varelse

136 ms·

961.
by varelse 13y ago
I'm a fan of subsidizing green electricity generation (and I include nuclear power as green) and electric cars with a progressively increasing tax on gasoline. Most (not all of course) of this nation's towns have been engineered
962.
by varelse 13y ago
So how does one propose to get around negotiating land usage permission in all the podunk districts between SF and LA? That's what kept BART from circling the Bay 20 years ago and I suspect it has a great deal to do with the ridiculou
963.
by varelse 13y ago
Richard Matheson was the Donald Knuth of horror and SF TV tropes. A giant has fallen. As for _I am Legend_, _The Last Man on Earth_ was debatably the closest adaptation.
964.
by varelse 13y ago
I think this comparison says everything that needs to be said about Xeon Phi (for now): http://clbenchmark.com/compare.jsp?config_0=15887974&config_... Better luck next generation...
965.
by varelse 13y ago
So basically, whatever party is in power tends to get a pass for its bad behavior... SSND...
966.
by varelse 13y ago
Not even remotely the same beyond for NVIDIA besides upping the clocks. On NVIDIA GPUs, all the cores are active at all speeds. You're not going to see the equivalent of the case where 1 core at 2.6 GHZ beats 2 cores at 1.3 GHz. Do
967.
by varelse 13y ago
I'm not sure that Amazon is an outlier, but rather that as you say, other companies fear they would get raided if they tried to reinvest all their profits (what most companies probably ought to do a lot more of IMO) instead of issuing them
968.
by varelse 13y ago
What's to stop someone from forking Chain World onto another USB stick? In fact, adding that ability plus some sort of unpredictable way to reconcile the subsequent divergent timelines when two USB sticks met in the wild would make a far m
969.
by varelse 13y ago
When I was 30, I was making a tenth of what I'm making now and in the middle of destroying my career by exposing faked data in the wonderful world of science. You think Google is bad? Try getting threatened by a prominent member of the Na
970.
by varelse 13y ago
It's not quite like it from what I just read, it almost seems like it's halfway between Quantum Monte Carlo and a multi-armed bandit.
971.
by varelse 13y ago
I had a bad manager at google too thanks to the wonders of blind allocation. And I have vented about that experience several times on here. It was cathartic, and now it's out of me. But the price of doing so is that I'll probably never b
972.
by varelse 13y ago
Sometimes Google can be a lot like Heathers with Wynona Ryder's character played by Jon Heder...
973.
by varelse 13y ago
A really long time ago when I was in grad school playing with genetic algorithms and simulated annealing, I implemented something that seems awfully similar to quantum adiabatic annealing. It worked as follows: 1. Assign all states of each
974.
by varelse 13y ago
You're in the minority and I'd suspect your anomalous data is lost in the noise. While I think Google+ came out the gate as a budding Facebook killer, and given the initial internal alignment of the company tying everyone's yearly bonus t
975.
by varelse 13y ago
Ex-googler: While I think open allocation is toxic and the reason I left, the best way to get into one of these projects is to get Acquihired into it. That said, obviously, many of those there are more than willing to do the dull work in ex
976.
by varelse 13y ago
Heh, this is one of the main arguments I make in support of deterministic results - hundreds of developers are nearly guaranteed to introduce race conditions - and if your output isn't deterministic, how else can you detect them? I don't ev
977.
by varelse 13y ago
PS in my experience, the existing automated GPU tools for finding race conditions Heisenberg the process sufficiently that the race conditions stop occurring. So in practice, I end up running the code as usual, but writing all the intermedi
978.
by varelse 13y ago
I run the code multiple times on different GPUs of the same generation (not necessarily the same model) in different physical machines and get the same bit-accurate output. You're right, technically, there could be a race condition somew
979.
by varelse 13y ago
I use deterministic results as my #1 debugging tool. I do this by building queue-based concurrent systems with a resulting nondeterministic order of execution but I then insure that the results themselves are deterministic. Consider it to b
980.
by varelse 14y ago
I post corrected! Still, relating this to storytelling, I continue to believe that the single most terrifying moment in a video game, to this day, occurs in _Rescue at Fractalus_. Everyone I know has literally fallen backwards in their cha
981.
by varelse 14y ago
All that and no mention of _Rescue at Fractalus_, _The Eidolon_ (An 8-bit FPS 11 years before Doom), _Koronis Rift_, or _Ballblazer_? I'm getting old...
982.
by varelse 14y ago
So what this implies to me is that the first year at Google is an additional hiring filter to weed out bad cultural fits who got past the previous hiring screen (which selects for walking encyclopedias of Comp Sci and recent Top Coder probl
983.
by varelse 14y ago
Work was assigned to me the first day and I'm a very self-motivated independent worker when given a task. That trait had served me well right up to joining Google whereupon it flamed out because I probably put too much time into learning
984.
by varelse 14y ago
If you're asking whether I was productive, yes, I was, just not sufficiently so apparently. My failing was explained to me that since I was a "Senior Engineer(tm)" I was supposed to arrive at google an expert in all their technologies so I
985.
by varelse 14y ago
I was a Senior Engineer with a manager who talked to me once on the day I joined the team, then forgot I existed for the next 3 months. If I had discreetly transferred without trying to repair this, I think I might have pulled it off. But
986.
by varelse 14y ago
"At Google, individual engineers are far more empowered than that. There is a strong internal process for simply changing projects." That assumes the Hogwarts Hat of random allocation doesn't land you with a stinker from the get-go. And I
987.
by varelse 14y ago
Let me see if I understood what you wrote... Summarizing: 1. I don't understand bitcoin mining but that's no obstacle to rendering a negative judgment on a comparison between building bitcoin mining farms and building an HFT firm. 2. The ma
988.
by varelse 14y ago
What perplexes me here is the enthusiastic response received for the development and deployment of bitcoin mining hardware (for which the motive is giving a hardware edge to people in order to make lots and lots of money) versus the condemn
989.
by varelse 14y ago
So if I read this correctly, you're effectively trying to solve the hybrid computing problem that everyone else is working on too (and the results so far are pretty disappointing IMO). To which I have to respond that better is often the ene
990.
by varelse 14y ago
Battery life alone should be reason enough to expose OpenCL on mobile...

More ›