[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 Avatar](https://lunarcrush.com/gi/w:26/cr:twitter::3436615413.png) @SebAaltonen Sebastian Aaltonen

Sebastian Aaltonen posts on X about voxels, all the, gpu, faster the most. They currently have XXXXXX followers and XXX posts still getting attention that total XXXXXX engagements in the last XX hours.

### Engagements: XXXXXX [#](/creator/twitter::3436615413/interactions)
![Engagements Line Chart](https://lunarcrush.com/gi/w:600/cr:twitter::3436615413/c:line/m:interactions.svg)

- X Week XXXXXXX -XX%
- X Month XXXXXXXXX +111%
- X Months XXXXXXXXX -XX%
- X Year XXXXXXXXXX -XX%

### Mentions: XX [#](/creator/twitter::3436615413/posts_active)
![Mentions Line Chart](https://lunarcrush.com/gi/w:600/cr:twitter::3436615413/c:line/m:posts_active.svg)

- X Week XXX +33%
- X Month XXX +55%
- X Months XXX -XX%
- X Year XXXXX -XX%

### Followers: XXXXXX [#](/creator/twitter::3436615413/followers)
![Followers Line Chart](https://lunarcrush.com/gi/w:600/cr:twitter::3436615413/c:line/m:followers.svg)

- X Week XXXXXX +0.30%
- X Month XXXXXX +3.60%
- X Months XXXXXX +11%
- X Year XXXXXX +20%

### CreatorRank: XXXXXXX [#](/creator/twitter::3436615413/influencer_rank)
![CreatorRank Line Chart](https://lunarcrush.com/gi/w:600/cr:twitter::3436615413/c:line/m:influencer_rank.svg)

### Social Influence [#](/creator/twitter::3436615413/influence)
---

**Social category influence**
[technology brands](/list/technology-brands)  XXXX% [stocks](/list/stocks)  XXXXX% [nfts](/list/nfts)  #781 [countries](/list/countries)  XXXX% [finance](/list/finance)  XXXX% [gaming](/list/gaming)  XXXX% [cryptocurrencies](/list/cryptocurrencies)  XXXX% [social networks](/list/social-networks)  XXXX% [automotive brands](/list/automotive-brands)  XXXX% [formula 1](/list/formula-1)  XXXX%

**Social topic influence**
[voxels](/topic/voxels) #3, [all the](/topic/all-the) 5.51%, [gpu](/topic/gpu) #25, [faster](/topic/faster) #1803, [finland](/topic/finland) #456, [fortnite](/topic/fortnite) #2027, [realtime](/topic/realtime) #690, [china](/topic/china) #4470, [tech companies](/topic/tech-companies) #5, [elements](/topic/elements) #1389

**Top accounts mentioned or mentioned by**
[@agilejebrim](/creator/undefined) [@jbrooksbsi](/creator/undefined) [@addictivecolors](/creator/undefined) [@0xe1](/creator/undefined) [@oswald_sn](/creator/undefined) [@opinali](/creator/undefined) [@zorgmonkey](/creator/undefined) [@cfanaticin2025](/creator/undefined) [@vmsousapereira](/creator/undefined) [@filpizlo](/creator/undefined) [@simplex_fx](/creator/undefined) [@customwetware](/creator/undefined) [@amd](/creator/undefined) [@lisasu](/creator/undefined) [@rouanijihad](/creator/undefined) [@atomlib](/creator/undefined) [@redneckleeusmc](/creator/undefined) [@csubagio](/creator/undefined) [@never_released](/creator/undefined) [@keyth0s](/creator/undefined)

**Top assets mentioned**
[Voxels (voxels)](/topic/voxels) [Alphabet Inc Class A (GOOGL)](/topic/$googl) [Perlin (PERL)](/topic/perlin) [Microsoft Corp. (MSFT)](/topic/microsoft) [Nokia Corporation (NOK)](/topic/$nok) [Tesla, Inc. (TSLA)](/topic/tesla)
### Top Social Posts [#](/creator/twitter::3436615413/posts)
---
Top posts by engagements in the last XX hours

