It's odd. The section headers says "Density matters" but then the experiment shows that, actually, density does not matter. Only the total number of branches encountered across the linear piece of code. (In the chart "block size" is what determines density.)
Also note in that chart what's being tested is unconditional jumps placed at unique addresses in a linear piece of code, not conditional jumps at the same address in a loop that fits entirely in L1i.
The atypical benchmark here is a manufactured worst case scenario for the purpose of quantifying the hardware capabilities. A deeper predictor means accommodating more complex program branching patterns. Obviously you'd expect to see diminishing returns versus silicone area at some point but I see no reason to assume that AMD would have made a poor allocation decision here.
It seems like that would struggle with detecting how many layers of branching to pay attention to. Imagine the two nested loops surrounded by a randomized one. Wouldn't that implementation keep hitting patterns it hadn't seen before?
Obviously that must be a solved problem; I'd be curious to know what the solution is.
Unfortunately that's your own misunderstanding. iOS (as well as modern android) quite effectively prevent phone theft while the electronics are in transit along the last mile of the supply chain. Anything beyond that is a happy accident.
(I'm being a bit overly cynical there but IMO only the tiniest bit.)
You don't need to. You just need to be in good condition yourself and actually paying attention. Professional delivery drivers routinely achieve seemingly absurd mileages per incident.
Well, if we were talking about forcing people to stop driving and transition to current waymos it's plausible that diligent sober drivers would be facing greater risk. Would that be acceptable to improve average statistics?
It also doesn't make sense because "get out of my way or I will ram you" is the default state of operating a motor vehicle. Not the goal but the physical reality of it.
That's unfortunate if true but it isn't a convincing argument to force the rest of society to live in proverbial padded cells. There's a minimum bar here. Some people probably shouldn't have online accounts and aren't responsible enough to manage their own finances. The rest of us are (hopefully at least marginally) functional adults.
This is actually a really interesting problem. Some portion of the public (nerds) are competent to understand what running software even means and the rest (let's call them "sheep") are naive and helpless. A portion of the nerds (Evil Hackers) are easily able to coach any sheep to do any action. Obviously everyone should default to being a Sheep, and obviously it would be ideal if Nerds could have root on their own damn hardware. But how can one ever self-certify that they're actually a Nerd in a way that an Evil Hacker can't coach a Sheep through? "Yes, now at the prompt that says 'Do not use this feature unless you are a software engineer. Especially don't click this button if someone contacts you and asks you to go through this process.'... type 'I am sure I know what I am doing' and click 'Enable dangerous mode.'"
> Obviously everyone should default to being a Sheep
This isn't actually that obvious, for a number of reasons.
The first is that it causes there to be more sheep. If you add friction to running your own software then fewer people start learning about it to begin with. Cynical cliches about the government wanting a stupid population aside, as a matter of policy that's bad. You don't want a default that erodes the inherent defenses of people to being victimized and forces them to rely on a corporate bureaucracy that doesn't always work. And it's not just bad because it makes people easier to scam. You don't want to be eroding your industrial base of nerds. They tend to be pretty important if you ever want anything new to be invented, or have to fight a war, or even just want to continue building bridges that don't fall down and planes that don't fall out of the sky.
Another major one is that it's massively anti-competitive. If the incumbents get a veto, guess what they're going to veto. This is, of course, the thing the incumbents are using the scams as an excuse to do on purpose. But destroying competition is also bad, even for sheep. Nobody benefits from an oligopoly except the incumbents.
And it's not just competition between platforms. Think about how "scratch that itch" apps get created: Some nerd writes the app and it has only one feature and is full of bugs, but they post it on the internet for other people to try. If trying it is easy, other people do, and then they get bug reports, other people contribute code, etc. Eventually it gets good enough that everyone, including the sheep, will want to use it, and by that point it might even be in the big app store. But if trying it is hard when it's still a pile of bugs and the original author isn't sure anybody else even wants to use it, then nobody else tries it and it never gets developed to the point that ordinary people can use it.
So maybe the scam we should most be worried about here is the one where scams are used as an excuse to justify making it hard for people to try new apps and competing app stores, and deal with the other scams in a different way. Like putting the people who commit fraud in prison.
No. This assumption is the core fault with the entire line of reasoning. The typical sheep will not do arbitrary things for a stranger such as sending you his entire bank account because you told him he needed to pay an IRS penalty in crypto to avoid being picked up by the state police who are already en route in 15 minutes.
It's a continuum. The question is how much of the low end needs to be protected by the system.
Binning into discreet blocks to match your example, the question is where to place the dividers between the three categories - nerd, sheep, and incompetent. We don't care to accommodate the third.
This is if nothing else an interesting postulate. Default all devices to nerd mode and sheep mode is an opt-in at setup time.
In theory I have no problem with the idea of hanging the incompetents out to dry, when I imagine them as unsympathetic idiots, the same people who litter, and can’t drive correctly. But actually I think most of us would be horrified when it turns out that category of incompetents includes our parents and grandparents, or, increasingly, our children (Gen Z has been increasingly falling victim to scams, partly because they have no idea how computers work since modern ones present only highly abstract surfaces to them, and I suspect Alpha will be the same).
> Default all devices to nerd mode and sheep mode is an opt-in at setup time.
The entire point here is that sheep do not need an overly protective mode. It's a false premise.
I know plenty of them. I help them navigate modern tech. I install fdroid on their phones. They lie on a continuum and none of them are going to turn on developer mode (or whatever BigTech wants to call it) because a stranger on the phone told them to.
There is a small sliver on the far end of the continuum that will do things like that. But in general they are sufficiently gullible that no measure that can be bypassed will ever work for them. They require a Fisher-Price device.
BigTech wants to hold that small sliver up as justification for their anticompetitive practices.
reply