@kareem_carr A lookup table of a+b is very useful, but doesn't constitute understanding.
If there are errors in the table you need to investigate the algorithm behind it and make fixes there.
@xriskology@davidmanheim@timnitGebru Are the notions of "effective altruism" and "longtermism" the same in your view? Can one support the one while discarding the other?
I can’t tell what’s more shocking so far about living in Germany 1. Uber doesn’t exist here or 2. Public transportation is so good that I don’t even notice
About the raging debate regarding the significance of recent progress in AI, it may be useful to (re)state a few obvious facts:
(0) there is no such thing as AGI. Reaching "Human Level AI" may be a useful goal, but even humans are specialized.
1/N
@Depl0reable@__mharrison__ `values` and `results` are changed to be numpy arrays instead of python lists. Essentially, this means that the computer can quickly access a specific index, say `values[10]`, since the precise location in RAM as well as its data type is known.
GPU Programming with Julia: Dr. Samuel Omlin and Dr. Tim Besard presented an 8.5 hour course on GPU #Programming with Julia at CSCS last year. This course is now available for free online https://t.co/whlsYv3Rke.
#julialang#GPU
@timnitGebru Can you provide backup for your claim that "[longtermism] has become one of the central research topics of the 'effective altruism' movement", especially in the form you describe it?
@Nokita50888531 @jdyeakel @karlrohe I find it difficult to come up with a kernel that would help us without using a lot of knowledge about the data (e.g. a learned kernel)...
Do you have anything particular in mind?
@CoryMSimon I assume this is pointer-comparison (z copies the pointer x) vs value-comparison. Therefore we would also expect @.time x === y to be much faster (O(1)) than @.time x == y?
@sp_monte_carlo Only seeing the left side I would expect it to be a sum, not a product.
Regular scalar product: <theta, k> =sum_i(theta_i * k_i)
This scalar product: <theta ^{, k}> = sum_i(theta_i^{k_i})
@karlrohe Agreed - you can find a vector for 0°, 10°, 20°, etc but what use is this if you cannot interpolate between them?
The problem is exaggerated even more if the "2" is drawn with a thinner line.