[GUEST ACCESS MODE: Data is scrambled or limited to provide examples. Make requests using your API key to unlock full data. Check https://lunarcrush.ai/auth for authentication information.]

@SebAaltonen "AI writing idiomatic C++ as always. Uses a lambda in low level code. Lambda does just a BSF/CTZ + bitmask. The lambda is used in one place so it doesn't even reduce repetition. When I generate pure C with AI there's no problems like this"
X Link @SebAaltonen 2025-10-27T13:40Z 46.1K followers, 13.4K engagements

"I just realized that I didn't actually want an iPhone Mini. I wanted a more lightweight phone. Tested iPhone Air in the local store and it feels awesome in my hand. 120Hz display too (which is a must have for me). Seems like a perfect phone for me. Managed to resist the urge :)"
X Link @SebAaltonen 2025-10-20T19:03Z 46.1K followers, 13.5K engagements

"iPhone Air has almost 2x screen area compared to iPhone X (8 years old flagship) but is only XX% heavier. They also crammed in a XX% bigger battery and a 4x brighter screen at 2x refresh rate. That's progress"
X Link @SebAaltonen 2025-10-21T19:09Z 46.1K followers, 13.2K engagements

"Made a X week refactoring to port our UI rendering to use our new RHI. Was using GLES2-style wrapper before. UI rendering still is using a global RenderPass object. Using Codex (AI) to refactor away the global. I instructed it to pass the RenderPass object around. Example:"
X Link @SebAaltonen 2025-10-29T08:40Z 46.1K followers, 5536 engagements

"2x2x2 = uint8. That's one byte. Link = uint32 = X bytes. Total = X bytes. 4x4x4 = uint64. That's X bytes. Total (with link) = XX bytes. 4x4x4 tree is half as deep as 2x2x2 tree. You get up/down twice as fast"
X Link @SebAaltonen 2025-10-18T19:09Z 46.1K followers, 4195 engagements

"Voxel mask (in non-leaf nodes) tells us which children are present. You can do a popcnt (full rate instruction on GPU) to count the children. Children are allocated next to each other. Children sub-address can be calculated with binary prefix sum (= AND prev bits + popcnt)"
X Link @SebAaltonen 2025-10-18T19:12Z 46.1K followers, 3948 engagements

"Why not 8x8x8 or 16x16x16 trees If you already allocate children next to each other 4x4x4 is good enough as it puts children in the same cache line implicitly. 4x4x4 fits nicely in a single uint64 so it's efficient to load and process in a single register. 8x8x8 = XXX bits"
X Link @SebAaltonen 2025-10-18T19:33Z 46.1K followers, 2293 engagements

"Since voxel octrees often encode just the surface a more narrow region is more storage efficient. Bigger nodes increase the ratio of payload:metadata but the ratio is already pretty good in 4x4x4 tree. Most nodes are leaf nodes and leaf nodes don't need the link"
X Link @SebAaltonen 2025-10-18T19:36Z 46.1K followers, 2865 engagements

"Seems that MacOS XX broke screenshot pasting to Slack. Previously you could see the screenshot directly in Slack as an image now you get a PNG file link instead (no visible image have to download to see it). Anybody figured out how to workaround"
X Link @SebAaltonen 2025-10-20T08:16Z 46.1K followers, 7575 engagements

"@atomlib OK. Let's compare iPhone Air to the iPhone X then. X years difference. iPhone Air has almost 2x screen area. 4x nits and 2x refresh rate. XX% weight difference"
X Link @SebAaltonen 2025-10-21T18:57Z 46.1K followers, XXX engagements

"This is the thing with AI coding. It's fine for unimportant leaf code but when you ask AI to do crucial performance critical inner loops it will not do what you want. I asked it to do sparse SDF brick tracer too and it had horrible SIMD loop/branch coherency perf issues"
X Link @SebAaltonen 2025-10-25T21:02Z 46.1K followers, 4914 engagements