"@henningsanden It's not technically hard. But it is labor intensive for artists to add all the small details. Labor intensive = adds cost = is hard to fit in the budget"  
[X Link](https://x.com/SebAaltonen/status/1979475068959838455) [@SebAaltonen](/creator/x/SebAaltonen) 2025-10-18T09:10Z 45.7K followers, 79.7K engagements


"Utilizing "GPT5 Thinking" model for graphics data model optimizations (adjoint vs inverse transpose). It's quite common that I have an idea and I use "GPT5 Thinking" model to validate my thoughts and calculate rough estimates for ALU cycles etc"  
[X Link](https://x.com/SebAaltonen/status/1979574710435279285) [@SebAaltonen](/creator/x/SebAaltonen) 2025-10-18T15:46Z 45.7K followers, 19.2K 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](https://x.com/SebAaltonen/status/1979630681396482143) [@SebAaltonen](/creator/x/SebAaltonen) 2025-10-18T19:28Z 45.7K followers, 2555 engagements


"There's also a lot of competing ways to spend the energy. AI training and inference is hitting a power wall. AI companies are buying all the cheap electricity they can. More demand = more expensive electricity. Hydrogen car dream relies on almost free electrify"  
[X Link](https://x.com/SebAaltonen/status/1979896923185356851) [@SebAaltonen](/creator/x/SebAaltonen) 2025-10-19T13:06Z 45.7K followers, 1291 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](https://x.com/SebAaltonen/status/1980343465725817330) [@SebAaltonen](/creator/x/SebAaltonen) 2025-10-20T18:40Z 45.7K followers, 13.7K engagements


"Tech companies are filled with programmers but the tools are used by artists. You need your own artists and technical artists to build good workflows. Programmers also need big scenes with lots of assets to test editor and runtime performance"  
[X Link](https://x.com/SebAaltonen/status/1980610359133094388) [@SebAaltonen](/creator/x/SebAaltonen) 2025-10-21T12:21Z 45.7K followers, 3773 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](https://x.com/SebAaltonen/status/1979929485177663972) [@SebAaltonen](/creator/x/SebAaltonen) 2025-10-19T15:15Z 45.7K followers, 7774 engagements


"This is not just a Unity problem. I have seen the same problem in many tech companies including HypeHype and Ubisoft. One of the reasons why Epic is doing so well is Fortnite. They have a big dogfood project. XXX player multiplayer dynamic world big environment"  
[X Link](https://x.com/SebAaltonen/status/1980611106449018938) [@SebAaltonen](/creator/x/SebAaltonen) 2025-10-21T12:24Z 45.7K followers, 3798 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](https://x.com/SebAaltonen/status/1927734858652012816) [@SebAaltonen](/creator/x/SebAaltonen) 2025-05-28T14:33Z 45.6K followers, 40.6K 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](https://x.com/SebAaltonen/status/1973266236881690872) [@SebAaltonen](/creator/x/SebAaltonen) 2025-10-01T05:58Z 45.6K followers, 26.4K engagements


"Performance was also the main reason Unity's Weta Digital acquisition failed. You can't just buy a movie company and believe that real time game engine + movies = real time movies. A massive amount of optimization and architecture refactoring work is required to make it happen"  
[X Link](https://x.com/SebAaltonen/status/1976922036082823297) [@SebAaltonen](/creator/x/SebAaltonen) 2025-10-11T08:05Z 45.6K followers, 7597 engagements


"This was also the main reason I left Unity. I got tired of pushing performance to middle/top level managers who didn't seem to understand that we NEED to prioritize high performance tech such as DOTS and GPU-driven rendering if we want to make real-time movies happen"  
[X Link](https://x.com/SebAaltonen/status/1976923309062471987) [@SebAaltonen](/creator/x/SebAaltonen) 2025-10-11T08:10Z 45.6K followers, 8434 engagements


"Wondering what the "universal compression" actually means. We already had DCC and samplers and display engine could read compressed textures. RT also had custom decoder. I guess this means that all buffer data is compressed in memory Of course seamless decode (no layouts)"  
[X Link](https://x.com/SebAaltonen/status/1976947851210162181) [@SebAaltonen](/creator/x/SebAaltonen) 2025-10-11T09:47Z 45.7K followers, 19.7K engagements


