Hacker News: nicebyte

New comment by nicebyte in "Waymo Premier"

nicebyte — Thu, 11 Jun 2026 17:14:41 +0000

If done right, this is more like a monthly bus pass

New comment by nicebyte in "Apple WWDC 2026"

nicebyte — Mon, 08 Jun 2026 19:40:08 +0000

when you re-crop a photo or use the perspective tool, you are literally distorting the image. not to mention, all of the processing that happens in modern cameras before you even see the image. in the case of a modern smartphone in particular, I think it's fair to say that you never really see the "real" image as captured by the CCD sensor, and even if you did you would not like it anyway.

New comment by nicebyte in "C++: The Documentary"

nicebyte — Fri, 05 Jun 2026 22:02:27 +0000

they're technically not wrong. C is literally an "abstraction" of the machine. As we know, the whole point of an abstraction is to ignore the multitude of details :-)

New comment by nicebyte in "Uber's $1,500/month AI limit is a useful signal for AI tool pricing"

nicebyte — Thu, 04 Jun 2026 18:15:33 +0000

250k for base pay is about in line with median I'd say.

If 250k was the total comp (taking into account bonus/stocks/what have you) then yeah, you definitely should have negotiated.

New comment by nicebyte in "Investigating how prompt politeness affects LLM accuracy (2025)"

nicebyte — Thu, 28 May 2026 18:05:54 +0000

"outgroup"? what outgroup? we're talking about inanimate objects here. by your own logic, you treat your home appliances as an outgroup so you must be secretly a dangerous psychopath. or do you thank your microwave after heating leftovers?

New comment by nicebyte in "Is my blue your blue?"

nicebyte — Mon, 27 Apr 2026 21:38:50 +0000

it's a neat experiment but I think it's ultimately flawed because color is usually perceived in context, and depending on context I could easily see anyone reinterpreting the hues they labeled "green" in this test as blue, and vice versa.

EDIT: in general, blue is a pretty fascinating color. yes, many cultures have a somewhat blurry distinction between blue and green. Some others seem to differentiate shades of blue that others don't (i.e. in Russian "голубой" and "синий" refer to distinct colors but in English those would be just shades of blue). I guess there's something about photons in that energy band that messes with perception. https://en.wikipedia.org/wiki/Non-photo_blue

New comment by nicebyte in "If more than 50% press blue, everyone survives. Red pressers always survive"

nicebyte — Sun, 26 Apr 2026 21:07:19 +0000

> There is no consequence for choosing blue

there are consequences in both cases.

Introduction to Spherical Harmonics for Graphics Programmers

nicebyte — Tue, 14 Apr 2026 05:16:18 +0000

Article URL: https://gpfault.net/posts/sph.html

Comments URL: https://news.ycombinator.com/item?id=47761517

Points: 4

# Comments: 0

New comment by nicebyte in "An autopsy of AI-generated 3D slop"

nicebyte — Thu, 26 Feb 2026 03:49:38 +0000

I've found Trellis specifically to be very "over-promise and under-deliver".

Nothing i tried with it got even close to th level of quality that they were advertising - felt like a bunch of examples were hand-picked, at best.

New comment by nicebyte in "We tasked Opus 4.6 using agent teams to build a C Compiler"

nicebyte — Tue, 10 Feb 2026 21:53:05 +0000

jumps is another one. jmp can have many encodings depending on where the target offset you're jumping to is. but often times, the offset is not yet known when you first encounter the jump insn and have to assemble it.

New comment by nicebyte in "Simplifying Vulkan one subsystem at a time"

nicebyte — Tue, 10 Feb 2026 21:50:51 +0000

> I want a single-line malloc with zero care about usage flags and which only produces one single pointer value

That's not realistic on non-UMA systems. I doubt you want to go over PCIe every time you sample a texture, so the allocator has to know what you're allocating memory _for_. Even with CUDA you have to do that.

And even with unified memory, only the implementation knows exactly how much space is needed for a texture with a given format and configuration (e.g. due to different alignment requirements and such). "just" malloc-ing gpu memory sounds nice and would be nice, but given many vendors and many devices the complexity becomes irreducible. If your only use case is compute on nvidia chips, you shouldn't be using vulkan in the first place.

New comment by nicebyte in "Simplifying Vulkan one subsystem at a time"

nicebyte — Tue, 10 Feb 2026 21:29:45 +0000

> Not sure if this is an "oh, no" event.

it's not.

descriptor sets are realistically never getting deprecated. old code doesn't have to be rewritten if it works. there's no point.

if you're doing bindless (which you most certainly arent if you're still stuck with descriptor sets) this offers a better way of handling that.

if you care to upgrade your descriptor set based path to use heaps, this extension offers a very nice pathway to doing so _without having to even recompile shaders_.

for new/future code, this is a solid improvement.

if you're happy where you are with your renderer, there isn't a need to do anything.

New comment by nicebyte in "Simplifying Vulkan one subsystem at a time"

nicebyte — Tue, 10 Feb 2026 21:22:08 +0000

just use the vma library. the low level memory allocation interface is for those who care to have precise control over allocations. vma has shipped in production software and is a safe choice for those who want to "just allocate memory".

New comment by nicebyte in "Simplifying Vulkan one subsystem at a time"

nicebyte — Tue, 10 Feb 2026 21:09:44 +0000

some of this is what's khronos standards are theoretically supposed to achieve.

surprise, it's very difficult to do across many hw vendors and classes of devices. it's not a coincidence that metal is much easier to program for.

maybe consider joining khronos since you apparently know exactly how to achieve this very simple goal...

New comment by nicebyte in "We tasked Opus 4.6 using agent teams to build a C Compiler"

nicebyte — Fri, 06 Feb 2026 19:26:49 +0000

assembler is far from trivial at least for x86 where there are many possible encodings for a given instruction. emitting the most optimal encoding that does the correct thing depends on surrounding context, and you'd have to do multiple passes over the input.

New comment by nicebyte in "Huge Binaries"

nicebyte — Mon, 29 Dec 2025 18:30:09 +0000

shameless plug: if you want to understand the content of this post better, first read the first half of my article on jumps [1] (up to syscall). goes into detail about relocations and position-independent code.

[1] https://gpfault.net/posts/asm-tut-4.html

New comment by nicebyte in "Speedrunners are vulnerability researchers, they just don't know it yet"

nicebyte — Mon, 03 Mar 2025 17:38:49 +0000

including ai generated illustrations in your articles or presentations is very cringe

New comment by nicebyte in "Suckless.org: software that sucks less"

nicebyte — Fri, 21 Feb 2025 21:26:54 +0000

my comment isn't about things being written using c/make/whatever, it's precisely about the faulty assumption that complexity is needless.

New comment by nicebyte in "Suckless.org: software that sucks less"

nicebyte — Fri, 21 Feb 2025 21:04:28 +0000

yeah no. I've mainlined dwm + dmenu all the way back in 200x, I've written tons of makefiles and have the scars to prove it.

These days I'm off of this minimalism crap. it looks good on paper, but never survives collision with reality [1] (funny that this post is on hn front-page today as well!).

[1] http://johnsalvatier.org/blog/2017/reality-has-a-surprising-...

New comment by nicebyte in "Johnny.Decimal – A system to organise your life"

nicebyte — Fri, 21 Feb 2025 18:16:44 +0000

I bet 90% of the reason this is on the front page is the Berkeley mono font. the system itself sucks.