dahart

> My understanding is that those who are, don’t use words as crude as “made” to describe a role in the making of a film.

Oh, no, film insiders use ‘made’ constantly, or even cruder - attribute credit without using a verb at all. Here’s a VFX person saying “Nolan’s CGI”: https://youtu.be/AFLcP-t0-kI (I’m ex-film industry, and still film industry tangential, and found this clip being discussed on a film industry mailing list.)

Film people are far more likely to pay attention to who the director is, and it’s very common for film people to talk about directors ‘making’ the film or otherwise casually assign ownership. I think this is especially true for famous directors and good films, and maybe a bit less true for unknown directors and mediocre films.

Nobody here advocated littering. What’s wrong with having trash cans, and why are you moralizing it?

Geometric is exponential; I goofed above. Geometric and exponential are sometimes used interchangeably, but sometimes people use geometric to refer to discrete intervals and exponential to refer to continuous functions.

No difference, you’re right. I meant polynomial not geometric. :P

That’s quadratic, not exponential. 2^n != n^2

I’m sure exponential is what parent meant, but geometric (edit: polynomial) is what actually happens, not exponential. Code complexity increases linearly when it doesn’t interact with other features, and geometrically when it does. In the worst case, complexity increases as the square of lines of code. That is still very, very bad, but it’s not exponential.

Re: throughput vs latency - what is the implication for having a signed floor divide sprinkled in with other math? Is Robison likely to be the better choice unless I’m doing divide operations on long vectors/arrays of numbers?

If the goal is wide vector divides, what about GPU? How does this algorithm compare on GPU? Are GPU uses less likely to rely on and benefit from throughput behavior if the vector of divides is spread across all threads?

I don’t do image gen regularly… so I’d love more context. I can see some less good features in the images but I’m not fluent with what today’s generators would do. I agree with GP that it would be quite useful to know about the prompts & iteration effort. If generation times changed, they could be relevant & helpful.

> good luck predicting where I’m going to have my next cuppa.

Why are you assuming there isn’t someone else trying to hack you at the next cafe? Actual damage can happen one person at a time, and can occasionally lead to larger scale compromises. Randomizing your location isn’t going to reduce your risk. Now that everyone who’s paying attention knows they can MITM your Mac, you shouldn’t trust any network.

Unfortunately, the goal with OpengGL is not to teach graphics and rendering concepts. Same goes for DirectX, Vulkan, WebGL, etc. These APIs are being designed primarily for performance, not for ease of learning. For teaching and learning, it’s a great idea to start with rasterization & ray tracing in software first! That can teach you the graphics concepts first, and also provide good motivation to learn a graphics API if you want it to go faster.

Vulkan, OpenGL and DirectX make GPU programming more difficult than it has to be for multiple reasons, some of which other people here have mentioned. They are also low level, focused on high performance, and are primarily designed and written for game engines. Making them easy to learn for newcomers is pretty low on the priority list. ShaderToy makes it easy by taking care of everything except the GLSL, you might like it. You can also try CUDA, which is generally a lot easier than Vulkan. For a really easy/fun time with CUDA, try Warp - it’s CUDA in Python.

Nope, they were just written in C. It was object oriented C, but without a C++ compiler. Do you know the story about the whole C++ toolset that was rolled out by R&D and promptly rejected by production? That happened before my time, but according to legend is what lead to PDI’s strict avoidance of C++… until a couple of years into the DWA merger. (Clearly there were some problems beyond C vs C++, but I don’t know the whole story.) Anyway, yeah I remember starting on the 64b transition, but left before that was finished.

Interesting, had never heard of Kono Express. It definitely looks a little ‘Unreal’, has video game vibes. The quality is pretty good at a glance, especially the environments and effects, but it doesn’t feel all the way there under scrutiny. This reminds me of all AI generated stuff - looks great from a distance but from up close you see glaring problems. The lipsync feels automated and robotic all the way through well past ep3, for example. Very little action in the whole series.

TBH I think it’s a big mistake to compare this to hand-animated Disney material. They are just very different things, and the debatable comparison will just cause people to focus on what’s wrong with Kono. If you think it’s good, just talk about the good qualities and share it. Also don’t forget the AI is trained on Disney material, and with heavy AI, it’s highly questionable whether you can credit one person for “cranking out stuff”.