"Second day WebGPU hobby proto: Asked AI to improve the terrain generator + add exponential fog. Now it starts to look a bit like a real terrain. I am writing a voxel renderer. Used AI to write a quick hmap generator + triangle raster comparison"
X Link @SebAaltonen 2025-10-26T21:03Z 46.1K followers, 16.5K engagements

"@therealdecross Genelec speakers have won countless of audio magazine tests and they have the DAC inside them. Digital signal to speakers. No inference. Use a top quality DAC manufactured by a real audio company instead of a phone company"
X Link @SebAaltonen 2025-10-28T10:54Z 46.1K followers, XXX engagements

"@treylorswift I never trust AI. I am using GPT5-Codex with maximum effort settings. This is basically a glorified search and replace operation. Using AI for it because it would take me couple of days to do because it's not a direct search and replace"
X Link @SebAaltonen 2025-10-28T14:40Z 46.1K followers, XXX engagements

"@Bearspacegoat @Tymscar Yeah. There's several comparisons around and the standard authors did proper scientific research. There's no visible difference except in some artificial test patterns that are designed to break it"
X Link @SebAaltonen 2025-10-28T17:55Z 46.1K followers, XXX engagements

"You need reference to sense scale. Above tiny humans are a good reference of scale. Birds offer a good reference for tall structures. In this legendary Jurassic Park scene the trees are a good reference. The dinosaur pulling/snapping the trees shows a physical response"
X Link @SebAaltonen 2025-10-17T12:33Z 46.1K followers, 1833 engagements

"If we generate a voxel terrain with some fancy perlin noise how do we know how big it is If we add trees or rocks we start sensing some scale. If we add 100000 players (see my previous MMO thread) we can use them as reference of scale. Animals vehicles etc also work well"
X Link @SebAaltonen 2025-10-17T12:39Z 46.1K followers, 1749 engagements

"Each brick has uint64 voxel mask (4x4x4) and XX bit shared child data pointer (can address 16GB of voxel data due to X byte alignment). A standard brick is XX bytes. Leaf bricks are just X bytes they don't have the child pointer (postfix doesn't cripple SIMD coherence)"
X Link @SebAaltonen 2025-10-18T09:29Z 46.1K followers, 5070 engagements

"The interesting realization is that octtree stores bits recursively in tree levels. If each level stores 4x4x4 bricks you get X bits of precision per level. Thus X levels = XX bits of precision in height. But the high bits are shared between lots of voxels"
X Link @SebAaltonen 2025-10-18T12:06Z 46.1K followers, 5471 engagements

"This is the same reason why 4x4x4 tracing ends up being faster. It's touches less different cache lines and transitions faster to near/far trace (fast empty space skip - near surface root finding)"
X Link @SebAaltonen 2025-10-18T19:19Z 46.1K followers, 3007 engagements

"2nd reason why 4x4x4 tree is better is the amount of metadata compared to payload. For each XX bits of voxels we have XX bits of link (offset). That's XX% payload / XX% link. 2x2x2 tree has XX% payload / XX% link. In data structure design you always want to minimize metadata"
X Link @SebAaltonen 2025-10-18T19:22Z 46.1K followers, 2855 engagements

"Thus even if we have all X children present the octree still doesn't utilize the cache lines fully. Partially used cache lines will be evicted from the tiny L1$ and you pay for data you didn't use. In realistic case we have less than X children so this issue is even worse"
X Link @SebAaltonen 2025-10-18T19:28Z 46.1K followers, 2615 engagements

"For the 4x4x4 tree the children often span multiple cache lines (up to 6) since they are continuously allocated. This is great for cache access pattern especially in cases of nearby voxels representing a continuous surface which is a common pattern for real content"
X Link @SebAaltonen 2025-10-18T19:30Z 46.1K followers, 2428 engagements

"And most important last: Do not mix material/texture data in your ray-tracing data structure. Keep it minimal. Find the ray hits first. Then do material/texture stuff. You don't want to load material/texture data to L1$ for voxels that you miss"
X Link @SebAaltonen 2025-10-18T20:47Z 46.1K followers, 4875 engagements

