Y
HN Search
Hacker News Search
new
|
comments
|
top
|
jobs
fwlr
searching PlanetScale…
1.
▲
2.
▲
3.
▲
4.
▲
5.
▲
6.
▲
42 ms
·
451.
▲
by
fwlr
3y ago
Yeah, I just threw in a few things I happened to know the specifics of from working with them in a factory. So it was just the intersection of "materials you might be familiar with" and "materials I know the specific grade fo
452.
▲
by
fwlr
3y ago
If it was possible to point to where the code is actually stored in Copilot (i.e. run strings on the server and it spits out the copylefted code), I would be a lot more sympathetic to the view that LLMs are stealing code. Even if you decode
453.
▲
by
fwlr
3y ago
It might be a bit difficult to navigate this site if you don’t know the intricate details of material sub-types - you’ll probably recognize the top-level categories and then get stumped by the subcategories (hybrid austenitic-ferritic wroug
454.
▲
by
fwlr
3y ago
While the accuracy is impressive (the article here says altitude difference of a few millimeters when the tunnels met, but comments have pointed out that’s an error and it is actually a few decimeters), the really cool part to me is the tec
455.
▲
by
fwlr
3y ago
There’s a related story, unfortunately not quite as neatly packaged, and it’s about taxi drivers. A businessman gets into a taxi and starts talking to the taxi driver about (of course) business. The taxi driver is in a great mood and explai
456.
▲
by
fwlr
3y ago
“Languages change over time” is often deployed as an argument against pedantry. I believe pedantry is a useful force (akin to friction) and it plays a necessary role in the change of language over time. A language change has proven its wort
457.
▲
by
fwlr
3y ago
I meant that it is limited by the size of the input layer. A 2000x2000 pixel image of a face has to be downscaled, because otherwise you’d need 4 million nodes on the input layer and that’s impractical.
458.
▲
by
fwlr
3y ago
If I’ve given the impression that I am defending the technology against anything other than claims about its accuracy, I apologize. The fact that I used to work on this technology and now I instead talk about how it harms minorities should
459.
▲
by
fwlr
3y ago
Respectfully, I do have a much better idea than most people about what is being deployed and used, as I built some of them. I am not saying it is bulletproof. I am saying it will reliably pull 10 photos of the same person out of 100,000,000
460.
▲
by
fwlr
3y ago
I have considered that, empirically speaking I’ve seen a few activists do exactly that, and I am definitely not going around believing that activists are all some kind of anti-all-surveillance, defund-the-cops, legalize-crime parody. But al
461.
▲
by
fwlr
3y ago
No, I do agree that Nest’s face rec (and face detection too, for that matter) is pretty awful. I don’t have first-hand knowledge of what Google is actually deploying but I suspect they’re performing so poorly because they’re training the re
462.
▲
by
fwlr
3y ago
I won’t say it’s impossible, but your suggestions do seem unlikely to me. We’re talking about unlabeled data, cropped to just the face, and the state of the art being found in commercial models in the real world rather than in papers compet
463.
▲
by
fwlr
3y ago
Addressing this misconception is the whole point of my original comment, actually. Typically the algorithms used work great on all populations and do not have heavily baked in bias (in the technical sense; they do still have bias in the soc
464.
▲
by
fwlr
3y ago
My usage of “state of the art” was misleading, I apologize. To be more precise I should say “most of the companies offering services in this area are this good and getting better”. This is not like batteries or carbon nanotubes where resear
465.
▲
by
fwlr
3y ago
Facial recognition is an algorithm applied to ordinary picture and video data, it does not require special or even particularly high-quality cameras. In fact, if the footage is too high-quality it actually has to be downscaled before being
466.
▲
by
fwlr
3y ago
With the 24/7 surveillance camera setup, the cost to the landlord of gathering enough violations to evict minority tenants is “watch all recorded video from all cameras, looking for the target”. Facial recognition allows landlords to s
467.
▲
by
fwlr
3y ago
If that was the criteria I would also be comfortable! However, it is not; the FTC has taken action* against many entities smaller than Facebook, and in point of fact Facebook might actually be one of the biggest entities the FTC has taken
468.
▲
by
fwlr
3y ago
I mean, I endorse your framing and don’t think it disagrees with my framing. I thought my comment contained a lot of “good cyborg, nice cyborg” for punishing a known bad actor while also cautioning against thinking government is “all friend
469.
▲
by
fwlr
3y ago
A minor technical nitpick: the article, and the activists it quotes, make the claim that facial recognition discriminates against minorities because it is inaccurate. The state of the art in facial recognition technology has been superhuman
470.
▲
by
fwlr
3y ago
I would caution against believing “the enemy of my enemy is my friend” when that would make the government your friend. Yes, Facebook has been awful and done great harm in some ways, but crucially it is not the case that Facebook was only
471.
▲
by
fwlr
3y ago
> while the vaccine may continue to confer long term benefits It had better! Otherwise (in my opinion) it has little claim to the label “vaccine”.
472.
▲
by
fwlr
3y ago
> although they tend to still disproportionally affect lower-income individual Any fair punishment will disproportionally impact lower-income individuals, since resources confer resilience and resilience resists punishment. If your goal
473.
▲
by
fwlr
3y ago
Imagine a government with eleven citizens, ten of which earn a thousand dollars a year and pay 50% income tax on each dollar, while the eleventh earns one million dollars a year and pays just 10% income tax on each dollar. If the government
474.
▲
by
fwlr
3y ago
If you dispense the same punishment to two individuals and one is less able to withstand that punishment, are you punishing that person more? I’m not sure that the answer is a simple “yes”. For one, this would mean if you gave someone a 5
475.
▲
by
fwlr
3y ago
> iCloud is objectively the best cloud service in your opinion? You're really not just picking that because you're using Apple products and thats the only one that allows backing up an iOS device (funny that) for example? Had m
476.
▲
by
fwlr
3y ago
There are good uses for app-specific back buttons, for sure! Two cases I’ve seen - one is a site with “single page app”ish flow for reading sequential posts, where rather than reloading the entire site it fetches just the JSON data for next
477.
▲
by
fwlr
3y ago
As bills like this move forward, the specifics begin to be debated and trialed. Can landlords use face rec? Can stadiums? What about banks, who have perhaps a more pressing ‘security concern’ argument? What about corner stores, supermarkets
478.
▲
by
fwlr
3y ago
Web pages have the ability to fire a “back event” that (I am pretty sure) works identically to clicking the back button on your browser. I imagine they’re using this because it’s very simple, you just put onClick(History.back) on the b
479.
▲
by
fwlr
3y ago
Highly suggest sorting by date rather than size. Putting the smallest failures at the top causes a little bit of an optical illusion making the first tower appear smaller than it really is. It also highlights that whatever is happening here
480.
▲
by
fwlr
3y ago
I’ve said many times that facial recognition inherits the moral issues of policing. Maybe I should generalize. Technology inherits the morality of whatever it empowers.
More ›