HNHacker News
TopNewBestAskShowJobs

ThinBold

54 karma · joined August 4, 2021

submissionscomments
ThinBold··on The Paywalled Garden: iOS is Adware (2020)
I find it difficult to position myself on this topic. On one hand, apps like stock can be deleted so you no longer see any ads/add-ons. It's an O(1) effort. On the other hand, the entire appstore is just a big junk of ads, and companies nowadays keep asking you to install a new app for, e.g., renting an e-scooter. Now that's O(1) effort per microservice IRL. I ended up aggressively deleting idle apps as a revenge and an attempt to keep my life simple.
ThinBold··on The Hidden Costs of Living Alone
I can happily live with my parents and save a fortune (assuming the logic in the article is correct), but the mating market wouldn't agree with we. The only way I can compensate living alone is to work extra harder to accumulate more professional currency; even if it means that I am so-called stealing jobs from people with families.
ThinBold··on A Cool Customer
Basically every party that is contributing to the global warming can defend like this.
ThinBold··on Apple can read your iMessages (even though they’re E2E encrypted)
Doesn't Apple simply happen to be both the chat provider and backup provider, so Apple-A does the E2E encryption and Apple-B sees your backup because you sort of want that?

And people worrying about the other end of the chat... come on, you talked to them in the first place. They can forward anything, even if it's via Signal.

The entire story is just hilarious and memeable. Users want backup; Apple open up the gate. Users want E2E; Apple shut up the gate. Users want iCloud recovery; Apple partially open the gate.

ThinBold··on Show HN: Neural-hash-collider – Find target hash collisions for NeuralHash
That NSFW picture is, in honor of Sean Lock, a challenging wank
ThinBold··on Show HN: Neural-hash-collider – Find target hash collisions for NeuralHash

  save hundreds of lives, would you?
Sure, why not.

But let me ask some questions, because at this point I am not sure if people want Apple's system to be robust or jammable. If our fear is that Apple will tune the system to detect pictures of two men kissing, wouldn't an easily jammable system works in our favor because we can DDoS it or threaten to do so anytime we want?

ThinBold··on Show HN: Neural-hash-collider – Find target hash collisions for NeuralHash
Unless Apple implements this with a backdoor on iCloud, the worst case scenario here is that they receive millions of false positives per day and terminate the program after two weeks.

The scalability issue seems to work in our favor because, perhaps, the normal usage will overwhelm the human reviewers Apple prepared and we don't even need to send troll images.

And for the entire time, our data remains untouched.

ThinBold··on Show HN: Neural-hash-collider – Find target hash collisions for NeuralHash
I did my best skimming comments the algorithm showed me, including related posts' comments. But man am I bad at reading comprehension.

If I miss anything then surely I am willing to be corrected. But so far I don't see comments that show us how to penetrate the four-layer system (local hash check, semantic check by user, on-server hash check, and human reviewer).

ThinBold··on Show HN: Neural-hash-collider – Find target hash collisions for NeuralHash
The first comment algorithm choses to show me contains substring "and this system is transparently broken". Isn't that a remark on the effectiveness of the idea or I misread?

(In case you have problem with me saying "so many", that I can fix.)

ThinBold··on Show HN: Neural-hash-collider – Find target hash collisions for NeuralHash
Despite that Apple scanning our images is a horrible privacy practice, I don't get why 𝚜̶𝚘̶ ̶𝚖̶𝚊̶𝚗̶𝚢̶ some people think this is an ineffective idea.

Surely you can easily fabricate innocent images whose NeuralHash matches the database. But in what way are you going to send them to victims and convince them to save them to their photo library? The moment you send it via WhatsApp FB will stop you because (they think) it is a problematic image. And Even if the image did land, it has to look like some cats and dogs or the receiver will just ignore. (Even worse, the receiver may report you.) And even if your image does look like cats and dogs, it has to pass another automatic test at the server side that uses another obfuscated, constantly-updating algorithm. After that, even more tests if Apple really wants to.

That means your image needs to collide ≥ three times, one open, one obfuscated, and one Turing.

Gmail scans your attachments and most people are cool with it. I highly doubt that Apple has any reason to withdraw this.

ThinBold··on Scientists find an effective solution for the three-body problem
While the tone sounds like a university webpage advertising its scholars' results, this does seem to be an interesting viewpoint. (Unless similar viewpoints have been proposed before.)

This also reminds me of the QR iteration, where you loose track of the matrix entries very quickly (after 2 or 3 steps into the iteration), but in the end the diagonal does converge to the eigenvalues.

ThinBold··on Good Design is Imperfect Design, Part 1: Honest Names
That

    (some date + 1 month) + 1 month ≠ some date + 2 months
has always been true for floating numbers. This is exactly why you never trust fast matrix multiplications---they rely on cancelations of the form (a + b) - b = a + (b - b) = a.

I would argue that 1 month is just having too few significant digits. You can even implement it as returning 30 with probability 60% and returning 31 with probability 40%. Then on average you would have

    1 year ≈ sum([1 month] * 12)
← PreviousPage 2 of 2