"Spent an extra hour on my WebGPU/WASM prototype. Compute shader fills the pixels to offscreen storage texture then there's a blit pass (full screen triangle)"  
[X Link](https://x.com/SebAaltonen/status/1976974577336304103) [@SebAaltonen](/creator/x/SebAaltonen) 2025-10-11T11:34Z 45.6K followers, 15.7K 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](https://x.com/SebAaltonen/status/1979163135815692594) [@SebAaltonen](/creator/x/SebAaltonen) 2025-10-17T12:30Z 45.6K followers, 16.9K 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](https://x.com/SebAaltonen/status/1979163919563284777) [@SebAaltonen](/creator/x/SebAaltonen) 2025-10-17T12:33Z 45.6K followers, 1805 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](https://x.com/SebAaltonen/status/1979165275015590100) [@SebAaltonen](/creator/x/SebAaltonen) 2025-10-17T12:39Z 45.6K followers, 1731 engagements


"SDF tiling equivalent:"  
[X Link](https://x.com/SebAaltonen/status/1979174411782459672) [@SebAaltonen](/creator/x/SebAaltonen) 2025-10-17T13:15Z 45.6K followers, 2536 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](https://x.com/SebAaltonen/status/1979476961010520207) [@SebAaltonen](/creator/x/SebAaltonen) 2025-10-18T09:17Z 45.7K followers, 22.3K engagements


"The above estimate is optimistic. If we have a rough terrain we end up having two bricks on top of each other in most places. Thus we have 64MB worth of leaf bricks. SVO/DAG upper levels don't increase much (as we use shared child pointers). Total is 70MB. Still a win"  
[X Link](https://x.com/SebAaltonen/status/1979477529141407962) [@SebAaltonen](/creator/x/SebAaltonen) 2025-10-18T09:19Z 45.6K followers, 4096 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](https://x.com/SebAaltonen/status/1979480028959121597) [@SebAaltonen](/creator/x/SebAaltonen) 2025-10-18T09:29Z 45.6K followers, 5026 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](https://x.com/SebAaltonen/status/1979519419060818034) [@SebAaltonen](/creator/x/SebAaltonen) 2025-10-18T12:06Z 45.7K followers, 5366 engagements


"@codyvasy @oisyn Oh yeah it's a 4x4x4 tree instead of 2x2x2 (oct) tree. So there's half the levels and each level does a 4-way split on each axis splitting the region to four thus two bits worth of extra precision per region in the next level"  
[X Link](https://x.com/SebAaltonen/status/1979566064439546310) [@SebAaltonen](/creator/x/SebAaltonen) 2025-10-18T15:11Z 45.7K followers, XX engagements


"@WuYixiang4084 @tinivule @henningsanden These things are usually done in decaling/polishing passes. If the game world is big (GTA X = biggest game project ever) then it's going to take time and an army of artists. Fingerprints in one pole is not a lot but achieving similar level of polish everywhere is significant"  
[X Link](https://x.com/SebAaltonen/status/1979609450399862796) [@SebAaltonen](/creator/x/SebAaltonen) 2025-10-18T18:04Z 45.6K followers, XX engagements


"@DubDublin @TechOperator EU passed a law forcing EV chargers to accept credit cards and not discriminate price based on payment method. Previously you had to download charger network app + register. I had X different apps on my phone. Now I don't need to do that anymore. Much better"  
[X Link](https://x.com/SebAaltonen/status/1979624738273251809) [@SebAaltonen](/creator/x/SebAaltonen) 2025-10-18T19:04Z 45.6K followers, XX engagements


"In a 4x4x4 tree our uint64 gives the ray at least X steps (DDA) worth of data using a single memory load. In best case XX steps. These extra steps allow the ray to get further from a surface faster. When you exit the 4x4x4 brick you are further away"  
[X Link](https://x.com/SebAaltonen/status/1979627358840930772) [@SebAaltonen](/creator/x/SebAaltonen) 2025-10-18T19:15Z 45.6K followers, 3641 engagements


"Same applies to all the information I read in books magazines news and in the social media. If the information is in direct conflict with my knowledge I need more sources to ensure the validity. AI is no exception. It's an information source like everything else"  
[X Link](https://x.com/SebAaltonen/status/1979871707486900473) [@SebAaltonen](/creator/x/SebAaltonen) 2025-10-19T11:26Z 45.7K followers, XXX engagements


