a-dub

59 ms·

901.
by a-dub 4y ago
i suppose the more interesting question is if the device allows a researcher to inject i2c/ddc/etc packets in order to simulate a hostile display and test the drivers.
902.
by a-dub 4y ago
you can do i2c/ddc over dvi as well... i think even old vga analog cabling supported it. has anybody actually hacked a monitor yet? has anybody actually hacked a monitor in a meaningful way yet?
903.
by a-dub 4y ago
i used to daydream about adding an ALLOW DOING IT LIVE option to cassandra's csql client. in the event that your where clause was incompatible with your table's key, it would just wholesale dump the table in question into a sqlite
904.
by a-dub 4y ago
thanks. can you recommend a good introduction to policy terms and associated wonkery for computer types? ... like some kind of policy in a nutshell ora book?
905.
by a-dub 4y ago
it seems to me like the obviously correct policy intervention would be to legislate a reduction in switching costs/friction across the industry, as that is literally a cornerstone of how free markets are supposed to work. that said, ea
906.
by a-dub 4y ago
> Specifically, when learning a new field of math, the proofs sometimes don't feel rigorous to me right away, but once I get used to the the basics and the linguistic conventions, I'm more able to fill in the holes in my head t
907.
by a-dub 4y ago
this is all true. and yes, one should never be rude and one should always be respectful of others' time and efforts. but there is indeed another side to the equation: software does not succeed in a vacuum. it is a symbiosis between use
908.
by a-dub 4y ago
hm... so maybe a better place would be one of these toolkits like jax where the entire computation is known at optimization time where a blas would potentially have to do some heroic heuristics to try and fully optimize underneath the blas
909.
by a-dub 4y ago
that's the obvious issue. another less intuitive one is that modern cpus and compilers are essentially optimized for iterative algorithms. state variables can be kept in registers, simd can be used for data that appears in contiguous a
910.
by a-dub 4y ago
any thoughts on trying to build a sort of vq-blas?
911.
by a-dub 4y ago
that's the classic excuse. but even azure, which was built from the ground up, is pretty weird compared to its major competitors. also, apple showed us how to do this. build the right thing for today and use virtualization to support y
912.
by a-dub 4y ago
yeah the acceptance of oss is cool but they still ship a lot of weird software and technology... which is fine assuming you can avoid it... but sometimes you can't and then it becomes torturous.
913.
by a-dub 4y ago
i think the problem is that most people try to follow the recursions in their head which is hard. the biggest lesson regarding recursion is that if you're lucky enough to have your problem fit with tco (or have guarantees your problem
914.
by a-dub 4y ago
> “We’re going to launch lots of tiny spacecraft into interstellar space, have them discover alien intelligence, then bring back its technology to advance human technology by a million years” "whartonite seeks space monkey"
915.
by a-dub 4y ago
i uninstalled twitter off my phone to stop the elon musk notifications. i guess i'll miss out on his future crowdsource pump and dumps, but that's pretty sketch anyhow. i have a lot of respect for tesla and what they've done,
916.
by a-dub 4y ago
most numerical algorithms are loops over arrays, accumulators and simple arithmetic. this is where numba shines. for the other cases, there's a python compatibility mode (on by default) that allows for use of arbitrary python. the hard
917.
by a-dub 4y ago
salvage faceplates from a bunch of old suns and netapps and use them to cover all your modern gear! put an atari console in the middle with a cartridge sticker with the name of your company on and a portable crt television with rotary chann
918.
by a-dub 4y ago
now that i read more carefully, looks like they're just making an attempt at a line breaking algorithm for their own edification and sharing the results. some parts of it make it sound like they're trying to beat tex, but on secon
919.
by a-dub 4y ago
> For each line measure the difference between the desired width and the actual width and square the value. Then add these values together. seems like you could add something to the loss function like sum (over hyphens) k*(1/1+d)^2
920.
by a-dub 4y ago
that's exactly right. psychophysics is fascinating! my favorite example of a shepard tone appears in an old wilco track that appears to be missing from the list. notably shepard wrapping forms the basis of octave invariant chroma scale
921.
by a-dub 4y ago
i don't know what the proper term is, so i'm just going to make bjork happy and call them aerodynamic scientifical tassles. are they just recorded visually by a camera and then inspected manually for a test flight or are they part
922.
by a-dub 4y ago
mutt was notable for it's automatic threading support. personally i found it interesting but had an emotional attachment to pine. i suspect that the threading support in gmail was probably inspired by mutt.
923.
by a-dub 4y ago
what's the performance cost vs. intel code on the host under rosetta2 and native arm code on the host?
924.
by a-dub 4y ago
a lot of the performance gains in rosetta 2 come from load time translation of executables and libraries. so when you run a binary on the host mac, rosetta jumps in, does a one time translation to something that can run natively on the mx p
925.
by a-dub 4y ago
when the m1 initially shipped, i was watching this issue like a hawk. rosetta was really cool technology, but they seemed to really miss the mark in terms of many developer workflows that involve intel vms and containers. it was particularl
926.
by a-dub 4y ago
how does m2 neural engine performance compare with popular contemporary nvidia laptop and desktop gpus?
927.
by a-dub 4y ago
interesting. so this fixes the major performance issue by allowing users to install a user installable rosetta inside user created vms. could still end up with some "works for me" but still broken on prod issues resulting from the
928.
by a-dub 4y ago
i think the big problem with ipv6 has been that there just hasn't been a compelling enough reason for most users to care. it's great for network operators and planners who worry about things like limited address space, but otherwi
929.
by a-dub 4y ago
sadly that's about all that i remember. that and all the aol ops people used aol screennames. i'm sure that there are some old school aol ops staff who hang out on here, no?
930.
by a-dub 4y ago
i'm envisioning more of a sort of ipvx, 64 bit addresses, encryption is non-optional and dos resistance is built-in.

More ›