Signal message keys seem to include an IV. [1] Message keys are generated by passing a chain key through a key derivation function. [2] So you can also expand a shared secret established through Diffie-Hellman.
You can search `known_hosts` for a fingerprint using `ssh-keygen -l -f ~/.ssh/known_hosts` and `grep`. If you disable `HashKnownHosts` in your ssh config, the command will also tell you the name of the host(s) for that fingerprint, rather than its hash.
I use this all the time, because my AWS/Azure/GCP ips change constantly, but their pubkey fingerprints do not.
> That is the most stupid thing I've read in a while.
Arrogance.
It takes 10 lines of Python to compute a multiplicative inverse mod p, using arbitrary precision integers. [1] Try doing the same thing with nothing but x86 assembly and Volume 2 of Knuth's AOCP.
You will be happy to learn that under Chair Khan, the FTC is going to prioritize enforcement actions against this specific practice. Only 1 Commissioner issued a dissenting statement. [1]
Comcast has probably already discontinued the practice, as it's been 2 weeks since the policy statement was issued.
> Crypto continues to help nobody and achieve nothing in the real world
It's kind of surprising how little progress has been made.
It is absolutely possible to use privacy-respecting low-fee tokens to purchase ebooks, purchase "no advertisement" article reads, or subscribe to periodicals.
Even something like Taler is better than Visa and Mastercard. I would love to know why lwn.net, ars, or the register don't try. I would absolutely pay $0.25-$2.00 per read for the ~50 articles I read per month on these websites.
Micropayments for tech readers seems like a good place to start. Tech readers use ad blockers, but are also willing to try new tools with janky UX.
The "inflation targeting era" began around 1995 [1]. The official story explaing how 2% became an (informal) international standard is kind of interesting [2].
Not too long ago, people could encounter a grumpy store clerk and forget about it by the time they got home. Sometimes they'd even turn it into a joke [1]. It was a simpler, less angry time.
You're probably correct in this case, but not in general. The authors of [1] could have quietly run a money printer but elected to seek reforms instead.
Hamming is great in general. Some years ago, I picked up a copy of "Numerical Methods for Scientists and Engineers" for like $2 at a used book store. I had no idea who Hamming was, but it was cheap so I bought it.
After the first chapter I thought, "this guy is pretty sharp let's see what else he's written." That is when I found "You and Your Research".
I believe the United States slow walked its own Prompt Global Strike [1] program because it was perceived to be needlessly escalatory [2]. That sounds like sensible policy making.
The formula for error bars under the traditional Binomial assumption is:
+/- sqrt(p * (1-p) / N)
So the errors are around +/- 1% for most values of p in this article. The article (rightly) points out that the binomial assumption is not reasonable given the survey method.
There was a recent thread [1] optimizing some very primitive trig functions.
I recently watched a djb interview where he talked about the importance of fully utilizing available hardware [2 at 5:15]. That can be a good starting point, although at work its usually easier to just consume more resources than to use what you've got more efficiently.
In nvidia land, `nvidia-smi` is like `top` for your gpus. If you're running compiled CUDA, `nvprof` is very useful. But I'm not sure how much work it would take to profile something like a pytorch model.
Someone please correct me if I'm wrong.
[1] https://github.com/signalapp/libsignal-protocol-java/blob/ma...
[2] https://signal.org/docs/specifications/doubleratchet/Set0_1....