"@Jonathan_Blow I have to admit that I didn't fully understand integrals before I did prefix sums based memory allocation on the GPU. Prefix sum is the discrete integral (sum) of all previous elements. Integrals are just addition with all the nice addition qualities (order composition etc)"  
[X Link](https://x.com/SebAaltonen/status/1979893782171763159) [@SebAaltonen](/creator/x/SebAaltonen) 2025-10-19T12:53Z 45.7K followers, XXX engagements


"Energy companies will never willingly build oversupply of energy. Power plants (especially nuclear) cost a lot to build. If the expected energy selling price is close to zero this is not a good business. Also lower energy prices reduces the profits of their current plants"  
[X Link](https://x.com/SebAaltonen/status/1979896422767181833) [@SebAaltonen](/creator/x/SebAaltonen) 2025-10-19T13:04Z 45.7K followers, 1786 engagements


"It's hard to see politicians and voters choosing to build 3x more nuclear plants for hydrogen cars and a massive effort for hydrogen fuel delivery network on top of that. The existing power grid + small extra effort is fine enough for standard EVs. Much less hassle"  
[X Link](https://x.com/SebAaltonen/status/1979897924793327778) [@SebAaltonen](/creator/x/SebAaltonen) 2025-10-19T13:10Z 45.7K followers, 1237 engagements


"What about the fluctuation caused by wind and solar There are couple of days in Finland every year with free electricity. Can't we use it to run a small batch of hydrogen cars The problem is that heavy industries NEED hydrogen and it's expensive and produced with fossil fuels"  
[X Link](https://x.com/SebAaltonen/status/1979898848160997401) [@SebAaltonen](/creator/x/SebAaltonen) 2025-10-19T13:14Z 45.7K followers, 1212 engagements


"@AddictiveColors @ZorgMonkey @AMD @LisaSu Isn't that the whole idea behind the UDNA It takes tensor cores and multi-chip scalability tech from the CDNA and the modern compute units schedulers compressors and gfx bits from the RDNA"  
[X Link](https://x.com/SebAaltonen/status/1979904881071697944) [@SebAaltonen](/creator/x/SebAaltonen) 2025-10-19T13:38Z 45.7K followers, XXX engagements


"@AddictiveColors @ZorgMonkey @AMD @LisaSu Volta was Nvidia's professional chip but they chose to dump all the new tech to Turing. Everybody got full tensor cores and ray-tracing (and all the big architectural improvements). The rest is history. CUDA won. Nvidia is a $3T company. Availability matters"  
[X Link](https://x.com/SebAaltonen/status/1979905446958792985) [@SebAaltonen](/creator/x/SebAaltonen) 2025-10-19T13:40Z 45.6K followers, XXX engagements


"@AddictiveColors @ZorgMonkey @AMD @LisaSu DLSS was the killer app. Otherwise there wouldn't have been any business sense to give tensor cores to consumers. Slow AI adaptation in games didn't matter since Nvidia driver was using the tensor cores in all games. As a bonus researchers got CUDA tensor cores in all laptops"  
[X Link](https://x.com/SebAaltonen/status/1979906096379691119) [@SebAaltonen](/creator/x/SebAaltonen) 2025-10-19T13:42Z 45.6K followers, XXX engagements


"@AddictiveColors @ZorgMonkey @AMD @LisaSu FSR4 and Sony PSSR are both AI based. They benefit from tensor cores. Apple also added tensor cores. They also need them for similar purposes. AMD can't afford to be left behind. And they need to focus on tensor cores anyways for AI business. They just made a big OpenAI deal"  
[X Link](https://x.com/SebAaltonen/status/1979906530020413685) [@SebAaltonen](/creator/x/SebAaltonen) 2025-10-19T13:44Z 45.6K followers, XX engagements


"@minitbnn @simplex_fx I use the formatting rules that the project uses. I am just a single person the team always has much more engineers. I use my own preferences in my own hobby projects"  
[X Link](https://x.com/SebAaltonen/status/1979913628653343086) [@SebAaltonen](/creator/x/SebAaltonen) 2025-10-19T14:12Z 45.6K followers, XXX engagements


