Y
HN Search
Hacker News Search
new
|
comments
|
top
|
jobs
alimw
searching PlanetScale…
1.
▲
2.
▲
3.
▲
4.
▲
5.
▲
6.
▲
40 ms
·
391.
▲
by
alimw
10y ago
Obviously it's self-interested for the species as a whole...
392.
▲
by
alimw
10y ago
This is just a lot of buzzwords strung together in no particular order
393.
▲
by
alimw
10y ago
There's a mechanics of research? Why is my supervisor hiding it from me?? It must be his plan to weaken and kill me...
394.
▲
by
alimw
10y ago
Humans probably have a monopoly on music-with-words for the foreseeable future...
395.
▲
by
alimw
10y ago
I'm no expert on ocean current modelling but it seems to me you should multiply the probabilities from each find, not add them up.
396.
▲
by
alimw
10y ago
Surely the value of this is that it provides a counterexample to a long-standing conjecture. The proof that it is a counterexample may not be too enlightening, but that's hardly the point.
397.
▲
by
alimw
10y ago
"But those huge powers you gave yourself, they still stay there, making all sorts of actions technically illegal. It's the state which decides when it wants to enforce them." It's not only to the state that the power acc
398.
▲
by
alimw
10y ago
That's exactly how I'd do it! What's supposed to be the better way? Modulo/division by three and five?
399.
▲
by
alimw
10y ago
Agreed it does seem unlikely that one could observe fundamental particles by looking inwards. Probably best to understand his use of language from particle physics as analogy. I don't suppose old Goenka had much of an education in mo
400.
▲
by
alimw
10y ago
Exactly what claim are you disputing about cloud chambers? According to Wikipedia they were extremely useful to the discoverers of the positron, the muon and the kaon.
401.
▲
by
alimw
10y ago
A corollary of Parkinson's Law is that, whatever language you start with, your neural net project will expand to fill the processing power available to it.
402.
▲
by
alimw
10y ago
I'll second this. I've plenty of work to do, but trying to do it from home has meant I'm not even on the road to recovery. If you're not too proud, you might be best off getting some part-time menial job on a temporary
403.
▲
by
alimw
11y ago
A link to Oxford isn't going to help house prices much, they are just as high there!
404.
▲
by
alimw
11y ago
Your conjecture as to the surface area of a hypersphere embedded in 4 dimensions (a "3-sphere") is not borne out :) The answer is in fact d^3 pi^2 / 4. See https://en.wikipedia.org/wiki/N-sphere#Volume_
405.
▲
by
alimw
11y ago
Or Windows. Anyway my point was, why force your readers to find and use a compiler when you could just give the answer below? Not sure why I got downvoted.
406.
▲
by
alimw
11y ago
> What does the C++ code print? If you guessed wrong, don't worry. You're in good company. If you guessed right, congrats! How do I knowwww????
407.
▲
by
alimw
11y ago
Is this your own insight or do you have references? It's a little hard for me to make sense of your position based only on what you've written here.
408.
▲
by
alimw
11y ago
You seem to be saying that meditation is not a reliable way to hang on to your ego. That should go without saying.
409.
▲
by
alimw
11y ago
"The United Kingdom responded with the most ambitious program of food aid and cash relief ever attempted by any nation to that point" "In fact, England exported far more food from Ireland than it provided in aid of all kinds&
410.
▲
by
alimw
11y ago
https://www.coursera.org/course/money https://www.coursera.org/course/money2 I took the first of these and intend to take the second, which I think came out as Coursera's top course at one po
411.
▲
by
alimw
11y ago
I think we agree :) Thanks for the reference btw, I'm supposed to be writing my thesis on such things.
412.
▲
by
alimw
11y ago
In the paper you reference, an 'optimal transport' problem does indeed arise in connection with Cournot-Nash equilibrium. However the naming is a coincidence, it is unrelated to the problem of finding the most efficient routing o
413.
▲
by
alimw
11y ago
Are you saying that an equilibrium is Cournot-Nash if it does manage to "simultaneously minimize both the expected commute cost for the group and for each individual"? Not sure that's right...