shaklee3

13 ms·

151.
by shaklee3 4y ago
take a look at @lionnetpiere on Twitter. he's done extensive analyses on the launch costs and revenue from service. most other estimates leave out big parts of the costs, so he's going the most accurate I've seen.
152.
by shaklee3 4y ago
launching more satellites has a tremendous cost, but a small, incremental capacity improvement since most of the capacity is stranded over water at any time.
153.
by shaklee3 4y ago
your history is a bit off. the fcc commissioner and fcc has generally loved SpaceX, and they were not appointed by Biden. what happened was SpaceX continually got worse in the fcc samknows testing, so much so that their original application
154.
by shaklee3 4y ago
bandwidth matters a lot. it's why SpaceX pays so much on legal fees for fcc filings
155.
by shaklee3 4y ago
no, it's not a strong sign of that. it's a strong sign that with VC money and government money you can offer unrealistic service at unrealistic prices. but when you need to start showing that you are profitable and launching satel
156.
by shaklee3 4y ago
this has been repeated countless times, but the only reason you are experiencing good service is because it's not at capacity. as soon as your area gets to capacity, just like it is all over the US, you too will face bad speeds and inc
157.
by shaklee3 4y ago
that's not true. satellite internet within the last 15 years has used adaptive modulation and coding, and rain doesn't have nearly the effect you're saying unless it's torrential and at an angle.
158.
by shaklee3 4y ago
if you read the article it doesn't just work for a lot of people, and is very slow. this is nothing that wasn't easily predictable, but you having a good experience where there's clearly a surplus of capacity doesn't mea
159.
by shaklee3 4y ago
superpeels work wonders. I've had mine for about 6 years.
160.
by shaklee3 4y ago
Nvidia had tensor cores before the tpu existed
161.
by shaklee3 4y ago
as pointed out, both Jetson and UM do this, but the upcoming Grace does it at higher bandwidth than the apple chip.
162.
by shaklee3 4y ago
from what I recall carbon doesn't have any working interop yet. it's just a feature on the roadmap.
163.
by shaklee3 4y ago
Sure, but the problem is that the places you might have high bandwidth might already have good access to internet. Hawaii in general has very good internet already, so asking them to pay a premium for satellite internet isn't always a
164.
by shaklee3 4y ago
This is a really poor way to compare. even with cross-links about 30% of their bandwidth is useable in the best case. starlink has no way of increasing capacity to their heaviest demand areas besides launching thousands more satellites. fib
165.
by shaklee3 4y ago
no, that's why they're losing money on starlink and keep raising prices
166.
by shaklee3 4y ago
besides the answers already given, the network is already the bottleneck on supercomputers on a local network. training that over a WAN would exacerbate that issue and make training even slower.
167.
by shaklee3 4y ago
I read it a few years ago too. Paul McKinney is the author of RCU in the Linux kernel, so a lot of the book is centered around it.
168.
by shaklee3 4y ago
It's slow, so while concurrency is great, it will never replace the core libraries written in C, C++, or FORTRAN.
169.
by shaklee3 4y ago
as opposed to zoom that doesn't even have a search?
170.
by shaklee3 4y ago
agreed. I use teams every day, and the meetings features are great. chat isn't as good, but it's useable. zoom is terrible and lacking basic features
171.
by shaklee3 4y ago
nice work! I tried a couple and they look great
172.
by shaklee3 4y ago
The claim being made in the original post and subsequent posts is that c++ is losing popularity. this claim is not true, and all evidence points to it gaining in popularity despite the HN echo chamber that it's too hard.
173.
by shaklee3 4y ago
again, use a better source if you don't like it.
174.
by shaklee3 4y ago
do you have a better source? stack overflow says the same thing.
175.
by shaklee3 4y ago
your link never said what you claim, and in fact he says it's gaining in popularity if he has to guess.
176.
by shaklee3 4y ago
unfortunately, the facts disagree with you: https://www.tiobe.com/tiobe-index/ it's still dominating and gaining. > but they’re limited by compatibility with existing feature no, they deprecate features as well
177.
by shaklee3 4y ago
someone posted a thorough benchmark on both compile and run times of rust vs C++ recently. c++ won on almost every test, so it's not as fast as rust people tend to spread.
178.
by shaklee3 4y ago
cmake is fine. just learn modern cmake. most of the complaints are from pre-3.11.
179.
by shaklee3 4y ago
you're correct. c++ isn't a dying language, and it's getting better every 3 years. the rust crowd likes to use that meme to gain adoption, but in the real world c++ is still dominating with no signs of stopping.
180.
by shaklee3 4y ago
V100 is 2 major generations old.

More ›