JaRail

I'd assume they're mostly fines against foreign or shell companies. What's effectively a shell company is set up to forward calls, then goes out of business when caught.

They'll buy tons of phone numbers from all over so they always have local numbers to call from. They just use VOIP software to route it to wherever the call center is located. As far as their provider is concerned, they're probably just a local business. Or, at most, a local number for a foreign company.

I'm excited to see improvements and results to the Google Assistant's on-device call screening. It's a step up from voicemail but I'm not sure how willing people will be to participate. For example, someone calling from a doctor's office might only be able to speak to a specific person. For privacy reasons, they call from blocked number and won't specify who they are until they've at least asked that they're speaking with the right person.

GV's block of phone numbers is probably considered low-value targets. Scammers get more value by talking to customers from big telcos. It's the whole "you don't have to outrun the bear, just your friend" kind of thing.

Interesting to see how much more excitement there is for AI in the developing world. They're the ones with the biggest problems so it makes sense. Seems NA/Europe feels like they have the most to lose. While the reality should hopefully be win-win, it does seem likely to level the playing field around the world. But a hyper-capitalist country like the US is definitely going to have the biggest social challenges adapting to job losses.

Automation + human classifiers + user reporting. The stuff automation gets wrong or isn't confident about goes into the training set and the system improves.

Even if they lose money by manually reviewing a small ad that was flagged as possibly political, it doesn't mean it costs more to review 1/100 low-volume ads. They're also paying to classify useful data to train their model.

Real people with real friends and families in mainland China, yes.

Running an outright dirty campaign would break ToS. So even if they were protected from deletion by virtue of being a politician, their tweets would no longer be included in feeds. It'd be political suicide.

Proper enforcement is hard. They won't get there soon. It'll come slowly for a few years to avoid mistakes. When AIs get reliable at detecting lies/harassment, enforcement will become the norm. (my prediction)

Those advertising methods don't allow you to micro-target based on people's private behavior. Micro-targeting and machine learning lets campaigns (like Cambridge Analytica's) generate what's effectively a thousand different campaigns, targeting exactly what triggers specific people the most. There's absolutely nothing stopping them from saying the exact opposite thing to two different people. That's what our current institutions may not be prepared to handle.

I have a hard time buying into that contrarian view in a post-Cambridge Analytica world. If anything, it feels like whataboutism to me. While Twitter is trying to deal with the mass corruption of democracy around the world from micro-targeted lies, they want to talk about some kid's high school student president campaign.

Seems pretty conflicted with their "think globally" core value. Maybe they should update that one to "divide and conquer."

And engineer some safe small-scale reactors for container/cruise ships. It's a crime we aren't hedging our bets on renewables with some next-gen prototype reactors.

They need to fix their reputation of awful privacy violations. While the reputation is more out of fear and speculation than it is of actual understanding, they still need to fix it.

So when companies like Apple are offering great security, people are scared Facebook is reading their conversations, continuously tracking their location, and listening to their microphones.

> In America, all we want is “drill, baby, drill”

Actually, it is much easier to build a pipeline, now that I think about it..

If they're worried about kids, maybe they should have asked to not enable E2E for Messenger Kids. There's already a feature to let parents see their kids' messages. I'm pretty sure they have no plans of breaking that. Government should be providing their expertise to help FB train AIs to help alert parents to suspicious chats. There are lots of good ways to protect kids. This ain't it.

Ask any woman/minority, domestic abuse victim, etc. They don't want people spying on their messages.

It'd be a really interesting project to data-mine a large number of IDEs at a company like google. Use machine learning to predict fixes to lint/compile errors, and surface those suggestions in the IDE. Would also be fun to do it with code reviews to see what kind of garbage suggestions you can generate there. Probably a lot of "I don't understand what this does" and "please write tests for this."