"@AgileJebrim @cfanaticin2025 @vmsousapereira @filpizlo I don't want abstractions to hide what the hardware actually does. Modern GPUs can all handle pointers efficiently. That's what is under the hood. I want to get rid of the buffer wrappers"  
[X Link](https://x.com/SebAaltonen/status/1979914756170437098) [@SebAaltonen](/creator/x/SebAaltonen) 2025-10-19T14:17Z 45.7K followers, 1471 engagements


"@AgileJebrim @cfanaticin2025 @vmsousapereira @filpizlo Same is true for all the vertex buffer abstractions. Much nicer just to use a C-style array indexing. Modern GPUs load vertices using standard raw loads. struct MyVertex fields ; MyVertex vertex = verticesglVertexId; where vertices is a 64-bit GPU pointer"  
[X Link](https://x.com/SebAaltonen/status/1979915456279371853) [@SebAaltonen](/creator/x/SebAaltonen) 2025-10-19T14:20Z 45.7K followers, 1102 engagements


"@AgileJebrim @cfanaticin2025 @vmsousapereira @filpizlo We can continue this discussion after I have released my blog post. The first topic is getting rid of the buffer abstraction and all the messy driver magic behind it. There's just GPU pointers. Shaders have a simple C-style ptr syntax. No messy binding model"  
[X Link](https://x.com/SebAaltonen/status/1979918523481624818) [@SebAaltonen](/creator/x/SebAaltonen) 2025-10-19T14:32Z 45.6K followers, XXX engagements


"@AgileJebrim @cfanaticin2025 @vmsousapereira @filpizlo I am not talking about stuff like the VK_KHR_buffer_device_address. That basically exposes the same GPU features but in a super messy way. GLSL and HLSL are not designed to handle pointers"  
[X Link](https://x.com/SebAaltonen/status/1979918819662393535) [@SebAaltonen](/creator/x/SebAaltonen) 2025-10-19T14:33Z 45.6K followers, XXX engagements


"@AgileJebrim @cfanaticin2025 @vmsousapereira @filpizlo Yes there are artificial language constraints. GLSL and HLSL are ancient languages designed for writing small 1:1 transform functions for vertex and pixel elements. These languages were designed before you could even load/store data to/from a buffer"  
[X Link](https://x.com/SebAaltonen/status/1979920692079419826) [@SebAaltonen](/creator/x/SebAaltonen) 2025-10-19T14:40Z 45.6K followers, XXX engagements


"@JBrooksBSI The scaling is indeed bad since we touch less and less voxels as the brick size increases. According to my math the sweet spot should be 4x4x4. It nicely fits into a single 64-bit register (efficient mem load efficient addressing with bit ops etc)"  
[X Link](https://x.com/SebAaltonen/status/1979972441708409216) [@SebAaltonen](/creator/x/SebAaltonen) 2025-10-19T18:06Z 45.7K followers, XXX engagements


"@never_released Of course it's nice to not having to do that kind of compiler gymnastics and waste instruction slots for it. But it's still SIMD. All lanes do single instruction or single texture fetch. Not multiple even if there was some hardware support for per-lane IP or per-lane heap index"  
[X Link](https://x.com/SebAaltonen/status/1980213715426586986) [@SebAaltonen](/creator/x/SebAaltonen) 2025-10-20T10:05Z 45.6K followers, XXX engagements


"@0xE1 Yes and this is nice. There's couple of summer days in Finland too every year when electricity is free. But is it cost efficient to build big hydrolysis plants and only use them couple of weeks every year"  
[X Link](https://x.com/SebAaltonen/status/1980214976494366812) [@SebAaltonen](/creator/x/SebAaltonen) 2025-10-20T10:10Z 45.6K followers, XXX engagements


"@0xE1 I discussed about the potential future in the thread. Maybe we have enough excess energy at some point so summers have free electricity. Building a new power plant at that point is bad for business. Why would they build extra oversupply. It kills energy company profits"  
[X Link](https://x.com/SebAaltonen/status/1980215312428716402) [@SebAaltonen](/creator/x/SebAaltonen) 2025-10-20T10:11Z 45.6K followers, XXX engagements


"@0xE1 Solar fields will be grown regardless at rapidly expanding rate. The need of electricity is growing rapidly due to AI training and inference. We don't need hydrogen cars for this equation at all"  
[X Link](https://x.com/SebAaltonen/status/1980216480936640887) [@SebAaltonen](/creator/x/SebAaltonen) 2025-10-20T10:16Z 45.6K followers, XX engagements


