dcanelhas

6 ms·

31.
by dcanelhas 1y ago
It depends on how intelligence is defined. In the traditional AI sense it is usually "doing things that, when done by people, would be thought of as requiring intelligence". So you get things like planning, forecasting, interpreti
32.
by dcanelhas 1y ago
If you look at timescales large enough you will find that plenty of extinction level events actually do happen (the anthropocene is right here). We are living in a historically excepcional time of geological, environmental, ecological stabi
33.
by dcanelhas 1y ago
I think that makes sense and I can tell that you understand the distinction. Formally I believe they're different concepts and think it may cause confusion in some cases to use them interchangeably. The chaotic nature of a system is on
34.
by dcanelhas 1y ago
I just did a mental substitution for the word "stochastic" and moved on, since it wasn't the main point of the article. But you are correct. Computing the future states of a chaotic system, given the same initial conditions,
35.
by dcanelhas 1y ago
https://www.johndcook.com/blog/2023/08/12/fibonacci-lattice/ If you want a spiral that covers the sphere with evenly spaced samples, consider this approach.
36.
by dcanelhas 1y ago
It has been suggested [citation needed] that the optical illusions of movement caused by gradients are there to to compensate for the time it takes to process the visual input. This should let you have an understanding of what is going on i
37.
by dcanelhas 1y ago
Sometimes I like to imagine early people inventing, forgetting about and inventing the wheel again.
38.
by dcanelhas 1y ago
Thanks for sharing. I don't see why there is an atomic add in the kernel there. It doesn't look like two separate threads should be able to modify the same pixel, based on the block/thread indices?
39.
by dcanelhas 1y ago
Does molten glass solidified in contaminated salt water have good insulating properties?
40.
by dcanelhas 1y ago
I sincerely appreciate the reply, but are you talking about Moore's law? Alexnet could run on a commercially available GPU in 2011(?). But that wasn't the peak compute platform being used at the time for DL inference, so it distor
41.
by dcanelhas 2y ago
> Once the new datacenters are up and running, they’ll be able to train a model with 10^28 FLOP—a thousand times more than GPT-4. Is there some theoretical substance or empirical evidence to suggest that the story doesn't just end h
42.
by dcanelhas 2y ago
On the topic of coercing bits into functioning as data storage: harder drive ( http://tom7.org/harder/ )
43.
by dcanelhas 2y ago
In what way was it unreliable? And how does Chronoptics ToF sensor perform better?
44.
by dcanelhas 2y ago
You can still buy primesense devices afaik http://xtionprolive.com/asus-xtion-pro-live
45.
by dcanelhas 2y ago
Going to assume you're genuinely interested here: Intel's RealSense cameras typically also have an active IR illumination component that projects texture onto otherwise featureless surfaces where stereo would not give any measurem
46.
by dcanelhas 2y ago
Is this really the case? It's very uncommon for an IP to resolve to an actual user these days due to widespread use of Carrier-Grade Networks Address Translation (CGNAT). Assuming you had the public IP of an actual user though, how wo
47.
by dcanelhas 2y ago
Seems it can use both, if available.
48.
by dcanelhas 2y ago
Human activity (e.g. oil production, underground nuclear testing) does impact seismological events to some degree so butterfly effects are possibly there too. If he can leave a bunch of computers in the past and there is a strong enough inc
49.
by dcanelhas 2y ago
SLI being "scan-line interleaving" at the time. Not to be confused with today's "Serial Link Interface"
50.
by dcanelhas 2y ago
Mathematically, the camera most often used in computer vision is a pinhole camera. When we talk about "distortions" I think it's usually with regards to how the real device systematically deviates from that model. Calibration
51.
by dcanelhas 2y ago
This is a good point. If I would have grown up today I would possibly have done other things that aligned with the same kind of interests, and perhaps even have gotten further. Kids these days can, and do, build actual robots and UAVs with
52.
by dcanelhas 2y ago
As a 12 yo I learned about IRC from reading the Descent FAQ and tried setting up IHHD to play over a 14,4k modem. I made levels in DEVIL. Today I'm an engineer and game developer. Thanks for putting Descent out into the world. It had a
53.
by dcanelhas 2y ago
There is sublevel zero, also. It's a similar experience, but made for VR and has a procedurally generated map system that makes the playthroughs a bit different from time to time. Gabbagabbahey
54.
by dcanelhas 3y ago
If CPU/GPU execution speed is the goal while simultaneously code golfing the source size, https://halide-lang.org/ might have come in handy.
55.
by dcanelhas 3y ago
In case you're interested in these kinds of relativistic effects, there is a toolkit called OpenRelativity http://gamelab.mit.edu/research/openrelativity/ that you can use, for your simulated near-luminal sp
56.
by dcanelhas 3y ago
VFX was Matte shots and wire removal, according to the doc.
57.
by dcanelhas 3y ago
From my time hanging out with Arne on IRC I would assume he's using Photoshop (or opencanvas version oc11b72 ;-) )... I'd be surprised if he's on a PS version newer than CS2 though. I think Adobe added a subscription model to
58.
by dcanelhas 3y ago
If you read past the initial uninformed hot takes that thread actually gets progressively more interesting and informative. Good find.
59.
by dcanelhas 3y ago
Would that generate even more billions in revenue, though?
60.
by dcanelhas 3y ago
You can probably find allies among those on the campaign against "more optimal".

More ›