Would also be cool for auto-complete suggestions. I'm thinking more along the lines of extracting larger patterns than for loops, such as stubbing out classes that follow a projects existing style. For example, adding an auth check at the start of a web request handler if that's what it sees elsewhere.

Might also have needed a higher sample rate.

For battery life concerns, I can see them implementing some of the fall detection logic in hardware as well. For example, a high shock or shaking could trigger the fall-detection software to wake. Although, with the 60-second period, they probably wake the main processor often enough anyway that hardware detection might not save much power.

Would just be called non-JS. However you don't need to avoid javascript completely to cater to non-js users. The term "progressive enhancement" is used for a site that works without javascript but gets enhanced when it loads. Hacker News is like that. The links/buttons would trigger full page loads except the javascript cancels those actions and handles them without. The vote button won't refresh a page, for example.

I personally don't think there's a market for non-js users. Most of the web is pretty broken without. I've never encountered someone in the real world who lives this way, even in the tech world. At most, someone will install a no-script browser plugin. That makes whitelisting and enabling javascript dead easy for the user. The only legitimate non-js use cases, IMHO, are for extreme political cases where you'd also be using TOR.

Of course. But within the JavaScript community, it has taken on specific meaning: no 3rd party libraries at runtime.

For "vanilla js," I can still have compilers, type checking, advanced ECMAScript 2019 features, local storage, webgl, etc. Without that understanding, you could qualify something as "vanilla js" in lots of different ways.

If someone said "vanilla c++," they might mean c++ without the recent language features, or they could mean the latest language spec but no standard library. Maybe it's still okay to link OpenGL but not okay to use a desktop GUI framework? Maybe OS frameworks are okay but nothing more? There's no communal understanding there that I'm aware of.

When you're not using frameworks/libraries, it's typically referred to as vanilla js. Probably the name stuck thanks to this amazing satire of a js library. http://vanilla-js.com/. This was basically around the time browsers got good enough that jQuery was no longer needed. It essentially emphasized that all the showcase features of popular 'must-have' js libraries were now a standard part of modern browsers.

The trick is to keep your karma below 500. xD

The actual like is pretty small, about half a KB. Just a light graphql query that responds with updated counts for different reactions (happy/sad/etc). They definitely batch it with some other stuff though. Batching is really important for browser performance.

The exact details would depend on what they do with the uploads and the specific countries they're uploaded to/from. I'd presume they do this for security and debugging purposes, not to 'steal' the libraries. Like a virus company uploading samples of 'suspicious' .dlls for analysis, this looks like a fair use exception.

Well you've basically got two other options

* webapps -- we tried this (iPhone v1) and they were awful. The web keeps improving but apps have consistently been years ahead.

* native software -- still too difficult to secure reliably.

We haven't really been suffering from fragmentation. We've essentially consolidated down to two platforms. You can also develop for the web or use a cross-platform framework.

Not really sure what you see as flawed. The app model has been wildly successful.

> Not wearing a seatbelt can put others at risk for injury liability.

Wow, I'm surprised that has never occurred to me before. I guess that to most people, such as myself, the trade-off is so obvious from a personal safety point of view that we don't think too deeply over it. This is a great point.

> having no automatic update mechanism

Ignoring the windows store, this is really more of a "let developers handle their own updates" situation. The vast majority of modern windows software updates itself directly or through a loader (steam, etc). Windows developers can easily tie into distribution services that offer update mechanisms. The default user experience is not 'no automatic updates.'

Google's employee shuttles/buses are a decent example too.

He offended a broad class of people. Women shouldn't have to pick and choose who they feel comfortable working with around the office. Or worse, be forced to work with that guy. Once he expressed views like that, his fate was sealed. You don't punish the victims.

The point of incognito was never to hide that fact from the sites you're visiting. The point was to not leave a trace on the machine you're doing it from. Basically, it's for porn, not news.

Exactly. If you look at the rating for a McDonald's, the rating is "how clean is the bathroom and does the ice cream machine work?"