What is the ink & paint comment about? The article doesn’t mention ink & paint…

Oh definitely - custom vs commercial. I remember the R&D team having to write a yearly report to justify the development expenditure, explain repeatedly why we were writing custom tools & renderer when these things were available off the shelf. I heard a few strong rants here and there about how the suits just didn’t understand. I remember the term “NextGen” being thrown around but I don’t remember if I lived through it or it was just getting started when I left (before S3). Jono I and tried for a while to stop having our shaders rewritten in C++ (without otherwise changing them nor adding any features). We eventually gave up and just kept building new things, but that was a fairly successful strategy to nudge the IP, tech, and support southward. Possibly the right strategy in retrospect, but felt a little rough at the time.

> The designers would fully texture out the inside of a dresser draw that will never be showing in the film. The idea was imagine how much attention to detail

Hehe, as an ex FX animator/TD, this sounds kinda funny to me, there’s likely more to the story. If they knew for sure the textures weren’t going to be shown, it’s a waste of time, not craftsmanship or attention to detail. But it’s plausible because often modeling & surfacing departments have to start work before the shots & camera are fully planned. And sometimes there are camera changes late in the production. It could be that someone’s shot that saw the inside of the dresser drawers got cut, that surfacing expected the drawers to be open in a shot, or that a scene inside the drawer was in storyboards but didn’t make it to layout, stuff like that happens a lot. But no productions are having artists work on stuff that they know will never be shown in the film.

“Better” is a loaded word. When you write your own renderer, you have a lot more control and a lot more support. Renderers at the time had not all converged on the same algorithms, so you could have a competitive advantage by writing a good one. PDI’s shader system was different than RenderMan’s and had some real advantages — compiled shaders, geometry & light shaders (not just materials), particles & particle shadowing, compositor integration, etc.. The renderer also needs to integrate into the pipeline deeply, so PDI’s renderer fits the PDI/DW pipeline better, whereas with RenderMan you have to try to mold the pipeline to fit the renderer in.

Hey I was on the PDI side as well, ahoy, nice to see you here! I had forgotten about Shark Tale using RenderMan, but I will say to the degree there was competition and to the degree that PDI won any pipeline battles, DWA eventually won the war. :P ;) There was plenty of cooperation as well, it just takes time for two companies to integrate, and plenty of people on both sides were quite attached to the cool tools they made. (I was certainly guilty!)

PDI’s toolset was pure magic for it’s time. When I joined, I was in heaven with how easy it was to make procedural animation, build new effects that crossed pipeline stages, render anything you could imagine… I wished I could share it with researchers and hobbyists.

Both companies each had their own renderer & rendering development team before the acquisition. And their own pipelines. A couple decades later, that’s still true.

> “Illegal” means a judge found you guilty.

No, “guilty” means a judge found you guilty. Illegal means you did something that’s not legal. People doing illegal things without getting caught or sentenced happens all the time. Speeding in your car is illegal, whether or not a judge finds you guilty.

> of course it can’t?

If obvious, can you prove it without reading Gödel’s proof or anyone else’s?

The important (2nd) part of Gödel’s proof is not that a model can’t contain all truth, it’s that a model cannot prove itself true, which is a far stronger statement than the intuition you describe. Of course, belief that your own intuition is independent of history is also highly questionable; you were raised and educated in an environment in which people were already aware of Gödel. It’s like saying the number zero is obvious (which it is now) despite the fact that it took thousands of years before it was used formally as a number. There was a time when it wasn’t obvious to the smartest people alive.

Note that formal axiomatic systems aren’t physical systems; what Gödel was talking about and what George Box was talking about are related but somewhat different concepts. It is obvious in physics that a parabolic ballistic trajectory model doesn’t account for air turbulence. It’s not at all obvious that all of math itself is guaranteed to produce contradictions and cannot prove itself, especially if you’re allowed to change math or use any system you want.

That’s not the meaning of “understand” that was being used. This article points out that nobody understands QM well enough, not even the experts. (Do note the author is an expert!) By “I Don’t Understand”, what he actually means is that QM is incomplete in the same way that Newtonian Mechanics is incomplete; it doesn’t explain physics well enough to understand or explain what’s really there. This is, of course, why there’s an ongoing debate among the experts about what QM even means and how to interpret it.

One problem with your analogy is that transformers do not even attempt to model the human brain, while Quantum Mechanics does attempt to model real physics.

