Google "mitigation of damages". You're supposed to stop the violation as soon as it has caused you harm. A party can't wait for greater harm to come to them in hopes of a bigger payout. That's not what the courts are for.

Fear is mostly an electrochemical process in a biological system. To call it silly is pretty dismissive of its role in our evolution. Now how supposedly rational, logical, thinking beings react to their fear is worthy of consideration.

The difficulty in replacing it is of course a factor. People could buy a CPU+memory package as an upgrade for a motherboard, though, and still use card-slot peripherals like extra network cards and such. Many motherboards even in the enthusiast space now have networking, sound, lots of USB, and lots of SATA onboard.

I'd pay a few hundred dollars for a mid-line CPU with mid-high GPU and 16 or 32 GB of ultrafast memory on a card that mounts onto the motherboard in a great big socket. Heck, that motherboard could then have DIMM slots for a few GB of slower DDR3 just as a RAM cache for the drives. Better yet for a server would be 32GB of HBM with 128GB of DDR3 or DDR4 swap space.

This all assumes HP or IBM doesn't make it to market with memristors / big racetrack memory that's much faster still. Any of that, too, though, would benefit from being more tightly coupled to the CPU.

People aren't so much scared of what could happen as they are about inability to do anything about it.

Being careful around traffic makes getting hit by a big truck less likely. Diet and exercise are not panaceas but are mitigating factors for lots of medical conditions even if you're genetically or epigenetically predisposed to them. Modern medicine, although imperfect, gives us far greater control over both of those types of things even after the fact.

In a plane crash, unless you're the pilot or mechanic, there's not much you can do. In an asteroid strike, unless it's a smaller one we know is coming and can evacuate people, there's not much anyone can do. With a disease with a high infection rate, high mortality rate and no known effective treatment, such as Ebola until recently was, there's not much that can be done other than avoiding high-risk areas and quarantining people in those high-risk areas.

Regardless of the odds, people fear loss of control sometimes as much or more than the actual negative impacts. Some competitive team sports athletes have a more difficult time watching their team from the sidelines after an injury than they have emotional difficulty in their recovery.

They could use separate heatsinks or mill a recess into a bigger one for the memory.

Now, imagine if the whole APU could use HBM as the combined system RAM.

What little part did you grab onto without reading the whole post? That's right, the very first sentence.

Read the whole thing. It's about balancing speed of development with other factors, including performance of the code.

Nobody's talking about blazing fast, careless typing into an IDE. The biggest cost in software development tends to be the software developer. Making the software developer more productive saves money. Most of us don't have unlimited budgets.

People who code in Python, Perl, Scheme, JavaScript, or whatever rather than assembly aren't all "sloppy" developers. They're producing code that does a lot more work per source line than Assembly does. If they produce the same number of lines, then they're more productive.

Assembly can call C libraries on most systems. Then again, so can Perl, Python, and Scheme.

The best reason for avoiding assembly is that it's almost never the fastest way to get the programming done, and with a good C compiler it may not be the fastest way for the program to run, either. Assembly is great for bootstrapping a compiler or OS and inside tight loops. It can be handy if your hardware has an advanced feature not yet exposed through a library for a higher-level language. Not everything needs it, though.

Even C or C++ is often not the fastest way to get a project done. Sometimes there's a pretty big runtime penalty for using something that enables faster development, though.

There are lots of cases out there of Lua, Python, or some other language being used to wrap around a small core of C or C++ with maybe the most time-critical parts in assembly. Sometimes a higher-level language is the right tool for the situation, but sometimes not.

In the Midwest it's common for a city or town to draw from the nearby major rivers like the Mississippi, Missouri, Illinois, Des Moines, and Ohio then treat that water for the tap. Then they take the sewage, separate out the solids, treat the liquids, and release that downstream.

I'm not sure I'd bitch so much about drinking what my own city or county was purifying on both ends. In the Midwest people are drinking what's been treated for drinking by their city, town, county, or water district but which was treated as wastewater by whoever was upstream. In the spring, sometimes the sewage treatment plants upstream flood. (A few cities and towns even continue to get fines from the EPA for their stormwater and sewage drains combined, so that flash flooding brings sewage up into their own streets.)

It's worked for decades elsewhere to re-treat wastewater as drinking water. California's supposed to be the progressive leader on this sort of thing. It's time they caught up.

This is what the attackers want. They want to erode the freedoms of Western secular societies. The Charlie Hebdo attackers have won.

Couldn't it be that those few are just very far to the right of a normal bell curve or a power curve rather than on an inverted bell?

I think you're conflating "talent" and "higher level of complex intellectual ability", AKA "intelligence". Logic, syntax, and calling conventions can be taught. Some of the things that make a programmer vastly more efficient are specific portions of intelligence: pattern recognition, long-term memory, holding a large working set in short-term memory, juxtaposition of one situation against another, and ability to abstract generalities.

This sounds an awful lot like one part of the group has been promoted to this level of schooling without strong fundamentals important to CS. There's little reason to believe there are people who just "get it" vs. people who don't. There's plenty of reason to believe that some of them have a background that prepared them to understand the material while some haven't.

As a hybrid solution, take shares quarterly along with your pay, and no BS about having to stay there through a layoff to get them vested.