"My favourite meme picture :)"
X Link @SebAaltonen 2025-10-21T15:57Z 46.1K followers, 90.7K engagements

"And why would you even think about generating Python with AI Why would you choose a 100x slower language if AI is writing it instead of you Same applies to Javascript and other common internet languages. Just generate C and compile to WASM. Nothing runs faster"
X Link @SebAaltonen 2025-10-23T22:09Z 46.1K followers, 28.6K engagements

"Why do we need a separate headphone plug Apple finally uses the USB-C standard. USB-C can carry the audio signal (both ways) and power the headphones (noise cancelling) transfer volume/next button events from headset to the phone. USB has become the standard audio connector"
X Link @SebAaltonen 2025-10-27T13:56Z 46.1K followers, 57.7K engagements

"@KuviacM @RafaTVaz Sure TSMC and Nvidia will need to evolve their architecture and process nodes. But there's no future in sight where you don't need super dense tensor processors for example"
X Link @SebAaltonen 2025-10-28T17:40Z 46.1K followers, XX engagements

"This refactoring is MUCH simpler than the one I did before. That X week refactoring had much more global state and iffy dependencies. AI is basically a glorified search-and-replace here. But it can do more complex transforms. This is monkey work so using AI is very helpful"
X Link @SebAaltonen 2025-10-29T08:42Z 46.1K followers, 2631 engagements

"Of course I will review and test every change the AI does. This refactoring will touch over XXX files. Mostly single line changes for passing the RenderPass around in function signatures. It's nice to have a better search-and-replace. Similarly AI is a better Google search too"
X Link @SebAaltonen 2025-10-29T08:43Z 46.1K followers, 2297 engagements

"@KimmyOverflow The downsides: X. Docking station is only compatible with that laptop. When you upgrade your laptop you need to buy a new dock. X. Docking station takes table space. A single USB-C cable to the monitor takes no space and you don't need to replace it when you upgrade your laptop"
X Link @SebAaltonen 2025-10-29T10:04Z 46.1K followers, XXX engagements

"@niccruzpatane @Kicom21 To make it financially a good proposal for the car owner Tesla could for example give them free FSD monthly. This does drain the car battery quite significantly and requires more charge cycles - battery wears out. Wouldn't work if the car owner is not compensated well enough"
X Link @SebAaltonen 2025-10-29T17:42Z 46.1K followers, XXX engagements

"@carygolomb But I didn't care. I wanted a tablet that fit in my pocket. Modern iPhone Air is so much better version of Galaxy Note. 6.5" screen = XX% bigger area. And it's lighter and has a 120Hz 3000nit 27361260 OLED screen. Note X had 1280x800 pentile screen. You could see the subpixels"
X Link @SebAaltonen 2025-10-30T12:50Z 46.1K followers, XXX engagements

"@carygolomb DISCLAIMER: I don't own an iPhone Air but I have tested it a few times. Feels notably nicer than my iPhone XX Pro. I will resist updating for now. Waiting for Air X (full speed USB-C camera improvements). No rush"
X Link @SebAaltonen 2025-10-30T12:52Z 46.1K followers, XXX engagements

"@AgileJebrim The hypehype codebase is old. The new code base we had at Ubisoft had much less problems like this. Renderer was much more flat and easy to parallelize. HypeHype has too deep call stacks"
X Link @SebAaltonen 2025-10-30T17:49Z 46.1K followers, XXX engagements

"The worst device I have ever used. Ergonomics sucked so bad the latency was horrible (the mouse made the computer feel slow) and the charging port was in the bottom so you couldn't use it while charging"
X Link @SebAaltonen 2025-10-21T15:54Z 46.1K followers, 17.5K engagements

"Seems that people buy a new computer every X years. Also M4 was a great jump for entry level models. 16GB RAM and nice CPU/GPU bump. M3 was bigger jump for Max models (2x MT perf vs M1 Max) but entry level models are much more popular"
X Link @SebAaltonen 2025-10-24T13:17Z 46.1K followers, 9919 engagements

