The compiled Chrome binary will be different than anything from the chromium repo because Chrome also includes the non-free portions. It's not going to be that easy to verify.

That said, I'm optimistic about this approach as it could easily be replicated by other browsers that are more open.

I am a bit concerned about the FLOC cohort calculation is currently done (see the open source repo). They're currently taking a hash of the domain names that a user visits and using that as the cohort identifier.

This seems like it would perform poorly for users who only browse large sites with content for users with diverse interests (such as news sites and large retailers).

Another concern is that more sophisticated AI techniques would likely be needed to target users effectively, which seems like it might give Google a competitive advantage due to their AI infrastructure.

Did you say antitrust?

Apple doesn't need to respond with a PR push when they are simply adding features users want to their tech platform.

Facebook is trying to get legislation to support it's outdated business model instead of adapting to the changing market. Hey Facebook, you should ask the buggy whip manufacturers how that worked out for them.

"In our own hiring, we will now treat these new career certificates as the equivalent of a four-year degree for related roles,"

Google is already notorious for ignoring degrees as qualifications for employment anyway. They have their own blessed interview process that they trust instead.

So I could see that these degrees would be worth the same amount to them, 0 == 0.

That's true. The article is almost entirely click bait, manufacturing a controversy where there probably isn't one.

That said, Apple has been pushing heavy privacy restrictions for web apps for a while, but has only recently started to apply some of those same standards to apps from their app store. If this article came out last year it might be accurate, but Apple's recent actions kind of paint a different picture.

RTFA:
"PWAs work even when you have no network connection."

Everything that would have been stored locally in an app is stored locally for a PWA. So a PWA that doesn't work when you're offline would have been an app that doesn't work offline.

I can guarantee that this feature is "not known internally as the quieter notification permission UI". Given the nature of engineers everywhere it probably has a name like Snorlax or silencio and the engineers quietly giggle as they discuss a trix of the weather in zork's borg cluster

(No, not that zork, the one for kids. No, the one that uses voice recognition. No, not the distributed voice recognition library, it's a service. No, not the zork that was cancelled last fall, the one that started 2 years ago....).

Yes, but in a regular flu season most people don't catch it, since they have partial immunity. The flu typically infects 15-20% of the population each year. We have no immunity to the coronavirus, so I've heard epidemiologists suggest 50-80% of the population is expected to get it.

Additionally flu cases are typically spread out over the entire flu season, so the demand for medical care is relatively flat. In the case of a pandemic, the growth in cases is exponential, so most of the cases happen at the same time, overloading the medical system. When this happened in China the CFR (case fatality rate) jumped from 2% to 15%, as the people who would have recovered with medical care started to die.

I think they can force Apple to pay them. The shareholders control the board of directors so at any time they could direct the company to liquidate assets and provide the proceeds as dividends (or use as part of a stock buy-back). If the share holders all decide to scuttle the company like that, then there's nothing the CEO or other employees of the company can do.

Copyright should require FRAND licensing. None of this hide it in the vault crap.

Given Google recently announced that they were deprecating the entire "Works with Nest" API, I wonder if Google already knew of some architectural security flaws with the API that could not be easily fixed. That might explain why they would be willing to take the bad PR that comes with deprecating such a popular API with such short notice.

To me it's the best argument for *any* single payer plan, whether it's individual or government run.

Of course the problem with individuals paying is that they may not be able to pay for it. And sometimes by the time you prove you do have money its too late to start some kinds of care (such as emergencies).

First world countries have "solved" this problem by requiring such emergency care to be given to anyone without requiring them to prove that they can pay. Enlightened first world countries have realized that only paying for emergency care is actually more expensive than completely paying for all healthcare (emergencies are more expensive than prevention), so they let the government be the single payer and then just cover everybody.

Of course if there is a way to tell immediately whether someone is able and willing to pay (possibly by borrowing our making some other arrangement), even if that person is unconscious, has pieces missing, etc. then you can stop covering everyone's emergency care and just leave those who can't pay to their own devices. At that point individual payer becomes practical again, and the poor can be left to die in ditches as they did in ye good olde days.

You clearly have never worked in the defense industry. Having a good acronym can make or break a project there.

And we can call these people "electors" and they can all go to an "electoral college" to vote.

In all seriousness, that's not a terrible idea, but we would still need everyone to hold their designated elector accountable. And that kind of gets back to the original problem (since they need to know the candidate's positions and the elector's vote to hold them accountable).

The reason they're doing this is not to track people more. They're doing this so more developers use the ping attribute for this functionality instead of hacky JavaScript or redirects (which prevent the user from seeing what URL the link goes to, increase navigation latency since everything ends up serialized, make it hard to copy the real URL or open the URL in a new window, etc.).

If things go as they typically do, browsers will start blocking the old behavior from working or otherwise disincentivising that behavior once enough of the internet has migrated.