I would argue it was both. No doubt this company was marketing it in a way to make it seem very reliable. And all of the procedural things afterwards made the error so much more damaging.
But imo this is why local police departments should not have access to this kind of tool. It is too powerful, and the statistical interpretation is too complicated for random North Dakota cops to use responsibly. Neither the company nor the PD have an incentive to be careful.
It's not an AI error. The face recognition AI simply said that it's a "potential match", which is correct. It's the humans' job to confirm that a potential match is in fact a match, especially when the suspect is 1,900 kms away.
I don’t know what tool they used, but it was very likely not an LLM. They probably have some database of drivers’ licenses and they ran a similarity search against the surveillance footage. This poor lady happened to be the top match.
Even if it also output a score, that score depends on how the model was trained. And the cops might ignore it anyways.
LLMs have significantly reduced the time I’ve spent chasing down cryptic errors on stack overflow, old github issues, or asking in random slack channels about it. Even if that’s all they did, they would be very valuable.
If that means I’m actually coding instead of figuring out why xyz random plugin isn’t doing its job right now, some subsystem that I need but don’t care to learn the internals of, then I am happy.
I am here to express my ideas and opinions. They might not always be popular, but they are my opinions (that is reason that I have 3x less karma than you but I was here 11 years longer). And some people will debate my opinions and try to convince me that I am wrong. And sometimes I learn soemthing.
But if we start ignoring ideas and opinions and instead focus on superficial things like how they are written or communicated, then the whole point of HN is lost.
If that is true you shouldn't have any objection to a rule against letting a chatbot express your ideas and options for you. Express yourself, because asking a chatbot to do your thinking and writing for you is not a superficial thing.
> But if we start ignoring ideas and opinions and instead focus on superficial things like how they are written or communicated, then the whole point of HN is lost.
How a message is communicated matters and always has. Even before this rule, I could express opinions here in ways that would get me banned from this website, and I could express those exact same opinions in ways that would not. Ideas and opinions still matter, but so does how we communicate them. It's a very small ask that you express your own thoughts in your own words while participating here.
The concept is the same: directed radio waves. They can be tuned and directed far better than a magnetron and waveguide pulled from a microwave. This is an adversary more sophisticated than some rando disassembling appliances in their garage, and even a moderately knowledgeable radio technician could improve upon it.
For it to do enough damage to make the inner ear fail temporarily and give you brain damage without burning anything seems a bit unbelievable though. Not even a first degree burn? I think even something as unlikely as remote-activated poison is a better explanation.
Different frequencies interact differently with wall materials or window materials or hydrous, organic matter. For example, many substances are pervious to microwave radiation and will remain cool while an egg beneath them cooks.
This is true for your skin in some cases, too: the frequencies which will quickly heat skin are a subset of the frequencies which can hurt you at high energy. For example, X-rays will do the latter a lot faster.
And that's only the considerations if you treat the skin and brain as just a hunk of stuff versus a ridiculously complex meat-computer. For an example, see the "microwave auditory effect"[0]. This effect originates inside the brain, not the ear. High levels of electromagnetic energy tends to harm even electronic computers, and they have no meat to deal with!
That's actually something I didn't think about! So the EM wave don't interact with the liquid, but the brain electric signal. I don't know if it is possible, but it is a far, far more likely than something that fry the inner ear/brain.
Ok, I'd rank it ahead of the poison theory, still behind the mass psychosis because of Hanlon though.
> But in another instance, Epstein was critical of misspelling. A contact forwarded the sex offender his daughter’s college application in 2013. “I wish you had let me review before sending…the grammatical errors and spelling mistakes will make it at least harder for early admission,” Epstein wrote.
It is funny that spelling and grammar matter more when writing to an admissions officer than to a potential business partner. But it’s also funny to imagine a world where you could send in an essay with a bunch of typos and grammar mistakes and expect it not to influence your application.
Spelling and grammar matter in the sense that they are a signal that you know a complicated and somewhat arbitrary set of rules and have agreed to follow them.
Proper grammar on formulaic language is a proof-of-work system. Difficult to achieve but easy to check. It suggests that the author cared enough to put in the time. When the cost of graduate labor is low, careful editing suggests that you can burn a student's time to demonstrate the message is worth reading.
Also just differing levels of relevance. You don't talk with a businessman or investor or famous people in general because of their writing; if you made a list of relevant skills, 'proper spelling when quickly texting from a phone' surely doesn't crack even the top ten thousand skills. In academia, on the other hand, writing a formal application properly is a core skill.
I would have to consider carefully if I thought I was a high-enough quality candidate that it would be interpreted as a countersignal rather than a signal.
If I, gwern, specifically, were to apply, I might; because I know I am widely read on HN and I've talked with any number of YC partners etc, and they all know I take care in writing, and so me not capitalizing is a deliberate message rather than laziness or incompetence. They may or may not appreciate the message, but they won't infer the usual things, at least.
If I were anyone else and my application just one of thousands in the flood? You'd better believe I'd capitalize and spellcheck my YC application: https://gwern.net/blog/2023/good-writing
The difference is the formality of the setting, not necessarily the recipient. Sending a quick note via email or text is one, and a pitch deck is another.
But imo this is why local police departments should not have access to this kind of tool. It is too powerful, and the statistical interpretation is too complicated for random North Dakota cops to use responsibly. Neither the company nor the PD have an incentive to be careful.
reply