You can certainly write a lot of verbiage to mask the bigotry, that is for certain. You seem to think you can win debates by typing endlessly and using ChatGPT to increase the number of words.

Plenty of couples cannot have children, and are allowed to marry. Unless you want to ban infertile heterosexual couples from marrying, this is a bizarre point.

The twisted logic bigots have to apply.

I'm sure many liberals, communists and moderates thought the same thing of the Brown Shirts in late Weimar Germany.

There was a very nasty surprise waiting for them. Once you gain the levers of power, and you are sufficiently motivated and unhinged from any kind of sense of obligation, decorum or constraint, you don't have to be a majority. You just have to be willing to use raw applications of power. Illegal immigrants are not the only people that are going to end up getting sent to Alligator Alcatraz. They're just the test subjects for the inevitable liquidation of all political opposition.

And even if he does, or doesn't make it until the end, do you actually think his allies and backers, and the Republican Party itself, is ever going to give the Democrats a chance to hold any kind of power again?

And this AC, even if trolling, exemplifies the kind of people that cheer on tyrants, always believing that only the people they hate will be targets.

But it costs 100x as much, if not more. Running an LLM can be done on a notebook these days. But training one requires an entire data center of expensive GPUs. Not to mention that the notebook will run a reduced (quantized) version. Go check huggingface how large the full models are.

And also, LLMs are still suffering from a number of issues. For example, on many non-trivial tasks, the LLM is still unable to follow simple instructions. If you use LLMs routinely, you likely found cases where it has zeroed in on one - wrong - answer and no amount of prompting can convince it to give you a different one. It'll even totally ignore very clear and explicit prompts to not give that same answer again.

A human will understand "if you give that answer again, you're fired". An LLM... well you can tell it that it'll get shot between the eyes if it repeats that once more and it'll tell you where to get help if you have suicidal thoughts.

These things are both amazing and amazingly dumb at the same time.

If I'm understanding the article correctly, the conference room window mitigation wouldn't work against this. It doesn't rely on vibrations of the windows. Instead, you'd just need a piece of paper inside the room, lit by ordinary lamps. As long as the light reflecting off the paper could pass through the windows unmodified (i.e. the windows provide clear visibility) the white noise vibrations of the windows would have no effect.

On the other hand, lightweight curtains that blocked the view through the window would stop this technique, but probably wouldn't significantly reduce what was detectable from a laser bounced off the windows (assuming no white noise).

I think we're talking past one another. I'll try to be clearer.

Sure, but that's not the point. The point is that Android does prevent most users from using anything other than Play. Not by actually blocking them from using other app stores but by simply not offering the option. And that's a good thing, because most users have no idea how to decide whether or not something is safe.

I think perhaps the confusion here is because you and I are looking at this from different directions. You seem to be looking at it from the perspective of what you or I might want to choose. I'm looking at it from the perspective of an engineer whose job is to keep 3B users safe, most of whom have no idea how to make judgments about what is safe and what isn't. Keeping them within the fenced garden (it's a low fence, but still a fence) allows them to do what they want without taking much risk. The fact that the fence is easily stepped over preserves the freedom of more clueful and/or adventurous users to take greater risks. I think this has been a good balance.

I'm pretty sure that the ability to allow unknown sources is required by the Android compliance definition document, and that a manufacturer who disables it is not allowed to call their device Android, or to pre-install the Google apps or Play.

Until Epic decides that they want their store to be able to install and update as seamlessly as Play can, and gets a court to order that. Still, your point is valid, there is still some friction for other stores. Is it enough? I guess we'll find out. Will it be allowed to remain? I guess we'll find that out, too.

goo.gl links are a significant abuse vector, so Google has to maintain a non-trivial team to monitor and mitigate the abuse. I'll bet there are several full-time employees working on that, and that the total annual cost is seven figures.

Even if it weren't an abuse vector, the nature of Google's internal development processes mean that no service can be left completely unstaffed. The environment and libraries are constantly evolving, and all the services require constant attention to prevent bit rot. A fraction of one engineer would probably be enough for something like goo.gl if it weren't abused, but that's still six figures per year, not pennies.

This isn't true for the vast majority of Android users. To a first approximation, all Android users are using devices that have "unknown sources" disabled, so they can only get stuff from Play. Of course, it's trivial to find out how to enable unknown sources and install stuff from other places and I'd expect that nearly all slashdotters who use Android have at least experimented with that, even if they don't use f-droid or whatever on a regular basis. But slashdotters are not remotely a good representative sample of Android users.

If you're talking about desktop/laptop software, sure... but most Android users don't use a desktop or a laptop and are accustomed to expecting that anything they can install is safe. And even among those who do use a non-mobile device, people expect mobile devices to be safer, because they are. This court ruling may change that, to some degree. The result will probably be good for Apple, since Android insecurity will drive people to the safety of Apple's walled garden.

Delegating security decisions to users is the best way to ensure that users have no security. I'm all for enabling users who understand what they're doing to make their own choices and are willing to accept the consequences, but the vast, vast majority don't understand security or the consequences of their security decisions, especially not in the face of clever attackers who are quite good at making malware appear completely innocuous. Even a knowledgeable security professional can't reliably distinguish malware from a legitimate app, not without deep and very specific expertise, and not always even then, and you think your grandma can?

There are three billion Android devices in the world; it's used by approximately 1/3 of all people living, and they put a lot of very important information about themselves in their devices. Android platform security decisions have enormous consequences. Android has gradually gotten more opinionated about user security because we've found time and again that if you ask users, they don't understand the implications and they make bad choices.

Many people think that the existence of unlockable bootloaders and the developer options are bad choices and suggest that we should push the Android ecosystem into the Apple model of closed, locked-down hardware and a closed app ecosystem. I disagree, and I've worked hard to make sure that the ability of people to run the software they want on the hardware they own is not restricted. For example, I have regular meetings with the leaders of various Android ROMs, including Lineage, Graphene, Calyx, etc., to help them navigate the security hardware changes that we make. This isn't something I do because my management tells me to, it's something I do on my own because I think it's important.

User freedom is deeply important to me... and so is user security, but these things are in tension. To a first approximation, increasing one decreases the other. IMO, Android has struck the right balance. By default, devices are locked down and software comes from a controlled source, but users who know what they're doing have the right and ability to remove the restrictions (mostly; low-level firmware is locked down -- I would like to see Android gain a "dev screw" capability like ChromeOS to completely open it up in a safe way). This court ruling seems likely to upset that balance in a direction that endangers users who don't know what they're doing -- and it doesn't provide any additional capabilities to users who do. It's all risk, no benefit.

Try a web search for my username and "Android". Or look for "swillden" in the AOSP codebase and commit logs. Seriously, why would you imply that I'm lying when it's extremely easy to verify? And if you think that I made up a /. username to match some rando Android engineer, look at my /. UID. I've been on /. since before Android even existed.

You have an odd definition of "subsidize".