deadc0de

"Men are judged by utility."

I didn't realize this until recently. Rude awakening. Even the most close relationships turn out to be quite transactional.

The problem with services like Airalo is that they almost never give you a local phone number. In many countries getting a phone number involves showing a physical id.

Not really. Look at google maps sat view. These are two different projects.

Tesla/PG&E are also building 182.5MW/730MWh presumably at the same site? https://www.teslarati.com/tesla-megapack-moss-landing-instal...

Also ICE engines themselves are pretty heavy compared to electric motors and require heavy gearboxes and shafts.

Prop 13 distorts that a lot. If you bought a house 25 years ago and now are earning 400k, yeah, you're an upper middle class.

140k is believable, but I think they mean it's per person. If you rent, it's very unlikely you'd be able to buy a 1.2M house on 140k if you started with 0 in savings.

EDIT: To clarify my previous post - I was talking about the household income.

The definition of the middle class would surely depend on the cost of living in the particular area. You can make $30k and live like royalty in the 99% of the world.

You can go to a public school but if you want it to be a quality one your rent of a house will not be a 3k.

Not in SV. $400k gets you to the lower middle class. That is you probably have a tiny chance of buying some real estate in your lifetime. House and expenses related to maintaining it is a huge chunk of change.

There are a lot of nonlinearities. Taxes, healthcare, eligibility for the financial aid when the kids go to college, etc.

No, compressed oops mode doesn't change at runtime. That's pretty hard to pull off. You'd need to change object layout dynamically and unless you want to reformat the whole heap you'd need tags or regions with different pointer rules. Any of these techniques would have a very substantial runtime overhead.

AOT implementation in the JDK, that we call the "HotSpot AOT" implements tiered compilation (look for $JAVA_HOME/bin/jaotc). It is also based on Graal, but with a different set of plugins. You get reasonably good startup and re-JITing afterwards. Of course it doesn't give you compactness on the SVM, but it supports regular java code with class loading, reflection, etc.

A talk about HotSpot AOT: https://youtu.be/n5DCg6M2MDM

JVM is pretty much running half of the internet.

Here is a talk about it at JVMLS'17: https://youtu.be/n5DCg6M2MDM

Really? I fly about two times a year to HI - never happened. But anyways, my favorite airline, the companion ticket thing rocks (especially to HI)!

You're missing the point that there is a positive feedback loop at the source of the problem. The reason that women (or minorities for that matter) don't get into some fields of STEM is because there are no role models that tell them it's a good idea to do so. So there is a chicken and egg problem, and the only way do break the cycle is to give priority to such minorities. Is it fair to the members of the majority? Perhaps not. But in the end, I believe, it will justify the investment. Consider it a small price to pay for the millennia of oppression.

Being passionate and dare I say obsessive about you work is good. I fact, it's the only way to achieve good results. That's what separates hackers from people that just go to work.

And we've had vectorization years ago in the JVM.

Except 150k in Bay Area is nothing and you won't even make it into the middle class (i.e. buy a house).

Exactly. It's fairly entertaining to play with this model: http://www.nytimes.com/interactive/business/buy-rent-calcula...;

Basically almost anything in the bay area is not going to be profitable, in the south bay for sure.

It was not ok then, and it's even more not ok now, considering Chrome's market share.

I mean use the monopoly to take the tech, clone, tweak, and then don't contribute back. Thankfully google will keep it open source since they're making money from it quite indirectly. But still.. A "fork you" to apple that would make the web suffer. Douchy.

A Microsoft-worthy move..

In my case the effect of getting a PhD was extremely positive. I felt a huge difference in attitude at work to doctors (I worked for a couple of top big american software companies), basically allowing doctors to do whatever they liked as long as it's related to the product. And that's in the product group, I'm not even talking about the labs.

Not really. Allocation and deallocation for a system that does not support object movement (and hence compaction) is inherently non-trivial. In most modern GCs allocation is bump-the-pointer with TLABs, you can get cheaper than that. If the first generation GC is a copying GC you pay absolutely nothing for collecting a dead object. Compare that to explicit allocation/deallocation schemes.

The biggest problem with GC however is not the cost of allocation/deallocation but the fact that it's pretty hard to make GCs incremental at a very fine-grain level.

Java, btw, at least with Hotspot JVM, does escape analysis, that will allocate non-escaping objects on stack. Heck, these objects may even just end up in a register (with unused data members discarded).

Totally agree, man! Education is always the key.

This describes the state of GC like it was 10 years ago (and in a few places incorrectly), at least in Hotspot. Shameless promotion: try out the new G1 (-XX:+UseG1GC), better to use the latest EA JDK7 build.

I don't know... I would personally feel a bit weird asking, for example, an experienced compiler developer with a CS PhD to reverse a linked list, or words in a string, or something like that.

The Facebook puzzles are pretty easy. That being said, I hate puzzles too. I think it's pretty degrading to ask a well-established researcher/engineer to do that kind of stuff. On interviews I typically ask people to give a talk about their previous projects in detail. A lot of nuances arise during these talks that allow you to quite easily judge the level of the candidate.