Hacker News: adgjlsfhk1

New comment by adgjlsfhk1 in "Woman charged with computer fraud due to scam CashApp account being in her name"

adgjlsfhk1 — Fri, 23 Oct 2026 05:34:51 +0000

It's not (convictable) perjury if it's what the government tells you to say.

New comment by adgjlsfhk1 in "Woman charged with computer fraud due to scam CashApp account being in her name"

adgjlsfhk1 — Thu, 22 Oct 2026 23:27:01 +0000

Extortion for testimony doesn't tend to produce reliable testimony.

New comment by adgjlsfhk1 in "The recently proposed o(n log n) integer multiplication algorithm [pdf]"

adgjlsfhk1 — Thu, 22 Oct 2026 19:38:26 +0000

You're dramatically underestimating. You need something like 2^2^2^5000 digits or something crazy like that.

New comment by adgjlsfhk1 in "The recently proposed o(n log n) integer multiplication algorithm [pdf]"

adgjlsfhk1 — Thu, 22 Oct 2026 03:29:51 +0000

> Lim f/g -> C a constant.

This isn't quite correct. O(g) doesn't require the limit to exist.

New comment by adgjlsfhk1 in "Κ = 0.003 for integer multiplication crossed"

adgjlsfhk1 — Thu, 22 Oct 2026 02:17:27 +0000

yeah in ram models, this turn into O(n^1-kappa) , but otherwise the result is pretty similar (but you can drop the bit network part since matrix transposition is easy in RAM)

New comment by adgjlsfhk1 in "Κ = 0.003 for integer multiplication crossed"

adgjlsfhk1 — Thu, 22 Oct 2026 02:00:09 +0000

The tldr is that you can turn a single fft into an n dimensional fft, into n 1d ffts and then share a very small amount of the work of the multiple ffts

New comment by adgjlsfhk1 in "Κ = 0.003 for integer multiplication crossed"

adgjlsfhk1 — Thu, 22 Oct 2026 01:56:53 +0000

Have you considered looking at this in ram model rather than multi tape? it makes things a lot simpler

Imaging Black Holes with Julia [video]

adgjlsfhk1 — Wed, 21 Oct 2026 22:22:35 +0000

Article URL: https://www.youtube.com/watch?v=eHTuWt3JnZc

Comments URL: https://news.ycombinator.com/item?id=50188486

Points: 2

# Comments: 0

New comment by adgjlsfhk1 in "Anthropic logged net loss of over $80B in first half of 2026"

adgjlsfhk1 — Wed, 21 Oct 2026 06:46:12 +0000

it's probably closer to 50k than 5k

New comment by adgjlsfhk1 in "Galois Finite Fields and the Advanced Encryption Standard (AES)"

adgjlsfhk1 — Tue, 20 Oct 2026 16:01:59 +0000

> have significant advantages.

Such as?

New comment by adgjlsfhk1 in "Galois Finite Fields and the Advanced Encryption Standard (AES)"

adgjlsfhk1 — Tue, 20 Oct 2026 05:05:56 +0000

CBC historically was the most popular, but modern applications mostly use GCM which is counter mode. Block mode is too easy to get wrong, and doing it right is ~2x more expensive and less paralellizable.

New comment by adgjlsfhk1 in "One Hash Bloom Filters"

adgjlsfhk1 — Tue, 20 Oct 2026 02:34:34 +0000

yeah. You also don't need complete independence. You can use partially overlapping bits to get semi-independent hashes.

New comment by adgjlsfhk1 in "Your C is not standard C"

adgjlsfhk1 — Thu, 15 Oct 2026 04:53:04 +0000

store to a global is typically 0 cycles (although subsequent reads may be slowed down)

New comment by adgjlsfhk1 in "Auto Research: Integer multiplication below n log n"

adgjlsfhk1 — Wed, 14 Oct 2026 06:21:42 +0000

this fft result works in that model (although it becomes slightly more complicated)

New comment by adgjlsfhk1 in "Auto Research: Integer multiplication below n log n"

adgjlsfhk1 — Wed, 14 Oct 2026 04:11:50 +0000

likely neither. The current bounds are very ad-hoc robot derived and definitely seem improvable. epsilon seems unlikely as there isn't an obvious reason why overhead should go to 0. I doubt we're anywhere near even being able to convincingly conjecture what the correct lower bound should be.

New comment by adgjlsfhk1 in "11 year old unsoundness bugs found in Rust borrow checker"

adgjlsfhk1 — Wed, 14 Oct 2026 03:06:13 +0000

> It leads to broken OOM behaviour

no it doesn't. the broken oom behavior is distros not configuring earlyoom/systemd-oom properly. The problem here is that the kernel doesn't know what software is critical to your OS so the distro needs to be responsible for what processes to protect.

For apps that need to handle oom, Linux is missing something a handle to tell apps to shrink their memory use. Apple does this and it seems to work quite well (and proves that Linux style overcommit works well since it's what they use)

New comment by adgjlsfhk1 in "Auto Research: Integer multiplication below n log n"

adgjlsfhk1 — Tue, 13 Oct 2026 06:13:22 +0000

no it doesn't. making the table takes O(2^n) time

New comment by adgjlsfhk1 in "Auto Research: Integer multiplication below n log n"

adgjlsfhk1 — Tue, 13 Oct 2026 06:11:18 +0000

this blog post is notably wrong on several counts. Firstly, the Also the 10^-13 isn't an optimized number. Current results are up about 10 orders of magnitude (10^-3) and show little sign of stopping. it's seems entirely plausible that we might get to nsqrt(log(n)) (with a likely still astronomical constant)

New comment by adgjlsfhk1 in "4-hour battery storage is cheaper to install than gas turbines all across globe"

adgjlsfhk1 — Thu, 08 Oct 2026 18:41:05 +0000

Those don't really happen at continent scale.

New comment by adgjlsfhk1 in "Claude Haiku 5.5"

adgjlsfhk1 — Wed, 07 Oct 2026 18:57:53 +0000

The cost per fixed level of intelligence is dropping, but we're also getting dramatically more intelligent models.