Y
HN Search
Hacker News Search
new
|
comments
|
top
|
jobs
hansvm
searching PlanetScale…
1.
▲
2.
▲
3.
▲
4.
▲
5.
▲
6.
▲
57 ms
·
961.
▲
by
hansvm
2y ago
It's something like 4 post-tax paychecks for people at minimum wage, even better if they weren't full-time (more common than you might imagine even amongst people trying to move up, since low-wage employers tend to shove hours dow
962.
▲
by
hansvm
2y ago
> You're making it sound like Drop is regularly not called Sorry if it came across that way. My claim is weaker, that it's sometimes not called even in cases where it could be with a smart enough programmer/language. >
963.
▲
by
hansvm
2y ago
Some of this looks great, but some of those benefits are literally never going to materialize. > barcodes highlighting allergens and other dangers Walgreens, CVS, and Walmart already sell homeopathic sugar pills labeled as closely as leg
964.
▲
by
hansvm
2y ago
It adds levity to the article and also introduces the reader to the sorts of things that can go wrong if they try it at home. The last paragraph highlights how they fixed one of the main pitfalls I normally see in this sort of thing, where
965.
▲
by
hansvm
2y ago
> You literally don't leak memory in safe rust, that is kinda of the whole point. That's a common misconception. Safe Rust prevents use-after-free, double-free, accessing uninitialized memory, (an extremely constrained form of)
966.
▲
by
hansvm
2y ago
> Well so Zig pointers, by default, has more UB than Rust pointers, at least when concerning null pointers. Not quite. In Zig, a (non-C) pointer type is closer to a Rust reference in its behavior. It isn't null, and that's enfo
967.
▲
by
hansvm
2y ago
> try between the alloc/free The idiomatic way to handle that is with some combination of `defer` and `errdefer`. Then, in code, there are no lines between alloc/free, and if a try would exit before the free then you would inst
968.
▲
by
hansvm
2y ago
Alright, let's take a look at that first link as if it's gospel. AWS charging excessively for inter-AZ networking is: 1. a microeconomic pricing strategy 2. where largely similar goods (AWS with or without substantial inter-AZ ban
969.
▲
by
hansvm
2y ago
The incremental cost of a pound of fuel is a rounding error in shipping. You might factor it in, but it'd be very far down the list of priorities. Interestingly, cars have similar economics. Details vary quite a bit, but every 100lb (n
970.
▲
by
hansvm
2y ago
It is, but it's not _that_ many. AWS pricing is complicated, but for fairly standard services and assuming bulk discounts at ~100TB level, your break-even points for requests/network vs storage happens at: 1. (modifications) 4200
971.
▲
by
hansvm
2y ago
It's a bit of a mix, but price discrimination isn't far off. It's like the SSO tax; all organizations are paying for effectively the same service, but the provider has found a minor way to cripple the service that selectively
972.
▲
by
hansvm
2y ago
I think so, and they're definitely more severe. Gear-up landings occur in something like 1/150k flights, and they're rarely fatal. Stalls happen in roughly 1/100k, near-stalls more frequently, and if unrecovered then the
973.
▲
by
hansvm
2y ago
Doing so complicates other hairy scenarios because of the increased hydraulic/power demands, reduced clearance, increased drag on the plane, worse handling, and the fact that in some emergencies you want a gear-up landing and would exp
974.
▲
by
hansvm
2y ago
I have no idea _why_ they wrote that up, but the points do separately make sense. They're willing to pay for a flight in the abstract, just not in the current timeline where employers know they're supposed to pay for it.
975.
▲
by
hansvm
2y ago
Did more than a few ask for feedback?
976.
▲
by
hansvm
2y ago
You (almost always) have to pay them for any work they actually did. If you catch a North Korean citizen day 1 of onboarding, you're obligated to kick them out immediately, and you might have to pay them for the few hours they were the
977.
▲
by
hansvm
2y ago
You can often work days to years before people catch on that you are (a) unqualified, (b) underqualified, (c) not legally allowed to work in a particular jursdiction, (d) overemployed, (e) leaking company secrets to ChatGPT, .... On top of
978.
▲
by
hansvm
2y ago
Floats are deterministic too (this winds up being helpful if you want to do something like test an algorithm on every single float); you just might get different deterministic outcomes on different compilation targets or with threaded inter
979.
▲
by
hansvm
2y ago
It only changes who actually gathered the CSAM they asked this person to label. OpenAI definitely gathered it.
980.
▲
by
hansvm
2y ago
With our current understanding of biology, if I model outcomes probabilistically (luck) and you model them any other non-equivalent way then I'll be more successful predicting those outcomes. The philosophy behind it is interesting, bu
981.
▲
by
hansvm
2y ago
Thanks for providing a concrete model to work with. Compared to GPT3.5, the number you're looking for is ~0.04%. I pointed out the napkin math because 0.00000001% was so obviously wrong even at a glance that it was hurting your claim.
982.
▲
by
hansvm
2y ago
A calorie surplus of 1800/day is ~190 lbs/yr. Is that something people actually do?
983.
▲
by
hansvm
2y ago
That's (obviously) a bit of an exaggeration. BERT is just another transformer architecture. Cut down from ~100 layers to 1, ~1k dimensions to ~10, and ~10k tokens to 100, and you're only 1e6 faster / more efficient, still a f
984.
▲
by
hansvm
2y ago
I have no clue in general. When I go through applications, I go through all the material and anything else I can find about the candidate, and cover letters are sometimes helpful. Put yourself in my shoes: - If the cover letter is a rehash
985.
▲
by
hansvm
2y ago
That touches on a couple related principles: - not doing extra work now if it's not necessary yet and if it's as cheap to do later - delaying building a thing till you know what that thing should be
986.
▲
by
hansvm
2y ago
I'm not sure that's at all the same thing with the extremely high rate of violent crime involving guns in Brazil. Pointing a gun straight at a city background, missing your specific target, and still injuring/killing somebody
987.
▲
by
hansvm
2y ago
> obvious reasons It's maybe worth mentioning that the obvious reasons almost entirely have to do with the things you'd hit in the air, not once the bullet descended. If everyone on earth suddenly shot upward at around 80 degre
988.
▲
by
hansvm
2y ago
Or not bad enough to overcome other benefits that came out around the same time.
989.
▲
by
hansvm
2y ago
Linters suffer from a false positive/negative tradeoff that AI can improve. If they falsely flag things then developers tend to automatically ignore or silence the linter. If they don't flag a thing then ... well ... they didn
990.
▲
by
hansvm
2y ago
If you think of the "AI" benefit as more of a linter then the value starts to become clear. E.g., I've never hesitated to add a linting rule requiring `except Exception:` in lieu of `except:` in Python, since the latter is ve
More ›