Y
HN Search
Hacker News Search
new
|
comments
|
top
|
jobs
a-dub
searching PlanetScale…
1.
▲
2.
▲
3.
▲
4.
▲
5.
▲
6.
▲
23 ms
·
451.
▲
by
a-dub
3y ago
because it's a tale of intrigue!
452.
▲
by
a-dub
3y ago
sounds like you want transmeta and the jvm to have a post-oop baby that natively understands matrix and vector math, simd and massive parallelism.
453.
▲
by
a-dub
3y ago
misery loves company. consider bringing on a business partner who knows product-market-fit, marketing and sales. disclaimer: i have never (yet) tried to build a saas business.
454.
▲
by
a-dub
3y ago
i believe that up until recent years replay/special features were illegal on machines in new york state as they were believed to be used to signal payouts in gambling settings.
455.
▲
by
a-dub
3y ago
my pet theory: consciousness, free will, the soul, the spice, whatever you want to call it is a sum of the evolutionary process which produces a living being plus the noise that its substrate and environment excite it with. i believe that t
456.
▲
by
a-dub
3y ago
signed code bundles with sts must-staple style semantics for preventing downgrades sounds reasonable. would probably also need some kind of protection in the browser runtime that prevents/limits scope of changes to execution that can b
457.
▲
by
a-dub
3y ago
cool library. although worth keeping in mind that strong typing doesn't solve all your problems. here's an example, taking the second example on this page https://docs.pydantic.dev/latest/ with the external_d
458.
▲
by
a-dub
3y ago
sortof. except two differences: 1) i'm primarily interested in data types, rather than objects, as i believe most runtime errors are the result of unexpected changes to data types. 2) i'm suggesting a type system that does away w
459.
▲
by
a-dub
3y ago
i've often thought a more interesting approach for python's duck typing would be if the interpreter looked for anomalous types on assignment. so build a compact distribution of type "shape" for every assignment and then
460.
▲
by
a-dub
3y ago
probably a lot of interesting work going on that looks like this for the virustotal db itself.
461.
▲
by
a-dub
3y ago
the world wide web was an internal tool.
462.
▲
by
a-dub
3y ago
reminds me of the days when s/keys over telnet were cool. there should be a vh1 where are they now special for assigned ports in /etc/services that shipped with early slackware.
463.
▲
by
a-dub
3y ago
very cool. reminds me of the alamo sculpture in astor place in front of the site of the old cooper union engineering building.
464.
▲
by
a-dub
3y ago
i seem to have vague recollection of a variant of this shirt that had a perl script on it? or was the perl script for decoding the barcode?
465.
▲
by
a-dub
3y ago
edit: nevermind, i need to read up on these.
466.
▲
by
a-dub
3y ago
numpy and scipy are largely just wrappers around native libraries that are already optimized machine code. numba is good for writing loops or doing other things you can pretty much imagine as simple C code. while it is pretty cool, it'
467.
▲
by
a-dub
3y ago
who remembers when the hip thing to do was to sign your webpages with a horizonal rule, followed by your name and your email address with a nice blue (on grey) mailto link, all in italics.
468.
▲
by
a-dub
3y ago
ahh yes, the rational r1000. an ada machine from the 70s that stored programs in a mixed ast/object data format called diana: https://insights.sei.cmu.edu/documents/948/1988_005_001_1565...
469.
▲
by
a-dub
3y ago
there's some machine from the 70s that does this. iirc it stores all source code in an ast like representation alongside binaries and has some kind of built in version control. wish i could remember the name...
470.
▲
by
a-dub
3y ago
"AGI: An autonomous system that surpasses human capabilities in the majority of economically valuable work." -- what an obscenely depressing reduction of a fascinating field of inquiry. who the hell snuck in and redefined the scie
471.
▲
by
a-dub
3y ago
i don't know enough about tax codes and how business works to comment on the actual policy, but it sure does seem to me that changes that result in immediate and massive new liabilities are unfair.
472.
▲
by
a-dub
3y ago
i think i liked listservs better. more verbal creativity.
473.
▲
by
a-dub
3y ago
> I spend most of my time thinking about code design, reading documentation, writing plans, discussing stuff with my colleagues—editing, writing code is maybe 30 – 40% of my time, at best. I also code fairly slowly anyway. ahh yes, the c
474.
▲
by
a-dub
3y ago
didn't terry gilliam make a movie about literal bugs in government computers ruining lives back in the early 80s?
475.
▲
by
a-dub
3y ago
> vi and vim have modes becaus they were created in an era where computing and terminal capabilities were very limited. that's the whole point. it was designed for limited terminals and high latency, very slow serial links. that mea
476.
▲
by
a-dub
3y ago
this is violence. i never knew it was possible to emit a shitpost in the form of a github project, i am simultaneously disgusted and humbled.
477.
▲
by
a-dub
3y ago
1) software grows and evolves in complexity over time while hardware remains constant until it is upgraded, this explains the macro or low frequency component of the slowdowns (ie, things that reinstalling the o/s won't totally so
478.
▲
by
a-dub
3y ago
> The 56k modem was a quasi-digital connection and as I recall asymmetric. ahh yeah i forgot about that. the other side was direct digital and often ISDN PRI. ascend (or what was left after lucent bought them) made a device called the po
479.
▲
by
a-dub
3y ago
i guess you're right about 33.6kbps... point stands, if you wanted to go faster than analog modems in residential settings, before DSL and DOCSIS, ISDN was often the only game in town. BRI did not have free local calling with pacific
480.
▲
by
a-dub
3y ago
bummer. i used to like the legend that it was all on one commodity linux pc implemented in some nice concise lisp running on sbcl. edit: my memory is crap. it was a single machine, but the codebase was written in a custom experimental langu
More ›