"Indexed geometry still matters. Indexed = XXXX fps Non-indexed = XXXX fps Naive non-indexed terrain renderer does 6x vertex transforms per grid cell (2 triangles). Indexed does just one. That's 6x less vertex shader work. My indexed uses 32x32 terrain patches (instanced)"
X Link @SebAaltonen 2025-10-26T08:46Z 46.1K followers, 26.4K engagements

"Patch size (grid cells) = frame rate (M3 Max): 2x2 = XXXX 4x4 = XXXX 8x8 = XXXX 16x16 = XXXX 32x32 = XXXX 64x64 = XXXX 128x128 = XXXX XXX million tris. Each patch has NxNx2 triangles and (N+1)x(N+1) vertices. Didn't analyze the cause of the differences but I have some ideas"
X Link @SebAaltonen 2025-10-26T18:42Z 46.1K followers, 5368 engagements

"Why would somebody today still want a separate audio plug that only supports one way audio (no microphone) and doesn't provide enough power for noise cancellation and can't carry volume/track button events Do you have legacy headphones"
X Link @SebAaltonen 2025-10-27T13:58Z 46.1K followers, 12.4K engagements

"Going to skip M5 Max. It's more powerful than my M3 Max but this time I am doing the upgrade for me. Last time M1-M3 Max halved the compile times = business reason = save money. M3 Max is almost perfect. The only problem is display latency. M5 Max doesn't fix it. M6 does"
X Link @SebAaltonen 2025-10-28T15:15Z 46.1K followers, 36.4K engagements

"I tried to convince the AI too many times to continue the big refactoring. It simply didn't want to do it. Asked me to do it instead. We had an argument :) "Codex ran out of room in the model's context window. Start a new conversation or clear earlier history before retrying.""
X Link @SebAaltonen 2025-10-28T20:44Z 46.1K followers, 5848 engagements

"People have been talking about the looming AI bubble for a few years now. This is how the past X months looked like. I am glad I am a believer"
X Link @SebAaltonen 2025-10-29T20:23Z 46.1K followers, 12.1K engagements

"@LukeMiani Standard professional photographer kit is 24-70mm + 70-200mm (f/2.8 both). The wide 26mm lens makes sense but a 13mm equivalent lens is niche. If you are selling houses you need one but it's not a requirement to have in your kit"
X Link @SebAaltonen 2025-10-30T13:02Z 46.1K followers, XXX engagements

"@ShishirShelke1 Also it's crazy that the 6.5" iPhone Air is just XX% heavier than iPhone 4s while it has 3.5x bigger screen area. Seems that manufactures are trying to stick at XXX gram weight for the Pro models. If bigger display is possible at that weight they increase the size"
X Link @SebAaltonen 2025-10-30T17:48Z 46.1K followers, 1396 engagements

"@cb_doge Of course he doesn't take salary. Salary = taxes. Rich people don't like to pay taxes"
X Link @SebAaltonen 2025-10-31T20:29Z 46.1K followers, XXX engagements

"@GustavSterbrant Yes indeed. On AMD you can store 256-bit descriptors inside any structs in GPU memory. On Nvidia and others you put descriptors in an array in GPU memory. Texture sampler has the descriptor cache instead of using SIMD scalar registers to store the descriptors"
X Link @SebAaltonen 2025-11-01T07:41Z 46.1K followers, XXX engagements

"I hate comparison shots like this. There's no shadows or screen space AO. A game designed without path tracing would definitely have both of those today. And the image would still look good"
X Link @SebAaltonen 2025-05-28T14:33Z 46.1K followers, 40.7K engagements

"Voxels are so easy. uint64 stores a 4x4x4 brick. uint64 a b; // two 4x4x4 bricks Union (add): a b Intersection (cut) a & b Both are a single full rate instruction on most CPUs and GPUs today. As @mmalex said a decade ago: I don't like polygons :)"
X Link @SebAaltonen 2025-10-01T05:58Z 46.1K followers, 26.5K engagements

