pca006132

Mesh is also a boundary representation. I think you meam NURBS.

LLM is just a tool. How the tool is used is also an important question. People vibe code these days, sometimes without proper review, but do you want them to vibe code a nuclear reactor controller without reviewing the code?

In principle we can just let anyone use LLM for medical advice provided that they should know LLMs are not reliable. But LLMs are engineered to sound reliable, and people often just believe its output. And cases showed that this can have severe consequences...

but how can you prevent the user from modifying the kernel?

By vendoring the code in, in this case I mean copying the related code into the project. You don't review everything. It is a bad way to deal with dependencies, but it feels similar to how people are using LLMs now for utility functions.

The problem is that it is natural to have code that is unreachable. Maybe you are trying to defend against potential cases that may be there in the future (e.g., things that are yet implemented), or algorithms written in a general way but are only used in a specific way. 100% test coverage requires removing these, and can hurt future development.

But this doesn't solve dependency hell. If the functionalities were loosely coupled, you can already vendor the code in and manually review them. If they are not, say it is a db, you still have to depend on that?

Or maybe you can use AI to vendor dependencies, review existing dependencies and updates. Never tried that, maybe that is better than the current approach, which is just trusting the upstream most of the time until something breaks.

Question: How many LoC do you let the AI write for each iteration? And do you review that? It sounds like you are letting it run off leash.

> as tasks that junior developers might perform don't match your skills, and are thus boring.

Yeah this sounds interesting, and matches my experience a bit. I was trying out AI for the Christmas cuz people I know are talking about it. I asked it to implement something (refactoring for better performance) that I think should be simple, it did that and looks amazing, all tests passed too! When I look into the implementation, AI got the shape right, but the internals were more complicated than needed and were wrong. Nonetheless it got me started into fixing things, and it got fixed quite quickly.

The performance of the model in this case is not great, perhaps it is also because I am new to this and don't know how to prompt it properly. But at least it is interesting.

> The idea of bitwise reproducibility for floating point computations is completely laughable in any part of the DL landscape. Meanwhile in just about every other area that uses fp computation it's been the defacto standard for decades.

It is quite annoying when you do parallelization, and idk if that many people cared about bitwise reproducibility, especially when it requires compromising a bit of performance.

Wonder if someone used effect handlers for error logging. Sounds like a natural and modular way of handling this problem.

It has many language bindings, including python and js. Though the js backend is not parallel because it uses wasm, and we had problem with mimalloc memory usage with pthread enabled.

At least 1 would not be enough. So how many branches are enough? And what about people with less money and time available?

But this is not related. You still have to pay the APC.

I thought this means for category theory people

anyway, quite cute :)

I remember people saying that chromium is better at sandboxing than firefox, so more secure.

If what they did is never revealed to someone else, what is the problem here? It is not like we have no way to hide stuff without cryptography, and people are not advocating for police to search every apartment once in a while to look for illegal stuff.

Authorities cannot tap into your brain, cannot tap into physical face-to-face conversations, and people can plan out crimes using these means. It is not like there is no way to hide stuff before the born of modern cryptography.

And who want everything to be open and transparent? I am not aware of anyone who wants this.

What I miss from vscode is the remote functionality, can you do it with emacs? For neovim there is distant.nvim, but idk if it is mature enough and configuration seems a bit annoying...

You can static link.

C/C++ dependency management is easy on windows? Seriously? What software did you build from source there?

I feel like these are stuff that the C-suite needs for justifying their pay. If it is "boring browser development", it will show that they are doing nothing, redundant, and cannot have bonuses and salary raise.

The nice thing is that a fully broken clock is accurate more often than a slightly deviated clock.

I am curious how the last algorithm is an order of magnitude faster than the one based on sorting. There is no benchmark data, and ideally there should be data for different mesh sizes, as that affects the timing a lot (cache vs RAM).

I work on https://github.com/elalish/manifold which works with triangular meshes, and one of the slowest operations we currently have is halfedge pairing, I am interested in making it faster. We are already using parallel merge sort for the stable sort, switching to parallel radix sort which works well on random distribution is not helping and I think we are currently bandwidth bound. If building an edge list for each vertex can improve cache locality and reduce bandwidth, that will be very interesting.

Yeah, in that case I guess it is usually memory-bound? When it is memory-bound you don't really care that much about scheduling etc.

They won't use it intentionally, but it is not like it can never happen. E.g., https://reproducible.nixos.org/nixos-iso-minimal-r13y/ is still only 96.59% reproducible.

While I think the OP did not mean the compilation process is nondeterministic, I won't be surprised if it is actually non-deterministic. A lot of algorithms and data structures rely on nondeterminism for performance or security (by default). It is too easy to introduce nondeterminism accidentally, and it is tempting to use that to speed up algorithms. Also, if it relies on floating point, results on different machines and environments may be different (depending on libm and hardware implementation), which is, in some sense, nondeterministic.

While it is decidable, people typically never produce optimal programs even for the hot path. It is just intractable and too slow to do right now.

For register allocation and instruction selection, there is hope because it is FPT and there are algorithms to do it optimally in polynomial time, albeit with a large constant factor (FPT), making it impractical to apply to compilers as of today. For instruction scheduling, it is just too hard. If you read literature on scheduling algorithms, it is NP-hard for apparently simple instances, e.g., 2 parallel identical machines with no preemption and bounding completion time (https://www2.informatik.uni-osnabrueck.de/knust/class/), while actual microarchitecture is much more complicated than this...

Needless to say, these are already the simpler problems. The longer the program or the more profiling data you can optimize for, the more tricks you can throw at it, and most of them are NP-hard to optimize optimally.

Being NP-hard doesn't imply that you can't obtain the optimal result, but compilers that I know of do not implement them, because most users are not willing to wait for days for such a compilation to complete. Ideally, one should make something that can run on clusters of CPUs or GPUs to optimize this, and people having those clusters will typically be willing to do this because they want to optimize the program they later run on the clusters. However, to my knowledge, no one is working on this at the moment.

Will they really get paid less? The feeling I have now is that people are paid a lot not because of what they do, but because of the potential damage they can do in case they fucked up. E.g. CEOs, lawyers, etc. Moving some of the work to AI doesn't reduce the risk, so they should have the same pay in my mental model.

Plus C-level executives typically don't lower their pay, and IMO investors apparently don't care that much about their pay, I can't see a reason why their pay will be reduced (significantly).

Not an artist myself. I think some artists may become more like head chefs in some Chinese restaurant, who is more like QA and give direction to cooks to improve their work. I think it is hard to notice the details and give concrete feedback if you are not working on it professionally for a long time.

Yeah, I remember reading someone saying you won't use a fork lift in a gym. I think this is the same idea.

The problem is really about how to evaluate performance or incentivize students to actually work on their exercise.