latino US citizen, who happens to be working someplace like a landscaping company. Or a kitchen.

The idea that citizens aren’t likely to be targets is now laughable. And we have ample reporting indicating that in fact, citizens are being detained, for hours and hours (if not longer).

[somenameforme]

Low doesn't mean zero, it means low. You might notice I used different terms for the different groupings, with the chance of a citizen being targeted by ICE as the highest overall at "low". ICE has so far deported more than 400,000 illegal aliens. [1] If they were "only" 99% accurate, you'd be able to find thousands of instances where things went wrong. Instead, you're looking more at tens to low hundreds of instances, so it's likely that their overall accuracy is somewhere in the 99.9% to 99.99% range.

And as I was demonstrating above, the conditional probabilities required for a false positive from this app mean that it's practical effective accuracy rate will likely be 100%.

[1] - https://www.dhs.gov/news/2025/09/23/new-milestone-over-2-mil...

[drewbug01]

> Instead, you're looking more at tens to low hundreds of instances

…based on what, the independent research ProPublica did? DHS doesn’t even keep statistics on how many citizens they detain so I’m not sure we should be assuming the numbers here are that low.