dathinab

57 ms·

781.
by dathinab 2y ago
- to structure your thoughts - document them for a future you - learn to better do technical communication with people which might not be experts on the specific topic (through without feedback this is just so so) - sometimes to give HR mor
782.
by dathinab 2y ago
the point here is that you need to be able to analyses things to get evidence and everything else is just a "seems to" and it is quite suspicious if someone decides to breach laws they had been complying with in the past just to p
783.
by dathinab 2y ago
this has nothing to do with privacy this isn't about private messages send on X It's about being able to factually analyses and judge how various entities try to __publicly__ infer with elections through various means like propaga
784.
by dathinab 2y ago
> If Apple can provide the UK government with foreign data, what's to stop Russia or China making them provide data on UK minister's phones, or more likely dissidents in exile? nothing the first precedence of not-draft law here
785.
by dathinab 2y ago
> but many other mutexes don't. yes and that's most times the better design decision > operating on an &mut T, while state is invalid. but you normally don't and for many use cases is in my experiment a non-issue pa
786.
by dathinab 2y ago
The problem starts earlier with the secret key which you can't place "into" a TKey. You can deterministically derive one between the TKey and a server using some thing like a (semi) static DH but that isn't how it is imp
787.
by dathinab 2y ago
their approach is a one which is fully stateless on the chip (see their documentation for why) theoretically this is quite desirable but various protocols aren't build in a way enabling it e.g. TOTP pushes a secret from the server to t
788.
by dathinab 2y ago
I know some do. But as long as you don't do anything unusual you are basically introducing a (potentially huge) availability risk to your service for no reason but except not liking panics. Like it's now enough for there to be a s
789.
by dathinab 2y ago
> The Cloud Act applies applies explicitly to daughter companies of US parents no matter which country they are based in
790.
by dathinab 2y ago
but it's a different legal structure US daughter companies have not means at all to direct their parent company to breach EU law, the other way around is more complicated
791.
by dathinab 2y ago
Technically they also have EU daughter companies operating this services etc. this are legally seen EU companies and have to strictly comply with EU law no matter what US law says. But US law like cloud act is a broad overreach of US law in
792.
by dathinab 2y ago
Also EU daughter companies of US tech giants are still legally EU companies (owned by US companies) legally they have to strictly comply with EU law and it matters shit what US law says (from the EU legislative POV) so this puts them into a
793.
by dathinab 2y ago
Honestly most non lobbyist people here in the EU has been expecting this to fall apart sooner or later, even without Trump. Some companies have gone further and not only are assuming data transfer to the EU will become illegal but also that
794.
by dathinab 2y ago
they are still anyway designed to be recoverable as a fundamental aspect of rust so that e.g. you web server doesn't fall over just because one request handler panics and it still is fundamental required that any code is safe in contex
795.
by dathinab 2y ago
> I would say that you are using them incorrectly if you assume them as recoverable. no it's them being recoverable at well defined boundaries is a _fundamental_ design aspect of rust > Overhead comes from the cleaning process. I
796.
by dathinab 2y ago
> fn check_invariant(&self) { > unsafe { assert_unchecked(self.ofs < > self.data.len()) } > } Is fundamentally unsound `check_invariant` needs to be unsafe as it doesn't actually check the invariant but tel
797.
by dathinab 2y ago
> Unrecoverable panics are very much designed to be recoverable at some well defined boundaries (e.g. the request handler of a web server, a thread in a thread pool etc.) this is where most of it's overhead comes from you can use pa
798.
by dathinab 2y ago
no it doesn't because here you are accessing their actions (approaching your front door in some unusual context like holding a crow bar) in addition, having front doors idk. calling the police on people just because it's a "u
799.
by dathinab 2y ago
ah yes countries doing their job of trying to protect their are a dystopia, sure
800.
by dathinab 2y ago
> just software, statistics, old ML techniques yes, and with the same problems if applied to the same use cases in the same way in turn they get regulated, too it would be strange to limited a law to some specific technical implementati
801.
by dathinab 2y ago
not only is Verilog comparatively "niche" the mind model behind it very different then that of "normal" programming languages, so less reuse of learned knowledge from other places ("knowledge" for a lack of bet
802.
by dathinab 2y ago
> Is there evidence more baking won't happen? No, actually there is a lot of evidence that it will still be worked on. Normally I would just say look at the issue linked in the nightly docs but due to an overlap of tacking PR and mo
803.
by dathinab 2y ago
but even then wouldn't the people using OpenAI still be the author/copyright holder and never OpenAI? (as no human on OpenAIs side is involved in the process of creating the works)
804.
by dathinab 2y ago
The US ruled that AI produced things are by themself not copyrightable. So no, it doesn't belong to OpenAI. You might be able to sue for penalties for breach of contract of the TOS, but that doesn't give them the right to the mode
805.
by dathinab 2y ago
they probably want to create enough legal instability to prevent companies from using this model internally for their use cases
806.
by dathinab 2y ago
a right the link below you can't read without loggin in because X penalizes links in posts no matter how much sense they make :shaking_head: (and because it got infected with the same kind of toxic design so common with apps today wher
807.
by dathinab 2y ago
There are different versions of the model as well as using it with different levels of quantization. Some variants of DeepSeek-R1 can be run on 2x H100 GPUs, and some people managed to get still quite decent results with a even stronger dis
808.
by dathinab 2y ago
Looking around a bit the price was ~70k$USD _in China_ around the time they where released in 2023, cheaper bulk sells where a thing, too. Note that this are the China prices with high markup due to export controls etc. The price of a H800
809.
by dathinab 2y ago
> Also you never replied about paypal not being a real bank, why should I reply on things which have nothing to do with the discussion? This never had been about weather PayPal is good or bad or anything, but that it's is very domin
810.
by dathinab 2y ago
In both the EU and US PayPal is very widely used to pay in online shops. To a point that disabling PayPal temporary can show noticeable decrease in sales for most online shops, especially if international orders are involved. When I sayed &

More ›