863 karma · joined September 1, 2010
I'm always amazed at companies that don't train their people. It's easier to say "they couldn't use HAVING in a query, pass" than to say "I believe I can teach them SQL effectively." But the freedom to think the latter lets you build a very strong team.
It's the genius of Facebook's OSS policy. Let everyone use their stuff so that new people can pop right in, learn on your own time before we're even paying you...
CSS loading/parsing is render blocking. By loading CSS that isn't needed immediately asynchronously in a way that will use the browser cache if it's available, it makes the page visible much faster. It feels like overkill in the simple example, but it's a tried and true method. ( https://github.com/filamentgroup/loadCSS has > 2k stars if that kinda thing means anything)
From my experience on a large website (m.trulia.com), On a 3G connection with an 'average' (nexus 5) phone, this technique made almost a 2 second difference in perceived load time.
So yeah, now I kinda get it.
Versioning APIs is hard. Anyone have any good reading / experience to share? How do you prevent too much code duplication across folders? How do you expediently insure that code that is shared doesn't break a distant dependency? Should you create lots of new versions or only for large changes? Does something like Semver help? I don't feel like I've seen a happy solution for this yet in my work and would like to learn more...
The website looks underwhelming, but the app is amazing. The more I use it, the more it's flexibility impresses me. There are lots of options to review the cards, and it'll use card data to make quizzes, etc for you.
Whenever I see a trolley go by in SF with dozens of people simultaneously filming video (and that trolley runs dozens of times a day) I always wonder about how well we could reconstruct our worlds digitally with everything we've got now.
The content may be indexed, but if your visitors are on a mobile network, that initial visit (or a visit with stale cache) is going to be crappy. It's great that they can read in they content (though bing cannot), but if it's buried on page two, does it even matter?
As someone who is a proponent of web perf, these kind of articles make me worried that server side rendering will be ignored because "SEO works now for Javascript", even if it's slow and google is only 70% desktop & 80% mobile search.
"Well you see, google had this idea that was gonna help everyone..."
Snark aside (gotta get it out of the way before commenting for real), I think this is a step in the right direction, and securing cookies, even though they still aren't "secure", is a big deal for everyone who uses a browser. I'm excited that they are willing to push forward something that's a little rough around the edges to start to get data and verify results.
There's a lot of money to be had (a well qualified mortgage lead is worth ~$90 a pop), but it seems inappropriate for google.
That sounds pretty awesome, actually. Who wouldn't want to add that wealth of experience to their web repo?
One uses money to manipulate the people in charge of socialist government programs. I'd bet that the dental lobbyists are stronger in their niche than the teachers union is because there's less public knowledge and interest there.
Also, I'd argue that capitalism is what causes the entrenchment of industries that get addicted to their profits and don't care about the customers.
When economists are unsure what exactly is going on, what does that mean? There's the quote by Buffet, "Be fearful when others are greedy and greedy when others are fearful".
What should one do when others have no idea what's going on?
If I can do everything with JS on the visitors browser, why not host some shell HTML on S3 and never worry about a server? Maybe hit AWS Lambda if need be for one specific thing? Dunno. The age of the do everything server seems to be coming to a close.
I'd love an app that gave me that ease of hosting and creation and generated a static site from there (hook it up to an S3 bucket or something). I'd pay for sure.
Currently I'm using Nginx w/ SSIs for lightly dynamic sites. Works well enough and is very very simple.
Never actually tested that out.
no kidding.
China's current reputation (fast and default to crappy quality without extreme oversight) doesn't really fit what I'd want out of an airliner. Optimistically, this project succeeds and changes a bit of the culture around quality and shortcuts in manufacturing. If "Made in China" could gain some equivalence to "Made in Germany" without exploding their cost structure. That'd be an interesting world.
This study has been done with people using names that are, black, foreign, feminine, etc. If you're different, some people are gonna think differently. When there's a "pipeline" little biases add up: Your teacher treats you differently, the college admission treats you differently, your professors treat you differently, recruiters treat you differently, your boss treats you differently. That might not happen in every step along the way, but the potential is there and there are many opportunities.
A white male Stanford grad from a well off family is flipping a coin once to see if they get a real chance to be successful in life. Someone who is discriminated against has to roll a D20 every step of the way to see if they make it.
If you want to make a difference, recognize these biases exist. Review CVs with all identifying information removed or generalized. Hopefully, eventually the bias will fade, but you can't ignore that it exists and that you can turn it off like a switch.
The presentation takes a while to get around to it, but it's a way to see your browser history using cert pinning/redirection. Specifically (ab)using features in the browser that even the paranoid don't disable. And then using that information to fingerprint you.
Demo: http://zyan.scripts.mit.edu/sniffly Code: https://github.com/diracdeltas/sniffly
It's a good find, even if it's not perfect the way the CSS link visited computed style checks was.
Seems perfect for deliveries around a large office campus where theft, maps, and smooth sidewalks are all very manageable. Though I don't know how much profit is in that market.
That was my first thought as well. Get a van and a sledgehammer and wait a few blocks away from their depot in the morning. Sure they could phone home, but what is the cost of maintaining a staff to protect the robots vs. just having the staff deliver things.
Reminds me of a friend who was a Marine Biologist. Their lab spent a lot of money to get an underwater probe that could go around and take automated readings based on a map. But the probe cost so much money that they had to follow it the whole time in a boat to make sure it didn't get lost.