tehnub

27 ms·

361.
by tehnub 5y ago
Well if you go high-enough resolution, then you can scale the display and it will be big enough to read, but also much sharper. You want integer scaling though, otherwise you get unclean lines, weird "rounding" errors, etc. On my
362.
by tehnub 5y ago
It's the return type that isn't decidable. You can't statically check, in the arbitrary case, whether the function returns an int or a string.
363.
by tehnub 5y ago
That's a good point. If they're never checked, then they're just like incorrect/outdated comments. They sort of get at this idea in the article, and sort of describe a compromise for it. They have a list of files that th
364.
by tehnub 5y ago
He just did derive it from the halting problem. In words: It's impossible to statically check the type of a function that takes a program P as its input, and returns the integer 1 as its output if P halts, and returns the string "
365.
by tehnub 5y ago
I'm not sure if 4K is enough to be Retina on a 27 inch at normal viewing distances. Apple made their 24 inch iMac 4.5K, and they call that Retina. And also their 27 inch collab with LG is a 5K screen.
366.
by tehnub 5y ago
Same wish here. Last I tried a few months ago, I was unable to compile UE4 as well. These machines would be great for UE4 dev if only the compatibility was there. I wonder if the politics between Epic and Apple has made efforts in this area
367.
by tehnub 5y ago
This looks like a good codebase to learn modern Python style and features from. I found it quite readable despite not being very familiar with this sort of programming.
368.
by tehnub 5y ago
Post from a little while back where someone really really thought this through. IIRC, they concluded that there's no good way to ask it, lol. https://news.ycombinator.com/item?id=21675717
369.
by tehnub 5y ago
I saw a Linus tech tips video where they made that mod, and they found that it heated the chassis well beyond the regulation limits, while Apple’s original design falls just within it.
370.
by tehnub 5y ago
Sincere question: Can mirrorless cameras match or do better than the result of the multiple exposure stuff that phones can do? (E.g. Apple’s “Deep Fusion” feature)
371.
by tehnub 5y ago
I would definitely encourage you to try digital libraries. I’m in the LA area and have access to both the LA county and LA city libraries, and between those two I have access to the vast majority of books I’m interested in. It’s easy to con
372.
by tehnub 5y ago
Isn’t it only for a few weeks or something?
373.
by tehnub 5y ago
But did the birthday paradox inform his procedure in any way? I guess he wouldn't have even attempted to do this at all if he didn't think it was likely to find collisions.
374.
by tehnub 5y ago
Apple did mention in their security thread model document [0] this: Apple will also refuse all requests to instruct human reviewers to file reports for anything other than CSAM materials for accounts that exceed the match thresh
375.
by tehnub 5y ago
> In order to test things, I decided to search the publicly available ImageNet dataset for collisions between semantically different images. I generated NeuralHashes for all 1.43 million images and searched for organic collisions. By tak
376.
by tehnub 5y ago
When my company first started using Zoom, people were fumbling the controls constantly. In fact, people still do ("Can you see my screen?"), and no one knows what "Optimize for video clip" even means. I don't think
377.
by tehnub 5y ago
Their use of the phrase "This claim is subject to code inspection by security researchers like all other iOS device-side security claims" stood out to me. Could someone tell me how that inspection works? Are there researchers who
378.
by tehnub 5y ago
They put out a new paper [0] describing the security thread model they were working with, and this paragraph on page 9 stood out to me: The perceptual CSAM hash database is included, in an encrypted form, as part of the signed operating sys
379.
by tehnub 5y ago
The neural net isn't exactly searching for CSAM itself. Its role is to extract perceptual features from the image, and it is applied to both the CSAM images and your iCloud images. If those were the same to start with, then the extract
380.
by tehnub 5y ago
To get the probability of a single false positive match, you need to look at the CDF of the binomial distribution [0]. Let X be the number of matches in your iCloud library. Assuming each photo's probability of a match is independent o
381.
by tehnub 5y ago
There will only be an alert if that photo is extremely similar to an image in the NCMEC database, AND there are numerous other such photos on the account that match. The threshold number of matches to trigger an alert is tuned for a 1/
382.
by tehnub 5y ago
It's true that Apple never _sees_ a hash that doesn't match, but the encrypted hash is included in the safety voucher. That is to say, all hashes are uploaded, but only the matches can ever be decrypted, and that's only if th
383.
by tehnub 5y ago
I always recommend the “Field Guide to Understanding Human Error” [0], which discusses a lot of similar ideas. Really helped refine my thinking about systems and errors. [0]: https://www.oreilly.com/library/view/th
384.
by tehnub 5y ago
All are uploaded but only when the secret sharing threshold is met are any of them decryptable, and then it’s only the ones that were a match that can be decrypted.
385.
by tehnub 5y ago
In the manual review failure case, you're saying that the reviewer would decrypt the steganographical encryption, discover the hidden offensive content, and thereby decide that it's a true positive? And related to that, I wonder h
386.
by tehnub 5y ago
They say that they address this issue through a mechanism outside of the cryptographic protocol, but don't say specifically how. The quote from the paper: A user might store multiple variants or near-duplicates of the same image on the
387.
by tehnub 5y ago
Presumably the reviewers would not know the identity of the user whose photos are under review, as they have no need to.
388.
by tehnub 5y ago
Maybe the carpet acts as a filter for dust that you would otherwise be breathing?
389.
by tehnub 5y ago
I was wondering about similar things a while back. I don't know if there are any easy options, but if you were, say, looking to do a custom house build, it seems that a passive house with mechanical ventilation heat recovery (MVHR), or
390.
by tehnub 5y ago
Anyone else find the crescendos on "A Day in the Life" to sound very different on the 2017 remix compared to older mixes? Not just more/better stereo mixing, but it almost sounds like the instruments are playing different not

More ›