Y
HN Search
Hacker News Search
new
|
comments
|
top
|
jobs
CarVac
searching PlanetScale…
1.
▲
2.
▲
3.
▲
4.
▲
5.
▲
6.
▲
50 ms
·
631.
▲
by
CarVac
6y ago
The only things I can possibly criticize is the global illumination lagging behind environment changes (presumably because it's iterative) and the water simulation not being movie-quality. Otherwise it's simply astonishing.
632.
▲
by
CarVac
6y ago
The only problem with this idea is self-organization of people into echo chambers, since misinformation doesn't spread randomly and uniformly like a disease.
633.
▲
by
CarVac
6y ago
They are also using robotic welding, such as when assembling the domes.
634.
▲
by
CarVac
6y ago
Payload capacity would not be diminished. Steel is denser but also stronger than aluminum, so you make the walls thinner and there's no mass penalty. Thin walls, though, are difficult to work with, though that's been done before w
635.
▲
by
CarVac
6y ago
To be mass-efficient, steel for a rocket the size of Falcon 9 would have to be so thin as to crumple easily—a "balloon tank". This makes manufacturing and handling difficult, as the structure must be either fully supported or pres
636.
▲
by
CarVac
6y ago
Steel needs less heat shielding than carbon fiber. Additionally the ease of manufacture is not just a manufacturing cost saver, it's a development cost and time saver. Falcon is a frozen design and won't change, beyond a proposed
637.
▲
by
CarVac
6y ago
I would attribute it to an inadequate equilibrium, a local minimum that's simply too hard to escape at this point.
638.
▲
by
CarVac
6y ago
How do you verify that ECC is actually functioning without just trusting what the hardware tells you though?
639.
▲
by
CarVac
6y ago
With constant exhaust constituents, exhaust velocity is rocket-engine efficiency.
640.
▲
by
CarVac
6y ago
It takes a couple seconds to start the first time you launch it. Very much not instant.
641.
▲
by
CarVac
6y ago
They went from individual stampings welded together to very large die-cast parts.
642.
▲
by
CarVac
6y ago
I never have this problem because it always takes me solid weeks to properly design and implement a single new feature in my photo editor's user interface. Sure, I could slap something shitty together in a weekend, but then you have a
643.
▲
by
CarVac
7y ago
The Pixel 2 uses true monocular depth estimation for the front camera.
644.
▲
by
CarVac
7y ago
When you use a full-frame lens on a crop sensor: * Vignetting always improves * First-order distortion improves, second-order distortion sometimes gets worse. * Corner sharpness usually improves, unless midway out on the frame was the weake
645.
▲
by
CarVac
7y ago
I think Google Pixels are individually calibrated.
646.
▲
by
CarVac
7y ago
Dispersion does come up; the last time I got new lenses for my glasses, one of the brochures on the table had the Abbe number [0] listed for various materials. Generally lenses with the highest refractive indices have fairly low Abbe number
647.
▲
by
CarVac
7y ago
My favorite site for explaining lens aberrations (unfortunately, it's currently under construction in a new URL) is here: https://web.archive.org/web/20090204211139/http://toothwalke... It's a
648.
▲
by
CarVac
7y ago
To convey the former meaning I'd probably use "haven't seen shit".
649.
▲
by
CarVac
7y ago
More margins make for more profit. It's always good to reduce costs regardless of how much you charge.
650.
▲
by
CarVac
7y ago
Attempting to achieve perfection is a problem, but at a minimum I believe any feature needs a good long time spent considering pain points for a wide variety of users. Achieving perfection is impossible, but you need to at least improve the
651.
▲
by
CarVac
7y ago
Google Lens has done this decently for me in the past.
652.
▲
by
CarVac
7y ago
No, I don't. I use Linux strictly on the desktop.
653.
▲
by
CarVac
7y ago
I don't understand how for some people it "never works", while for me it's literally never once failed, all the way from 6.10 up until now, my latest desktop install lasting through 12 updates, plus motherboard and GPU c
654.
▲
by
CarVac
7y ago
I got a very similar reply from Bob Menendez. >Thank you for contacting me to express your concern for the Eliminating Abusive and Rampant Neglect of Interactive Technologies (EARN IT) Act. Your opinion is very important to me, and I ap
655.
▲
by
CarVac
7y ago
Alcohol will destroy the charges that make the fabric effective for small particle sizes, according to the article.
656.
▲
by
CarVac
7y ago
Alas, I never got the chance to give a presentation and draw on the screen with fire. I always wanted to. KDE has zooming built in nowadays.
657.
▲
by
CarVac
7y ago
To add to this, make sure you do a varety of sites and source texts. Try transcribing articles as you read them.
658.
▲
by
CarVac
7y ago
Oh, neat. The annuals make for a stunning field. See my top-level comment for an album of photos I took in a cloverleaf.
659.
▲
by
CarVac
7y ago
That's real bokeh from a 36×24mm sensor and either moderately fast apertures or long focal lengths.
660.
▲
by
CarVac
7y ago
The NJ Turnpike has been planting lots of wildflowers, which make an attractive subject for photography. One cloverleaf near me had cosmos: https://flic.kr/s/aHsmG1mX94
More ›