"Sense of scale plays a huge role in perception. 81923 map at 1m3/voxel is 8km3. 12.5cm3 voxels = 1km3. Why do they look different I've been traveling seeing natural and architectural wonders. Would be nice to build a game with this kind of scale. Thread"
X Link @SebAaltonen 2025-10-17T12:30Z 46.1K followers, 17.1K engagements

"People often think voxels take a lot of storage. Let's compare a smooth terrain. Height map vs voxels. Height map: 16bit 8192x8192 = 128MB Voxels: 4x4x4 brick = uin64 = X bytes. We need 2048x2048 bricks to cover the 8k2 terrain surface = 32MB. SVO/DAG upper levels add 10%"
X Link @SebAaltonen 2025-10-18T09:17Z 46.1K followers, 22.6K engagements

"Let's discuss why I think 4x4x4 tree is better than 2x2x2 (oct) tree for voxel storage. It all boils down to link overhead and memory access patterns. L1$ hit rate is the most important thing for GPU performance nowadays. Thread"
X Link @SebAaltonen 2025-10-18T19:07Z 46.1K followers, 50K engagements

"Another hour of progress in my WebGPU compute shader voxel ray-tracer: Hint: There's no ray-tracing or voxels yet :) It works fine on my iPhone too iOS XX finally added WebGPU support. Seems to be working great so far"
X Link @SebAaltonen 2025-10-19T15:15Z 46.1K followers, 7937 engagements

"WTF is this Some windows in MacOS XX have different curve radius on the corners. Steve would have never approved this"
X Link @SebAaltonen 2025-10-20T11:51Z 46.1K followers, 12K engagements

"Awesome new feature in Xcode GPU debugger. You can expand passes to see stage dependencies now. As you can see our G-buffer pass vertex shading has no dependencies to prior GPU work meaning that it can already start during the previous frame. This gives XX% perf increase"
X Link @SebAaltonen 2025-10-20T18:40Z 46.1K followers, 15.3K engagements

"C++ ranges are safe Avoids buffer overflows and all raw loop issues Not exactly. The nice thing with raw loops is that I can easily debug them and our vector class asserts if operator is OOB. I prefer loud crash ASAP instead of silent memory corruption"
X Link @SebAaltonen 2025-10-21T10:02Z 46.1K followers, 29.9K engagements

"@theteknosaur Solving is a strong word. This is not operating on quark scale yet or even at atomic scale. Most of the physics we learn at school are large scale statistical approximations of how large clouds of tiny particles work. You can call it fake if you want at any level after the bottom"
X Link @SebAaltonen 2025-10-23T09:08Z 46.1K followers, 1605 engagements

"I chose OLED for my phone TV and PC monitor. Laptop is still LCD (but has 2010 dimming zones). Waiting for M6 Max with dual-layer OLED. Old iPad Pro is still LCD. Reason: I have never personally seen burn-in in any of my devices. Used a Panasonic Plasma for XX years too"
X Link @SebAaltonen 2025-10-23T12:55Z 46.1K followers, 12.6K engagements

"AI generated C is a real deal. C coders wrote fast & simple code. No high freq heap allocs no abstractions slowing the compiler down. Lots of good C example code around. Ai workflows need a language with fast iteration time. Why waste compile time and perf on modern languages"
X Link @SebAaltonen 2025-10-23T22:00Z 46.1K followers, 70.5K engagements

"Bug report: Car wheels don't rotate. Looks bad. Repro steps: X. Pick the Phantom X. Drive it. Expected behavior: Wheels should rotate"
X Link @SebAaltonen 2025-10-24T14:21Z 46.1K followers, 19.7K engagements

