I memorized it. I really don't want to install a client even though I wrote one. I wanted the protocol to be simpler so the lazy client would be simpler but I realize it needs to check for request forgery.
Authentication is by preshared keys and the client prompts for them or you can set environment variables. The preshared keys are static, yes.
$ time python youtube-dl-original --version
2020.11.01.1
real 0m1.474s
user 0m1.356s
sys 0m0.111s
$ time python youtube-dl-compiled --version
2020.11.01.1
real 0m0.380s
user 0m0.328s
sys 0m0.049s
It might be nice if youtube-dl would be compiled to bytecode before release.
Interesting. On the one hand Starlink costs 10x as much for equipment and 10x as much for service as compared to LTE. On the other hand Starlink is 10x faster and half the latency of LTE.
A simple factor of 10 increase in both price and speed is not impressive. I'm gonna say I'm not overwhelmingly convinced until Starlink delivers more for less.
> On the other hand Starlink is 10x faster and half the latency of LTE.
Hmm, are you sure you're actually on LTE? That sounds more like HSDPA numbers.
Just did a speed test on an iPhone 11 with Vodafone Ireland's not-very-inspiring LTE, in an urban area. Latency: 21ms, 147/19mbit. This costs 20 euro with a cap, mind you. Capless LTE is also available, but on a generally somewhat worse network.
Tell first: I've been reading Excession by Iain M Banks again. Not because Elon Musk pays superficial lip service to Iain Banks. Not because of peer pressure among posers to read Excession. Because I genuinely identify with some of the characters.
A theme in Excession is: nerds collect things. Gestra collects scale models of sailing ships. Pittance collects warships in stasis. Sleeper Service collects people in stasis. Grey Area collects torture devices.
Another theme in Excession is: nerds are not welcome in the Culture. I mentioned four nerd characters. Three of them leave the galaxy by the end of the story. Grey Area goes a step further and leaves the universe.
Yeah that's right. Too much literary analysis for Hacker News.
Show second: I've been collecting song metadata from Jango Radio. By which I mean I wrote code to collect metadata. It took a few months but I collected metadata for every song and made it searchable. I did it because I was looking for a few specific songs which had become unsearchable on Jango Radio. Without knowing the song IDs I was looking for so I had to grab every song ID and once I started I might as well finish the collection.
Bash and JSON and PHP and SQLite are tech nerd tools. Too technical for Hacker News. We all know Hacker News is named ironically because it's really a marketing site.
Finally: Coding is a worthless activity for nerds like Gestra and Grey Area and just as unprofitable as collecting scale model sailing ships. There's nothing to monetize about some nerd's collection of song metadata. There's nothing about coding to monetize. Marketing is what matters always.
Yeah that's right. Hacker News is not a tech site. Hacker News is a marketing site.
Well let's see. It's a reply to a Show HN about a coding project. The reply doesn't contain any feedback about the coding project. The reply doesn't contain any technical content at all.
I discovered a way to open random UDP ports on xfinitywifi. Then I modified AndIodine to enable raw mode on random ports. So much speedier than DNS mode. Fun times were had.
I know Hacker News is The Number 0x01 Best Place because when I search for "xfinitywifi" and "iodine" Hacker News comes right up.
So to do this the way I did it you would need to build AndIodine which is open source and Android SDK is easy to install right? And you need an iodine server and a cloud account somewhere to host it and a DNS zone you can delegate. Oh and the host running your iodine server needs to redirect all incoming UDP traffic into port 53 with iptables. Whew! Too much preparation?
Or just adapt my changes to your favorite censorship circumventing UDP based VPN app. You know you want to.
There should be like a whole startup dedicated to a "residential neighborhood guest network road warrior" app for no reason.
This is cool, but does stock AndIodine not work out of the box with raw UDP mode where you live?
The local xfinitywifi networks in my area work with raw UDP. That being said, I've been to some areas where the local xfinitywifi networks did not, and it reverted to DNS tunneling.
I'm also curious who you are using for a cloud provider, as many throttle/block inbound traffic on udp/53 to prevent DDoS attacks against improperly-configured DNS servers.
As for commercializing/publicizing this, I wouldn't recommend it. Judging from the effective MTU of xfinitywifi hotspots, it seems as though the gateways broadcasting xfinitywifi are establishing an IPSEC tunnel to a local Comcast-operated server. It would be trivial for them to shut this down in one fell swoop by transparently redirecting all outbound traffic on port 53 to an internal resolver.
Comcast didn't configure xfinitywifi the same way everywhere in Comcast country? Then what's the point of having one big internet service provider if it doesn't provide homogeneity as a service?
There's nothing to monetize here. I think censorship circumvention services like for example Psiphon should try harder than just hoping ports 53 are open. But Comcast country is part of the allegedly uncensored first world where there's no market for censorship circumvention.
Think probably though if your solution did become more widely adopted via a startup either Comcast would sue it out of existence or figure out away to block it right ?
> Doing something because other people are doing it is the path to misery and competition.
So very very wrong.
Doing everything everyone else is doing is the road to group membership and social acceptance and great cultural fit and the best jobs and the most lucrative opportunities and loads and loads of money.
Doing things specifically because nobody else is doing them is a ticket to a pit of despair full of poverty and loneliness and isolation.
I know the difference very well because I never do popular things. I always choose unpopular projects that nobody cares about because my attitude is if a project is too popular then too many people are doing it already. I want to do things that nobody else is doing because if not me then nobody else will do them.
Doing something that nobody else is doing is very rewarding. Talking to other people about it is misery. People do not care and you are competing for their attention with other things that they actually care about. And you lose the competition.
Thanks for sharing your perspective and the example about how you choose your projects. I don't disagree with you. This applies to some things but not others. I mostly referred to working on important problems, startups, and businesses. Even in startups and businesses, this can have the opposite effect (like the one you mentioned). In fact, in this very same quote I added a footnote below to explain this. I really liked your perspective about how if you end up being too extreme, you end up competing anyway but this time for the attention.
I kind of agree. Doing things that other people aren't doing (e.g. because it's hard) even though it needs to be done would be a better heuristic and can lead to an interesting career and possibly even success.
Authentication is by preshared keys and the client prompts for them or you can set environment variables. The preshared keys are static, yes.