Y
HN Search
Hacker News Search
new
|
comments
|
top
|
jobs
o11c
searching PlanetScale…
1.
▲
2.
▲
3.
▲
4.
▲
5.
▲
6.
▲
8 ms
·
61.
▲
by
o11c
11mo ago
Okay, so apparently HN has a comment size limit. Normally I customize this a bit based on exactly where my audience is coming from, and sometimes what question was asked. For HN I guess I'll assume someone who is aware of the pop-cultu
62.
▲
by
o11c
11mo ago
Note that "feeds" is a confusing word choice; I immediately thought of something RSS-like which often links to long-form content. For news, my rule is to check major headlines at most once a day (often less in practice), so I am a
63.
▲
by
o11c
11mo ago
Remember that even ELIZA fooled people. That doesn't make it useful, unless you think fooling people is itself a goal.
64.
▲
by
o11c
11mo ago
The one caveat to this is that you must perform a sufficiently-deep clone that you can actually reach the tag.
65.
▲
by
o11c
11mo ago
Eh, kind of. Peering in Europe is such a mess that even Cloudflare can be pretty bad. Sometimes you have to manually calculate "okay, there's a colo in this particular city that will force the correct route if we proxy all our t
66.
▲
by
o11c
11mo ago
Obligatory https://xkcd.com/1200/ Just replace "someone steals my laptop" with "Microsoft installs malware"
67.
▲
by
o11c
11mo ago
> [0] e.g. some lock-free algorithms depend on the presence of GC If this is what I'm thinking of, it does not require GC, only "defer deallocation until all running threads check in", which is ... not trivial , but certa
68.
▲
by
o11c
11mo ago
No, SCTP only fixes half of a problem, but also gratuitously introduces several additional flaws, even ignoring the "router support" problem. The only good answer is "a reliability layer on top of UDP"; fortunately every
69.
▲
by
o11c
11mo ago
TCP has another unfixable flaw - it cannot be properly secured. Writing a security layer on top of TCP can at most detect , not avoid , attacks. It is very easy for a malicious actor anywhere in the network to inject data into a connect
70.
▲
by
o11c
11mo ago
It's true for any C/C++ project which bothers to write 10 lines of GNU `make` code. The problem is that many projects still pander to inferior 1980s-era `make` implementations, and as such rely heavily on the abominations that are
71.
▲
by
o11c
11mo ago
... ugh, they actually made an `o11[a-z]` abbreviation? When I picked this nick, the only term I ever saw in the wild was `i18n`.
72.
▲
by
o11c
11mo ago
Yes, and most DBMS's are full of historical mistakes. In a sane world, statements outside `BEGIN` would be an unconditional error.
73.
▲
by
o11c
11mo ago
I would argue that it's a bug for it even to be possible to autocommit.
74.
▲
by
o11c
11mo ago
The thing the "just optimize AI" crowd misses is that this isn't like optimizing a programming language implementation, where even the worst implementation is likely only 100x slower than a good implementation. AI is million
75.
▲
by
o11c
11mo ago
The part you're missing is that a very large number of voters (on both political sides) expect nothing but outright grift and corruption from both parties. And they're not wrong to do so. Remember, Trump won both times against a
76.
▲
by
o11c
11mo ago
Another fundamental difference: OpenAI explicitly markets their tool as a replacement for the copyrighted material it was trained on. This is most explicit for image generation, but applies to text as well. As a reminder, the 4 factors of &
77.
▲
by
o11c
11mo ago
Typescript doesn't even support notions like "unsigned integer". It is not a serious attempt at type-safety; its main claim to fame is "better than raw Javascript" which is not saying much.
78.
▲
by
o11c
11mo ago
Hopefully, we'll just blacklist SEO spam at the same time. Slop is slop regardless of origin.
79.
▲
by
o11c
11mo ago
Does `mosh` work better than `ssh` for this?
80.
▲
by
o11c
11mo ago
And it is very important to remember: being able to do this is the reason why companies have brainwashed the Internet into choosing the MIT license for everything. With GPL-only code, the world would be much nicer for all of us.
81.
▲
by
o11c
11mo ago
Sounds exactly like my mom's Windows computer. Flaky wifi/power issues are not a Linux problem.
82.
▲
by
o11c
11mo ago
Even RMW instructions are very RISC compared to what people used to mean by CISC.
83.
▲
by
o11c
11mo ago
The C++ implementation, sure. But there is plenty of other implementation space without giving up like Java. With a trait-first implementation that mostly defers monomorphization and prefers "static if" over C++-style specialized
84.
▲
by
o11c
11mo ago
One helpful rule is: if you use `__del__`, it should be on a separate class which doesn't contain any methods or data except the native handle. You can't call inappropriate functions if you don't have any way to reach them!
85.
▲
by
o11c
11mo ago
> The mistake they made was buying not from Amazon, but from "Maestro Technology" listing on Amazon. If you understand that Amazon is a marketplace and you take 10 seconds to read who you're buying from, it's not a pr
86.
▲
by
o11c
11mo ago
Recently I've found Youtube to just show a white screen if it detects Firefox+uBO.
87.
▲
by
o11c
11mo ago
The NYT has problems with being a stooge of the military-industrial complex, but I really don't see them doing anything wrong in this case.
88.
▲
by
o11c
11mo ago
No, because OpenAI can change their server at any time - and does, to patch over the cases where their copyright infringement is obvious.
89.
▲
by
o11c
11mo ago
In practice, artifacts become common past something like 16 kHz. I'm not sure how much of this is math and how much is that almost all speakers are made very cheaply.
90.
▲
by
o11c
11mo ago
I do not have any good advice, only pain. It turns out that a lot of old tape players have either failed completely or else add excessive background hissing. Most tapes don't record all the way to the end so you will have to cut it
More ›