"Asked AI (Codex) to write me a 8k2 perlin noise heightmap generator and a simple triangle terrain renderer. I will use this as a comparison to develop my sparse voxel ray-tracer. AI written terrain shader calculates xy from vertex index as I asked but it's not optimal at all"
X Link @SebAaltonen 2025-10-25T20:56Z 46.1K followers, 12.6K engagements

"8192x8192 terrain. 64M grid cells. X triangles per grid cell. 128M triangles. Non-indexed = 128M * X = 384M vertex shader lanes Indexed = 64M vertex shader lanes Savings = XXX million vertex shader invocations"
X Link @SebAaltonen 2025-10-26T08:49Z 46.1K followers, 2754 engagements

"Assuming perfect vertex reuse. Instancing plays a big role here. First reason for instancing is to avoid having a 8k223*4 = 1.5GB index buffer. 32x32 patch = 1024x2 triangles = 6K indices. 16-bit indices are fine here so that's 12KB. Fits in L1$"
X Link @SebAaltonen 2025-10-26T08:54Z 46.1K followers, 2530 engagements

"Second reason for instancing is vertex reuse. If you render a big terrain (such as 8192x8192) at scanline order the second row vertices are already out of the vertex cache when you start rendering them. Thus you have to transform that row again. Which amortizes to X vx per tri"
X Link @SebAaltonen 2025-10-26T08:55Z 46.1K followers, 4582 engagements

"If you however have short enough vertex stride or if you do morton order (or similar) to your vertices the scanline is still in cache for the next row. This way you only need to transform X vertex per X triangles = XXX per triangle. Which is the optimal for vertex grids"
X Link @SebAaltonen 2025-10-26T08:56Z 46.1K followers, 6526 engagements

"I don't think AI is a bubble. Latest AI tools offer a real productivity boost. Not just for programmers but for everybody. X billion consumers paying $20/month = $XXX billion/year. Companies are willing to pay much more. 100$/month for professional AI is considered cheap"
X Link @SebAaltonen 2025-10-27T13:35Z 46.1K followers, 21.8K engagements

"Testing whether AI can write me a good changed (dirty) bitmask generator for a 64-byte struct (16x 4-byte elements). Apparently NEON doesn't have _mm_movemask_epi8 equivalent. That sucks. Also AI codegen extracts lanes X times. Could do it once at the end (bitpack in NEON)"
X Link @SebAaltonen 2025-10-28T08:21Z 46.1K followers, 9919 engagements

"I plug in a single 40Gbps USB-C cable between my computer and my monitor in the morning. Monitor has USB-hub connecting mouse and keyboard. Single USB-C cable carries the 4K 240Hz HDR image charges laptop at 65W and connects keyboard and mouse. Don't want a cable hell"
X Link @SebAaltonen 2025-10-28T14:13Z 46.1K followers, 227.3K engagements

"I had the urge to write my own graphics API too. I understand :)"
X Link @SebAaltonen 2025-10-28T16:05Z 46.1K followers, 8300 engagements

"I told the AI that I don't want a global variable because I need to multithread the algorithm. It wrote a thread local variable instead :D"
X Link @SebAaltonen 2025-10-28T17:36Z 46.1K followers, 16.9K engagements

"We were missing instructions in our file about our project structure. Added more info to avoid AI modifying external libs in the future. This also seems to speed up the AI. Finds code faster as it doesn't need to scan all the data/build/libs folders"
X Link @SebAaltonen 2025-10-29T08:45Z 46.1K followers, 3994 engagements

"@Tuhun2012 Yes I hate phones that only have wide and ultrawide lenses. Wide lens is already 26mm equivalent. I almost never need wider than that. Professional photographers usually don't carry a wider lens with them. Ultrawide is mostly useful if you are selling your home"
X Link @SebAaltonen 2025-10-29T10:02Z 46.1K followers, XXX engagements

"I prefer the iPhone Air design. It's only XX% heavier than iPhone 4s yet still has a 3.5x bigger screen. iPhone X had a horrible XX% screen-to-body ratio. Such a horrible design. The massive button and microphone wasted lots of valuable space. I didn't like that design at all"
X Link @SebAaltonen 2025-10-29T20:27Z 46.1K followers, 8840 engagements

