bippihippi1

9 ms·

61.
by bippihippi1 3y ago
interesting investigation and js. Though I'm wondering why does windows rely on the software to uninstall itself?
62.
by bippihippi1 3y ago
surveillance into private places is illegal. you can't film in someone's window, you can't wifi vision through someones walls. whether or not that is prevented from happening by manufacturers and that the law is enforced depe
63.
by bippihippi1 3y ago
cite source?
64.
by bippihippi1 3y ago
I would guess that's different because people who follow a page want to see that content, and ads are shown to people who might not want to see that.
65.
by bippihippi1 3y ago
huggingface integration?
66.
by bippihippi1 3y ago
even 1% of a billion dollars is a lot of money
67.
by bippihippi1 3y ago
no, foss means free and open source. you're still able to use a license that dictates how it can be used
68.
by bippihippi1 3y ago
if you train a model on modular arithmatic, it can only learn what's in the training data. if all of the examples are of the form a + b mod 10, it isn't likely to generalize to be able to solve a + b mod 12. a human can learn the
69.
by bippihippi1 3y ago
it's been proven that all models learned by gradient descent are equivalent to kernel machines. interpolation isn't generalization. if theres a new input sufficiently different from the training data the behaviour is unknown
70.
by bippihippi1 3y ago
if you wait to fight a fire until you're breathing smoke, it'll be too late
71.
by bippihippi1 3y ago
google shows ads not good content. let's hope those are more aligned in the future
72.
by bippihippi1 3y ago
for security, all dangerous malware runs on encrypted traffic
73.
by bippihippi1 3y ago
the reason the uk wants an encryption backdoor is because it's expensive to do statistical analysis of encrypted traffic. there's ways to make it more difficult, but if you own the certificate that a tls endpoint uses you can just
74.
by bippihippi1 3y ago
i think it's part marketing hype to convince people it's cooler than it is, and that ai companies are not stealing your work by training models on it. ai doesn't generate content, all models learned by gradient descent are ke
75.
by bippihippi1 3y ago
good luck finding the telemetry in the decompiled meeting connector server
76.
by bippihippi1 3y ago
i get text messages with links and they don't ever respond
77.
by bippihippi1 3y ago
the analogy to compelled speech is a little weak, since google doesn't claim ownership of any speech they display, and since google search is an effective monopoly on information discovery. it's more like an isp choosing not to se
78.
by bippihippi1 3y ago
how do you prove it's the same people if they use different numbers?
79.
by bippihippi1 3y ago
they don't want to rush it and waste the evidence before a full case is ready. if you lose a suit you can't charge them again with the same evidence
80.
by bippihippi1 3y ago
it is not
81.
by bippihippi1 3y ago
if a private person does that for commercial reasons, it's harassment. if an ad company stalks you and send you bad ads, it's harassment.
82.
by bippihippi1 3y ago
would you really not expect that people shouldn't follow you around? if i drive around the city taking pictures of you in your car, that's harassment. why can a company do it? it's one thing for it to be available for polling
83.
by bippihippi1 3y ago
creating an LLM (kernel machine) from art is certainly fair use, as it's transformative. using a kernel machine to output original works or combinations of original works, and publishing that is certainly copyright infringement. buildi
84.
by bippihippi1 3y ago
doesn't ot have to be opt in? you need to give consent. you can't gove consent until they've told you what cookies there are and click a button. if you don't click those can't be added if you're in the eu
85.
by bippihippi1 3y ago
attention is turing complete https://news.ycombinator.com/item?id=36332033 I'd guess it's not as efficient as a native algorithm i many cases though
86.
by bippihippi1 3y ago
At least for me, naturally having experienced needing to context switch often (getting disctracted or multithreading or refreshing forgotten context), I have gotten good at it not by remebering more but by storing the context in an easily r
87.
by bippihippi1 3y ago
all gradient learning creates kernel machines
88.
by bippihippi1 3y ago
speculation olive oil is made from the fruit like avocado and coconut. they are easier to extract with fewer chemicals
89.
by bippihippi1 3y ago
if you don't have the problem it solves, don't use it. you need clusters services that scale up and down quickly. it's not the best way to deploy one server, it's the best way to deploy 10k servers, turn them off, and de
90.
by bippihippi1 3y ago
that demographic is also likely to have an outaized effect on evangelizing cool new features and getting people to swap over from other products. how much google is trying to grow vs curb competition is the question.

More ›