They are probably working towards cleaning out the 30 series stock given that crypto demand is dying off. It takes a lot to put 3060 prices on a marketing slide for 40 series.
But you do? You expose some information on the range of the value via the time it took to sample assuming for example that the attacker knows the rejection sampling method in use.
I would say that people who regularly invest in top end hardware don't care as much about power bills. Otherwise power efficient chips is the norm (laptops, phones, etc).
4.5x faster seems perfectly reasonable to me, I always see the x as meaning a multiplier (or divider) over the original. Saying that this is 450% faster would be blasphemous.
Come on, burying side loading in a setting and having a giant red warning is good enough to keep the vast majority of users out of using it. No reason they can't offer that technically.
I absolutely love that NASA is doing this AND encouraging the RISC-V ecosystem. Maybe it's impossible, I don't know, but if we have an open soruce alternative for hardware with the kind of reach that Linux has in software i think that's the best way forward.
20 kids x 25mbps streams each is still only half a gig. You are way off here. The only way you saturate a gig is if you download big games from Steam, etc. If each of your kid is doing that you would probably run into fair use limits and have to set up a central PC for catching downloads and stuff but that's still a niche use case.
CUDA is their moat and their drivers do add significant value, seemingly more than AMD. So it's reasonable to expect that these won't be open sourced any time soon. Honestly I'm surprised why people expect them to just open source core IP like this.
Any reference for this claim? In my opinion this is most certainly not the case - random forests are hard to overfit compared to gradient boosted trees but you can overfit with them too if you don't tune your parameters right.
Overfitting is generally a function of the the size of your data and the complexity expressible in your model.
This is NOT aimed at people like you. Tons of folks out there who either 1) don't have the skills to be competitive without pay2win, or 2) don't want to waste time on useless grinding and have the money to buy items outright.
The point is that the relevant feedback is that Wikipedia is not a reliable source, not that you should not cite Wikipedia. They sound similar but they are fundamentally very different.
For former encourages the behaviour of finding other sources that are reliable. The latter encourages quoting Wikipedia without citing it.
Not really. pbfid needn't be a hash. You can take the plain text url, which can be ".../random.user/post/post_number", happened a random salt, and encrypt it using a key which is a function of "f(random.user)". That way you get unique encodings for each shared URL and every time you decrypt just discard the random bit. Defeats all pooling/reverse engineering efforts and offers perfect user isolation (each user has their own key).
I mean sure, but no one is forcing users to install software from APT groups. That's a choice people make. The general point is that you should be able to use hardware for general computing. There can be a switch that by default limits to software signed by your original manufacturer but that should toggleable by the end user.