"I was vocally against Apple phones back then. Preferred an Android phone as Apple had so small screens and awful screen to weight ratio. But iPhone X changed it all. XX% screen-to-body ratio and two cameras on a lightweight phone. 50mm equivalent standard lens. It was awesome"
X Link @SebAaltonen 2025-10-29T20:30Z 46.1K followers, 2619 engagements

"@Thornskade I never have to unplug the mouse and keyboard from my monitor. And those cables go directly from mouse/keyboard to the back of the table and fall off. I can't see them beyond that point. It's all nice and clean. Docking computer = connect one USB-C cable"
X Link @SebAaltonen 2025-10-29T21:43Z 46.1K followers, XXX engagements

"@badguy974 There's not many EU companies worth investing into. ARM and ASML are of course super important for the whole AI industry. TSMC uses ASML machines for producing all their chips for Nvidia/AMD/Apple. And ARM instruction set used in all phones tablets and Nvidia server CPUs"
X Link @SebAaltonen 2025-10-29T21:47Z 46.1K followers, XXX engagements

"@badguy974 Yes I know. EU has been doing badly recently in tech. But ARM and ASML are still good investments. Finland had the biggest phone company back in the day: Nokia. Sold more phones year than Apple + Samsung combined back in the day"
X Link @SebAaltonen 2025-10-30T07:22Z 46.1K followers, XXX engagements

"@badguy974 Nokia also got caught in the AI craze this week as Nvidia invested $1B in Nokia"
X Link @SebAaltonen 2025-10-30T07:23Z 46.1K followers, XX engagements

"@pirpirikos They have means to make money. The pro license is $200/month. I just hit my consumer license limit while doing a massive refactoring: "You've hit your usage limit. Upgrade to Pro or try again in X days XX hours XX minutes.""
X Link @SebAaltonen 2025-10-30T07:25Z 46.1K followers, XXX engagements

"@JeffGraw That's because a modern phone is just a portable touchscreen. People want as much screen area as possible with least amount of weight and dimensions. Thus the front must be XXX% display. Not much innovating in physical body. New innovation: 120Hz 3000 nits Liquid Glass UI etc"
X Link @SebAaltonen 2025-10-30T08:26Z 46.1K followers, XX engagements

"@carygolomb I had the original Galaxy Note. People laughed at how big the 5.3" screen was. That's tiny by modern standards. Nobody sells under 6" screen phones. Steve Jobs was vocal that phone must be usable with single hand. Thumb radius = max size"
X Link @SebAaltonen 2025-10-30T12:47Z 46.1K followers, XXX engagements

"@LukeMiani Give me just a single 50mm equivalent standard lens. I can live with that. Or a 35mm. Professional street photographers can live with that so can I. 26mm is already too wide. I don't need ultrawide"
X Link @SebAaltonen 2025-10-30T13:00Z 46.1K followers, XXX engagements

"The massive RenderPass refactoring (get rid of global render pass object in the UI rendering) made Codex (AI) too scared after XXX changed files so I continued using traditional methods. XXX files changed so far in XXX days of refactoring"
X Link @SebAaltonen 2025-10-30T13:44Z 46.1K followers, 7409 engagements

"@tomcr2100 Yes but Yaris is a much smaller car. Model X is not smaller than Octavia. It is actually 5cm longer. Octavia RS is a more fair comparison for Tesla Model X regarding to performance. A bigger engine to get closer to Tesla Model X acceleration. It's 1629kg. 0100 km/h: 6.4s"
X Link @SebAaltonen 2025-10-31T14:35Z 46.1K followers, XXX engagements

"Nowadays most GPUs support per lane heap index for texture sampling instructions. SIMD sends to sampler per-lane index in addition to UVs and a shared 64-bit base address if GPU uses 32-bit per-lane indices AMD is the exception. They send 256bit (4xSGPR) descriptor to sampler"
X Link @SebAaltonen 2025-10-31T15:11Z 46.1K followers, 11.1K engagements

