HN user

rickard

96 karma

meet.hn/city/59.2664372,18.1330727/Skarpnäck

Posts0
Comments22
View on HN
No posts found.
CVE program faces swift end after DHS fails to renew contract [updated] 1 year ago

I agree that it’s not zero, but according to CDC, the US sees about 1.35 million cases per year in a population of about 346 million, which is about 390 cases per 100,000 people. Your figure for the EU over a population of 447 million in 2022 gives 14.5 cases per 100,000 people, or more than a factor of 26 less.

Being 26 times less worried about something translates, at least for most things, for me, to not being worried about it any more.

0ThreadHN
Asking robots to design stained glass windows 4 years ago

Not DALL-E 2, but Disco Diffusion (https://colab.research.google.com/github/alembics/disco-diff...):

* https://jossi.avkrok.net/Escher_0.png

* https://jossi.avkrok.net/Escher_1.png

* https://jossi.avkrok.net/Escher_2.png

The prompt was "A stained glass image of M.C. Escher working in his studio in non-Euclidean space", 600 iterations, 1024x768, and these are the first three ones it came up with. It's not perfect, but it definitely knows who he was!

0ThreadHN
Julia: Faster than Fortran, cleaner than Numpy 5 years ago

Yep, completely agree. For the project I'm currently doing, it seems like a fairly good fit though. Lots of prototyping different approximations, and needs to be faster than plain numpy.

Also, the jitclass things help somewhat. I use them as plain data containers, to work around the hideously long argument lists that otherwise would be required, but with no methods. jitclass breaks the GPU option, though.

0ThreadHN
Julia: Faster than Fortran, cleaner than Numpy 5 years ago

I recently (last week) started using numba, for similar reasons to why the author seems to like Julia. I tested translating his example to numba:

  @numba.njit(parallel=True, fastmath=True)
  def w(M, a):
      n = len(a)
      for i in numba.prange(n):
          for j in range(n):
              M[i,j] = np.exp(1j*k * np.sqrt(a[i]**2 + a[j]**2))
and timed it like this:
  %%timeit
  n = len(a)
  M = np.zeros((n,n), dtype=complex)
  w(M, a)
On my 8-core system, this ends up more than 10x as fast as the numpy version he listed (which seems to lack the sqrt, though), which would place it close to the multithreaded Julia, even considering that ran it on a 4-core system. As an added bonus, it can also pretty much automatically translate to GPU as well.
0ThreadHN
Sweden's unusual response to coronavirus 6 years ago

The number of people in intensive care is most certainly not growing exponentially. Please see the official statistics on https://portal.icuregswe.org/siri/report/vtfstart-corona . The number of patients admitted to intensive care daily has gone from 43 on the 23rd of March to 10 today.

The cumulative number of patients admitted to intensive care is 278, and since we had 510 intensive care before the start of the pandemic, that is not a problem at all.

Unless you have reason to believe that the ratio of patients requiring intensive care somehow has decreased radically during the course of the pandemic, the growth of cases is most certainly not exponential.

0ThreadHN
What’s Really Killing Digital Health Startups 11 years ago

I don't quite agree with your complaints.

Regarding #2, what do you consider the problem with representing arbitrary precision decimals as numbers? That your javascript json parser converts json numbers to 64-bit floats? I'm not sure that that is really a problem with FHIR - see e.g. https://www.npmjs.com/package/json-bigint . Or is the problem that exponents aren't allowed?

About #3, how would you standardise addresses? ISO has been at it for several years and still not produced a standard. See e.g. http://stackoverflow.com/questions/4840928/iso-standard-stre... for relevant links.

Your fourth complaint is plain invalid. A period consists of two dateTimes, both of which specify time zones.

Number one and five I can somewhat sympathise with, though.

0ThreadHN
Boot a linux kernel right inside your browser. 15 years ago

1): The disk doesn't seem to be emulated; it's just a rootfs in RAM.

2): See another thread here.

3): Check out cpux86.js. In the start() function at the very end, the following section might be enlighting (even though it is a bit obfuscated by a javascript compressor):

  function start() {
  [...]
    If=32*1024*1024;
    ya.phys_mem_resize(If);
    ya.load_binary("vmlinux26.bin",0x00100000);
    Jf=ya.load_binary("root.bin",0x00400000);
    start=0x10000;
    ya.load_binary("linuxstart.bin",start);
    ya.eip=start;
The files vmlinux26.bin, root.bin and linuxstart.bin are fetched from the server.
0ThreadHN
Boot a linux kernel right inside your browser. 15 years ago

I guess a tun/tap to some server-side service might be possible too, but perhaps not very interesting. The x86 dynamic library api suggested in the tech notes sounds more useful.

Edit: Easier idea might be adding ttyS1 and connecting that to a websocket on the server. Combined with something like socat or ppp, that ought to work. Since the console is already ttyS0 and connected to the js terminal, it might be doable even with the obfuscated source.

0ThreadHN