dathinab

9 ms·

91.
by dathinab 4mo ago
> I don't understand how Codex can blunder so badly. because they trust the AI too much (and seem to be fin with acting knowingly negligent) the problem is - AI tends to produces very convincing looking code, even if fully wrong - A
92.
by dathinab 4mo ago
racial slurs are an impossible problem you can only approximate based on non trivial linguistic analysis. Regexes can't handle that without you accidentally messing up badly including, discriminating against minorities in a potentially
93.
by dathinab 4mo ago
Lol "fix this code" is beautiful. Like it basically jail broke the "no security vul guard rails" not in any clever way but just by fixing them, producing exploit code just by writing test cases making sure it's fixe
94.
by dathinab 4mo ago
32bit WASM doesn't have a strict 4GiB limit (if the runtime and OS it runs on is 64bit, which is normally the case) the thing is in WASM "memory" is more or less a resizable ArrayBuffer and while each has an effective 4GiB li
95.
by dathinab 4mo ago
or they increasingly copy what a lot of their training data does, discussion for discussion sake and arguments for the sake of winning them instead of productive outcomes. probably the truth is somewhere in between
96.
by dathinab 4mo ago
That the AIs where trained on what humans wrote on the internet forms is increasingly sowing as they incresingly mirror all the bad things which are so common on such forums, like: - non stop, non productive discussions - gaslighting - val
97.
by dathinab 4mo ago
> beside-the-point semantic nits all over the place. This is also a problem with Copilot Reviews on GitHub. We have them enabled (but opt in) and they have, multiple times, spotted quite useful things. Sure often the thing they spot is j
98.
by dathinab 4mo ago
> Canadian government directly liquidated their property and put them into camps. While the mechanism was different that was what _effectively_ happened to US Japanese people, too. The taxes part is more like a separate mechanism of disc
99.
by dathinab 4mo ago
yeah, and implicitly force them to sell the land they own for less then it's worth, which in combination with setting up very messed up tax related laws in some states (1) which highly benefit you if you bought land longer in the past
100.
by dathinab 4mo ago
Yes. Extremists or in general any fraction willing to engage in systematic discrimination, harassment, terrorizing or similar love highly detailed non anonymized census data. Why? Because it gives them the perfect layout for which areas to
101.
by dathinab 4mo ago
For completeness another trick to deceive people can be to have (git/http) sources from places other then just the official repo, like in the example you linked. When changed they will just show up as a a "hash" change... whi
102.
by dathinab 4mo ago
> hasn't happened earlier. it happens all the time Just not always on this scale and doesn't always end up on HN. Similar to how you don't see every npm supply chain attack or malicious github action or similar on HN. In g
103.
by dathinab 4mo ago
both allow it but only if you use quoted text AFIK through the message does allow an additional display name (like `display name <email>`) which has it's own rules.
104.
by dathinab 4mo ago
> Note: I have struggled to verify this one, and it’s possible I’m actually misreading the RFC. Is correct, you can have quoted local parts and (I guess?) theoretically "foo"@mail and foo@mail should even be treated the same. B
105.
by dathinab 4mo ago
> Punycode [...] and the local-part was still limited to ASCII. the funny part is this is only half true The true part: Punycode has never be standardized for the localpart and as such taking a email address with non us-ascii characters
106.
by dathinab 4mo ago
I really wonder how legal that is. Or more precisely suspect it is very much illegal. like think about it it's pretty much a tool which intentionally silently sabotages you if you try to compete with the tool maker It is like selling a
107.
by dathinab 4mo ago
no, it's very much compatible with GDPR and other laws, as it clearly (enough, kinda) communicated 1. what data they keep/collect 2. what they do with it (and that there is a reason to have it) 3. with whom they share it 4. how lo
108.
by dathinab 4mo ago
but wrt. the points of the law suite it is already regulated, that is why there is a law suite think about it, it would be absurd if you had to redefine deceptive, misleading or negligent business practices for every single kind of product
109.
by dathinab 5mo ago
yep and a lot of "smooth" aerodynamic surfaces have "microscopic"/"very small" surface patterns to make the surface less perfect smooth as if it is too perfect smooth the air kinda "sticks" to it
110.
by dathinab 5mo ago
or it is pretty much over before it even began, because he sued way too late...
111.
by dathinab 5mo ago
one one had Japan seem to have quite bad luck with the specific tree(s) mass planted but also on the other hand in Germany problems with allergies are very common and a pretty big deal for many people, it's just that we got used to it
112.
by dathinab 5mo ago
I'm confused by your answer, the previous post doesn't seem to be about vibe-coding at all. It seems to be more about: 1. auto grouping duplicate security reports 2. auto validating if they are likely viable or likely nonsense 3.
113.
by dathinab 5mo ago
the problem here is that many of the submissions are not "make-believe work" but actual existing security issues it's just that in the past people most times didn't find security vulnerabilities independently of each oth
114.
by dathinab 5mo ago
putting aside what others have commented about the truastability of studies and similar - what all studies show is some vague "craving" for something generic, e.g. the link between iron deficiency and craving for ice - but what yo
115.
by dathinab 5mo ago
no and yes autists have often a much much stronger need for habits and avoidance of change. This includes a change of, or a less repeting/habitual diet. The effect if applified due to autism being commonly comorbid with ADHD and hyper
116.
by dathinab 5mo ago
Grate example of how misleading just reading headlines can be, even if the headline isn't intentional misleading/click bait. > reduce autism symptoms > chronic gastrointestinal issues are a harsh reality > boosting microb
117.
by dathinab 5mo ago
> Have you ever seen what comes out of c2rust? It's awful. It relies on a library of functions which emulate unsafe C pointer semantics with unsafe Rust. which is somewhat close to what their port produced... like their goal was fro
118.
by dathinab 5mo ago
> Did they even claim it was "memory-safe"? they didn't, actually the port is trying to be mostly 1:1 and in turn is mostly unsafe rust, which means no benefits initially but also doing the 1:1 port to mostly unsafe rust i
119.
by dathinab 5mo ago
> many multiuser Linux systems nowadays not relevant IMHO we don't live anymore in a time where you can trust that local apps do not misbehave, and in such a context LPE is pretty bad even in a single user system just thing about al
120.
by dathinab 5mo ago
most times it's done by (reliably re-)rooting a attested phone in a way which bypasses detection of the attestation system so not really useful for 3rd party ROMs

More ›