"Also AMD has to waste some ALU cycles for the loop around the sampling instruction. And they have to call the texture sampler multiple times which reduces throughput. Just one of many reasons why Nvidia is faster in ray-tracing and other non-uniform workloads"
X Link @SebAaltonen 2025-10-31T15:16Z 46.1K followers, 1980 engagements

"@karanjagtiani04 It can actually be reduced latency too in the uniform access case as the descriptor goes directly to the sampler the sampler doesn't need to fetch it. But non-uniform access case is not great"
X Link @SebAaltonen 2025-10-31T18:18Z 46.1K followers, XXX engagements

"@kilianhekhuis Apple and Nvidia. New Qualcomm Adreno and ARM Mali also have non-uniform sampler heap index support. Intel uses heap index like the others but they don't have support for per-lane index yet. Thus Intel must do a scalarization loop like AMD"
X Link @SebAaltonen 2025-10-31T18:47Z 46.1K followers, XXX engagements

"Just landed a XXX file refactoring. Used AI in the beginning to help with me but it got too anxious of the scale of the refactoring. It started reverting all the changes when it saw HypeHype's old UI code (virtual function heavy OOP soup)"
X Link @SebAaltonen 2025-10-31T18:52Z 46.1K followers, 4832 engagements

"AI did mostly a good job. Routing the RenderPass object references to all the functions that needed it. Legacy UI code had a global RenderPass. But it did some strange moves like starting a new RenderPass for editor outline rendering instead of rendering to the existing pass"
X Link @SebAaltonen 2025-10-31T18:54Z 46.1K followers, XXX engagements

"Preordering a Tesla Roadster XXX years ago was extremely stupid move. No interest for $50k for XXX years. Huge opportunity cost. If you instead invested that $50k to Tesla stock back then you would have one million today. Even more if you invested in Nvidia"
X Link @SebAaltonen 2025-10-31T19:07Z 46.1K followers, 8475 engagements

"@Teslaconomics X years $50k deposit = X million if you invested in Tesla stock instead of the Roadster deposit. Which one is wiser"
X Link @SebAaltonen 2025-10-31T20:12Z 46.1K followers, 9984 engagements

"@Teslaconomics When the Roadster is eventually available you can immediately buy four of them in different colors. No further payment. Just that initial $50k in Tesla stock XXX years ago"
X Link @SebAaltonen 2025-10-31T20:15Z 46.1K followers, 1065 engagements

"100k players sending their state to server is not a problem. That's O(n). The big problem is updating every other player state to every player. That's O(n2). And at 100k players that's 100k*100k = 10G. Server can't obviously send 10G player state infos at acceptable rate"
X Link @SebAaltonen 2025-11-01T09:03Z 46.1K followers, 1053 engagements

"There must be a fixed budget per player. Otherwise the server will choke. This is similar to fixed bandwidth rate in the video compressors. If there's too much hard to compress new information then the quality automatically drops"
X Link @SebAaltonen 2025-11-01T09:05Z 46.1K followers, 1058 engagements

"Just like in video codecs movement prediction is a key for the algorithm. The server and the players should have the same knowledge about the motion predictor state so that the server can calculate the (projected) error of each player (for each player)"
X Link @SebAaltonen 2025-11-01T09:08Z 46.1K followers, 1175 engagements

"For very far away players quality/latency can be as bad as Uber Wolt and similar apps that show the delivery car position. You don't need pixel precise information of players that you can't interact with the moment (due to their distance to you). You only need coarse info"
X Link @SebAaltonen 2025-11-01T09:14Z 46.1K followers, 2230 engagements

"I am writing a quick WASM/WebGPU prototype of my ideas soon. Would be very difficult to get 100k real people to show up so I am likely writing some test bot to test my assumptions with"
X Link @SebAaltonen 2025-11-01T09:16Z 46.1K followers, 2398 engagements