"@G_melo_ding @bernaferrari @ecoezen Yeah. ROG Ally has a 7" 1080p 120Hz touchscreen. Also it has a 4x bigger SSD and XX% more RAM than the entry level Mac Mini. And a Windows OS license. ASUS is not Microsoft. They have to pay for the OS too"  
[X Link](https://x.com/SebAaltonen/status/1980277328417538475) [@SebAaltonen](/creator/x/SebAaltonen) 2025-10-20T14:18Z 45.6K followers, XXX engagements


"@Meito_1 But there's no guarantee for iPhone Air X release next year. It has a different numbering scheme. How knows when they update it. And iPhone Fold might be coming. I have no interest for a 250+ gram fold. Would be exactly the opposite I want"  
[X Link](https://x.com/SebAaltonen/status/1980366768788107330) [@SebAaltonen](/creator/x/SebAaltonen) 2025-10-20T20:13Z 45.7K followers, XXX engagements


"@Pirazel Kelvin mega peta henrys :D"  
[X Link](https://x.com/SebAaltonen/status/1980568889143750851) [@SebAaltonen](/creator/x/SebAaltonen) 2025-10-21T09:36Z 45.7K followers, XX engagements


"Apple M5 GPU has tensor cores similar to Nvidia and AMD's professional CDNA. Older Apple GPUs already had tensor intrinsics but these intrinsics ran using the vector pipelines similar to AMD RDNA3 WMMA. AMD UDNA is also expected to have tensor cores"  
[X Link](https://x.com/SebAaltonen/status/1979469715933462927) [@SebAaltonen](/creator/x/SebAaltonen) 2025-10-18T08:48Z 45.7K followers, 25K engagements


"Of course we have to ask: Why not a single cache line aligned nodes. That's iffy because A) are storing sparse data (surface near band is not a tile) B) GPU SIMD branch/loop coherency issues (homogenerous tiles are important). Allocating children contiguously is almost as good"  
[X Link](https://x.com/SebAaltonen/status/1979648280570311038) [@SebAaltonen](/creator/x/SebAaltonen) 2025-10-18T20:38Z 45.7K followers, 2056 engagements


"To electrify all cars in Finland we would need one additional nuclear power plant. If we do the same with cars based on hydrogen electrolysis we need three additional nuclear power plants"  
[X Link](https://x.com/SebAaltonen/status/1979897471523230038) [@SebAaltonen](/creator/x/SebAaltonen) 2025-10-19T13:08Z 45.7K followers, 1299 engagements


"The only reason China is not yet conquering the USA and EU car market is massive protective tariffs. They are rapidly conquering every other country. Soon the same will happen for all computing devices. USA will lose this massive market that they always owned. Great move indeed"  
[X Link](https://x.com/SebAaltonen/status/1979921384097693801) [@SebAaltonen](/creator/x/SebAaltonen) 2025-10-19T14:43Z 45.7K followers, 4813 engagements


"@stgewehr @theonecid The funny thing in Finland is that companies can give as many phones they want to employees tax free but legally a telephone contract (data/speak/SMS) is worth of 20/month of taxable income"  
[X Link](https://x.com/SebAaltonen/status/1980658507272814645) [@SebAaltonen](/creator/x/SebAaltonen) 2025-10-21T15:32Z 45.7K followers, XXX 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](https://x.com/SebAaltonen/status/1980713159783588315) [@SebAaltonen](/creator/x/SebAaltonen) 2025-10-21T19:09Z 45.7K followers, 5317 engagements


"@AgileJebrim The idea is to have a fixed cost regardless of overdraw. This requires very fine grained occlusion culling and LOD. If you put a new object in the view that object occludes as many pixels behind it. Thus the total amount of pixels always stays the same. That's the main idea"  
[X Link](https://x.com/SebAaltonen/status/1979889917456203833) [@SebAaltonen](/creator/x/SebAaltonen) 2025-10-19T12:38Z 45.7K followers, XXX engagements