> All models are wrong

That famous quote could use attribution: https://en.wikipedia.org/wiki/All_models_are_wrong

> If you make the error of believing that quantum mechanics is ontological god’s honest truth of nature

This is why science & experiment exists. The author holds no such belief, and the paper makes that very clear. That is what he actually meant by “I Don’t Understand Quantum Physics”. Jump to the very last section, “Is Quantum Theory Correct and Complete?” He’s pointing out the confusing and contradictory nature of the model so far, and saying that Quantum is incomplete and needs work. He is not making any category error.

You’re right, my definition sounds different. Your comment left me with the impression that you meant you weren’t talking to anyone. Meetings in my book are any scheduled discussions, regardless of how many people are involved. 1:1s are still meetings.

This depends entirely on how well the requirements have been communicated, right? If you could go for long periods without meetings on large projects, it means either you had great PMs or you’re very good at collecting relevant info and planning in advance. Or, it could mean that your target was quite flexible and open to interpretation. I don’t doubt it, and I’ve known some developers who could do what you did, deliver something great without that much back and forth. I’ve also known many developers (probably most TBH) who will focus on problems the customers don’t have and endlessly go down rabbit holes that have no bottom. Heads down software engineering for big-O and data structures seems to be a lot more fun for some people than fixing weird hard to repro bugs with vague repro instructions that customers complain about. Many very good devs still need to check in and iterate and get some steering, especially if the requirements have not been well or fully explained… which in my own experience has been more or less always. ;)

The false positive rate isn’t the problem, the false negative rate is, which is why I mentioned the false negative rate above, and not the false positive rate. Obviously it would result in fewer colonoscopies and that’s the problem. The whole reason people take a FIT test is to avoid getting a colonoscopy. False negatives means a larger percentage of people will have cancer and not be aware of it, and not schedule a more reliable test until the cancer develops further and gets caught later by some other test maybe. The FIT test is better than nothing, but worse than getting a colonoscopy in terms of safety and reliability. It’s popular because it’s convenient, not because it’s highly reliable.

The error rate could be higher, which I mentioned already, but who knows how much higher. I don’t see how speculating about the error rate helps anything. Did you mean to say the error rate is usually much lower than 11%?

> except for things that are life threatening for which no better test exists.

This was precisely why I chose the FIT test example. There is, in fact, a more reliable test than the FIT test. Several, it turns out. Reliability isn’t the reason for the FIT test’s popularity.

> therefore all tests with error rates are equally bad

I did not say that, nor even imply it.

Oh I don’t doubt there was funny business or fraud. What you’ve added is what the parent comment was missing; the reasons beyond the error rates.

> the test you Google is a false negative rate, which a simple fact that test can’t reliably diagnose everything even when run properly.

Yes I mentioned above it’s a false negative rate, right? And false negatives are potentially more dangerous for survival than false positives. Again, I was just reacting to the narrative about people getting scared by false positives, as if that was the primary problem. Your comment supports my claim that the error rate was not the primary problem, no?

I don’t know a thing about Holmes, I’m just reacting to your framing - don’t all medical tests that real people rely on have error rates? Doesn’t an 11% error rate suggest something that was real and mostly works but maybe wasn’t ready for prime time? The error rate could have been higher, perhaps. You only listed false positives, while the much more concerning problem would be false negatives - people who didn’t get scared or act when they should have. Still, a medical test error rate on its own doesn’t suggest deceit or fraud. People receiving terrifying but wrong results is a common problem that already exists in medicine.

Google tells me the FIT test for cancer has a false negative rate of between 8% and 26%. That appears to be likely worse than the Theranos numbers you cited here.

This has gotta be a reference to something, amirite? Do I need to watch the movie, or is it something from Nathan For You or somewhere else?

> you will never render something really big. And CPU look is much much better.

Why do you believe these things? In VFX and film, the GPU shaders and rendering is identical to CPU, they use the exact same programmable shaders on either side, so there is no difference in look or capability. Today’s GPUs have out of core streaming and caching and access to host memory. Pixar showed off their pixel-matching between GPU and CPU at Siggraph this year, and mentioned that today’s consumer GPUs are big enough to render IIRC ~98% of their scenes without much ‘out-of-core’ tech. Lots of film and VFX studios are using GPUs in production today to render pretty big shots. I think your story is a bit out of date; GPUs being size limited was somewhat true in the past, but is quickly going away. The CPU vs GPU look question has never been true for VFX or film, and was only true in games if you chose different rendering algorithms.