a-dub

22 ms·

391.
by a-dub 2y ago
it's true, the techies of the '90s and '00s who lived in sf embraced the culture (and suffered the commute). what happened in the '10s kinda steamrolled it.
392.
by a-dub 2y ago
historically there was a lot of media and some internet in sf. more interesting hardware type tech companies generally clustered in the south bay with some exceptions (oqo, sega). twitter was notable because they put a big campus in midmark
393.
by a-dub 2y ago
> No you can’t roll back bad driver updates in any OS, if you could then by definition they do not sit in the kernel space. drivers and kernel binaries are typically installed and maintained by user space programs that run with some sort
394.
by a-dub 2y ago
i would have thought that in 2024 a bad driver update is something that windows would automatically roll back. or at least provided some level of protection against crashes in third party kernel code.
395.
by a-dub 2y ago
no, it is not. beyond overall poor software choices, there is no notion of persistence in a bad software update.
396.
by a-dub 2y ago
yes, but does it actually work end to end and actually deliver meaningful speedups that make it actually useful?
397.
by a-dub 2y ago
i wonder if simd is working. that would be cool.
398.
by a-dub 2y ago
there's also numpy and scipy in the webassembly python distro (pyodide). but the "kinda not" part more refers to first class scientific/numerical computing support. it's possible, but the libraries are all disjoint
399.
by a-dub 2y ago
so when do we get WebBLAS and WebFORTRAN? kinda joking, kinda not.
400.
by a-dub 2y ago
that's kind of the beauty of the system. we actually have no way of knowing if it was just random fill. maybe they were just random digits with an off by one error or some other problem with the symbol set missing one symbol. or maybe
401.
by a-dub 2y ago
see also: the conet project. home recordings of numbers stations from around the world: https://en.wikipedia.org/wiki/The_Conet_Project audio now free on the internet archive: https://archive.org/detail
402.
by a-dub 2y ago
he gives awesome talks as well!
403.
by a-dub 2y ago
lyons is a good intro but maybe a bit handwavey at times (although my copy is an edition from the 90s). consider maybe backing it up with one of the textbooks like oppenheim (the classic) or manolakis (one that i think i remember liking).
404.
by a-dub 2y ago
http://www.ee.columbia.edu/~dpwe/ https://github.com/librosa/librosa
405.
by a-dub 2y ago
might be neat if there were acceleration wins to be had with instructions for computing hash ints from machine ints and strings. or even better, a full fledged linear probing hash lookup instruction.
406.
by a-dub 2y ago
i think the key insight, at least for me, is that if you break things into piles of lots of little pieces and then you come up with n different ways of ordering those piles, similar things are going to have the same piece show up on top acr
407.
by a-dub 2y ago
hm. haven't seen camelCase hard real-time task definitions since the vxworks days!
408.
by a-dub 2y ago
> The “technical skill” everyone talks about is coding not really necessary or sufficient in many cases. the real question is whether or not someone has the background and interest to understand the details at hand and how they fit toget
409.
by a-dub 2y ago
> It's hard to imagine why they'd distract themselves with that, except perhaps with a token small-run demo for pure brand marketing purposes. haven't google and samsung started making their own socs?
410.
by a-dub 2y ago
a little backstory (that seems to be missing from the blog post). this is a technique that was invented in the early days of google for deduplicating the crawl set (turns out that building an llm and building a plain ol' web text index
411.
by a-dub 2y ago
more interesting explanation and benchmarks: https://shipilev.net/jvm/anatomy-quarks/25-implicit-null-che... hotspot's always-on runtime profiling (and corresponding runtime dynamic optimization) is pretty ne
412.
by a-dub 2y ago
i think the fact that they put a digital signature of any kind on the paper transcript is pretty cool. would be kinda neat if there were a combined standard where there were a deterministic scheme for ocr'ing the text (specifically its
413.
by a-dub 2y ago
will be interesting to see if google ships a decent desktop mode for the pixel line in the next revision.
414.
by a-dub 2y ago
idk. aren't they tier 1? what if it breaks peering? is all their management capability strictly "inside" their network? what would happen to the rest of the internet if they dropped off?
415.
by a-dub 2y ago
> Its not really going to affect their services, maybe email in/out of it but thats about it idk. how do you know that their management capability can run sans dns?
416.
by a-dub 2y ago
resource and data forks were hfs(+) features that appeared in pre-osx versions of macos. post-osx made use of the bsd fast filesystem and a rather nice unix style convention from nextstep where the on-disk representation of a .app or .pkg
417.
by a-dub 2y ago
too bad they've been kneecapped by the business casual crowd...
418.
by a-dub 2y ago
wild to think that the web is old enough now that major media properties are shutting down websites.
419.
by a-dub 2y ago
> Crypto isn't private as traceability is an inherent requirement for crypto to operate. monero is a counterexample to this. i wonder how many organizations that do payroll see a timestamp and location when you tap a credit card at
420.
by a-dub 2y ago
NYC OMNY allows for direct payment from a debit/credit card via NFC at the turnstile. this has privacy issues but i think they're ameliorated if you use a smartphone. i think you can buy a transit payment card if you need one. no

More ›