They could have made a version just for (E)AX. "general purpose" registers in x86 are not the same. AX is the accumulator, for arithmetic, BX is for indexing, CX is the loop counter and DX is for data and extending AX in divisions. You don't have to use them for that purpose, but you will have access more optimized instructions if you do. Out of these 4, AX is the most likely you would want to set to zero.
For loops, it is generally expected that you count down, with CX. The "LOOP" instruction is designed for this, so no special need to zero CX. SI and DI, the index registers may benefit from an optimized zeroing, for use with the "string" instructions.
Here I think Intel engineers didn't see the need and not having a special instruction to zero AX must simplify the decoder.
It is obvious that energy to weight ratio is one of the most important characteristic of power banks. Reputable manufacturers will optimize for this, and if they are particularly good at it, they can ask for a premium.
If a generic powerbank beats the big names, sells for cheaper and is associated with a brand that has no reason to be associated with powerbanks, then it is very likely that the weight savings come from omitting something important rather than an optimized design. If the specs are true that is.
Say using 20% more electrolyte would prevent insufficient spacing and separation of the plates. You get a safer/higher quality product but it's heavier. And maintaining the same safety with less electrolyte would involve higher development costs. There's a sliding scale where something is just too cheap/dangerous to sell at high volume to a general audience.
Hopefully these high quality CT scans show the battery makers that people are going to notice when too many corners have been cut, even if there isn't a flood of reports of their product causing fires (yet).
Probably, however, it makes me think of the Galaxy Note 7 problem. For those who don't remember, these phones had a battery problem that can cause them to catch fire, resulting in one of the most infamous smartphone recall campaign.
The cause of the problem is that the electrodes were pinched inside the pouch cell, in some cases causing a short circuit and a fire. To put it simply, it was too tight in there.
So here it is possible that these defects are a result of having too small safety margins for how these batteries are manufactured.
In addition, they only focused on the electrode alignment issue, because it is visible on their CT scans, they didn't study the electronics, the casing, the connectors, etc... There may be some other issues there.
Yeah. I remember seeing the circuit board for official Xbox controllers vs cheap 3rd party ones. The official controllers had about 10x as many components. I don’t know what all that stuff does, but I’m sure it all contributes to the controllers feeling and working better.
I wonder the same thing about chargers. I’ve recently moved from a 3rd party charger for my camera batteries I got on amazon to an official Sony charger. The 3rd party charger seemed to work great - but it was practically weightless. The Sony charger is clearly a way more complex (and more expensive) product. I don’t know if all that complexity is actually worth it. What does it all do? But I assume so.
Signal has always seen some controversy, usually centered around centralization. Also the MOB cryptocurrency, the use of phone numbers, contact discovery,... It has led to the promotion of alternatives such as Matrix and third party applications such as Molly.
But these alternatives are all niches compared to Signal. Which is to say something considering that Signal itself is a niche compared to Whatsapp.
Palantir sells software for analyzing data, like Excel but on a large scale. If "Chat Control" passes, they will need software to analyze the data they collect, which is exactly what Palantir sells. It is just business.
I don't know about Thorn but it looks like the same: they sell software that may be of use for implementing "Chat Control".
"It is a just a business" is crazy to say if your founder was Peter Thiel and you ostensibly merged already halfway with the operating goverment (US, DoD)
Of course if your organization have connections to a criminal organization, you are going to be in trouble. Same thing for refusing to cooperate with law enforcement, this is not some abstract thing, it is about following the law, for example relating to evidence tampering or search warrants.
I don't think France is anything special in that regard.
I wonder how much of it is just Arthur Conon Doyle hating his character. Which has was known for as the stories progressed. He even killed him, just to resurrect him later because of public demand.
He accumulated character flaws along the way, as if Doyle wanted to make Holmes as unsympathetic as he could without changing his core traits.
This is the kind of stories where I would like another angle.
First, to make things straight, neither Signal nor AdBlock are illegal in France, they are used by many without any problem. As in some other countries E2E encryption is sometimes attacked, and usually, it makes the news, but for now there is no problem using any privacy-focused app in France.
But, and I think it is what happened here, there is nothing stopping a prosecutor mentioning the use of such apps when making an argument in combination with other facts. It is like the use of cash. It is not illegal to use cash and possess large amounts of it (there are limits for large transactions though), but a prosecutor will certainly mention it if they can and it makes you more suspicious to justify an arrest. Not a conviction though, for a conviction, you need actual proof.
And of course, the defense attorney will point out the many legitimate uses of Signal as a way to discredit the accusation, this is the side we have here. A judge will hear both sides and settle the case in the end.
Not to say that there are not deeper political motivations, or that justice can't be corrupted, but I prefer not to make an opinion without hearing both sides.
Homan cognition is more advanced than in any other animal. I think it is clear enough. Humans are not the only animals that evolved higher intelligence, but we have a combination of attributes that made it really effective: we are larger animals (with room for a big brain) with a social structure and a relatively long lifespan (good for passing knowledge).
Ants beat us when it comes to society, but in a sense, we may also consider multicellular organisms as a society of single cells. Still impressive, and there is a good chance for ants to outlive us as a species, but we are still orders of magnitude more intelligent than ants, including collective intelligence.
By intelligence, I mean things like adaptability and problem solving, both collective and individual. It is evident in our ability to exploit resources no animals could, or our ability to live in places that would normally be unsurvivable to us. It doesn't mean we are the pinnacle of evolution, we have some pretty good competitors (including ants) but we are certainly the most advanced in one very imporant area.
Usual rant to news outlets: Cite your damn sources! (if they are public, which is the case here). This is a web page, not printed paper, web pages support links, links are not just for ads!
I understand not having a link to the primary source at the top of an article because you might be trying to write an explainer, but if you want to avoid confusing your readers but still do the right thing at the very least include a bunch of source links at the bottom of the article!
I know all the cynical reasons not to do it, but I feel like even the good faith objection has a pretty straightforward solution.
For loops, it is generally expected that you count down, with CX. The "LOOP" instruction is designed for this, so no special need to zero CX. SI and DI, the index registers may benefit from an optimized zeroing, for use with the "string" instructions.
Here I think Intel engineers didn't see the need and not having a special instruction to zero AX must simplify the decoder.
reply