"@AgileJebrim Aren't you splatting (scattering) all voxels to the froxel viewport If you scatter 100x overdraw worth of buildings (city street view) that's 100x more expensive than scattering only the first visible surface"  
[X Link](https://x.com/SebAaltonen/status/1979891056566256015) [@SebAaltonen](/creator/x/SebAaltonen) 2025-10-19T12:43Z 45.7K followers, XXX engagements


"@AgileJebrim Voxels and triangles behave similarly if you cluster the meshes to small spatially local clusters. Voxel bricks (let's say 8x8x8 bricks) are similar to geometry clusters and can utilize similar occlusion culling strategy such as a hierarchical Z buffer"  
[X Link](https://x.com/SebAaltonen/status/1979891822584561775) [@SebAaltonen](/creator/x/SebAaltonen) 2025-10-19T12:46Z 45.7K followers, XX engagements


"@ecoezen Apple doesn't do cheap products. One of the reasons they are leading in efficiency is that they can afford to use only the leading edge process nodes. That's only possible with premium products and high profit margins"  
[X Link](https://x.com/SebAaltonen/status/1979909983291461966) [@SebAaltonen](/creator/x/SebAaltonen) 2025-10-19T13:58Z 45.7K followers, 1380 engagements


"@midnight_john1 M4 Max was tied with 5090 laptop when running on battery. Nvidia GPUs often throttle to around XX% when unplugged. Apple maintains full CPU and GPU clock rates and has reasonable battery life time too despite heavy CPU and GPU workloads"  
[X Link](https://x.com/SebAaltonen/status/1979936004954783783) [@SebAaltonen](/creator/x/SebAaltonen) 2025-10-19T15:41Z 45.7K followers, 1008 engagements


""It's much faster performance and to load things" Zuck's reasoning started with performance. Performance matters. Google maps won because of performance Nokia lost because of Symbian OS not being designed for real-time systems (touchscreen needs it). Unity has to wake up"  
[X Link](https://x.com/SebAaltonen/status/1976919371823169765) [@SebAaltonen](/creator/x/SebAaltonen) 2025-10-11T07:54Z 45.7K followers, 68.9K engagements


"Asserts are a double-edged sword. One should only use asserts for code bugs never for data. Data should be validated at load/receive time before inserted into runtime data structures. Error handling should not be in the low level processing/transform code"  
[X Link](https://x.com/SebAaltonen/status/1979230610435481991) [@SebAaltonen](/creator/x/SebAaltonen) 2025-10-17T16:58Z 45.7K followers, 37K engagements


"@theredpix This is nice I am personally waiting for M6 since I want the new OLED screen that's rumored to be released in the next year model"  
[X Link](https://x.com/SebAaltonen/status/1979567912525697209) [@SebAaltonen](/creator/x/SebAaltonen) 2025-10-18T15:19Z 45.7K followers, XXX 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](https://x.com/SebAaltonen/status/1979625344748728652) [@SebAaltonen](/creator/x/SebAaltonen) 2025-10-18T19:07Z 45.7K followers, 49K 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](https://x.com/SebAaltonen/status/1979628423976034541) [@SebAaltonen](/creator/x/SebAaltonen) 2025-10-18T19:19Z 45.7K followers, 2952 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](https://x.com/SebAaltonen/status/1979629227873063129) [@SebAaltonen](/creator/x/SebAaltonen) 2025-10-18T19:22Z 45.7K followers, 2801 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](https://x.com/SebAaltonen/status/1979631144066023594) [@SebAaltonen](/creator/x/SebAaltonen) 2025-10-18T19:30Z 45.7K followers, 2367 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](https://x.com/SebAaltonen/status/1979650498321055902) [@SebAaltonen](/creator/x/SebAaltonen) 2025-10-18T20:47Z 45.7K followers, 4248 engagements


"A person asked me: How can I use AI to validate my thoughts when AI is known to lie. I treat AI as a single source of information. If I agree with it then I have more confidence in my thoughts. If I disagree with it then I need more sources. I never solely trust AI"  
[X Link](https://x.com/SebAaltonen/status/1979871235564867614) [@SebAaltonen](/creator/x/SebAaltonen) 2025-10-19T11:24Z 45.7K followers, 10.7K engagements


"Apple has a great GPU architecture. M3 added innovative new register file. M5 was another nice generational jump in GPU performance and finally added tensor cores for 4x higher AI throughput"  
[X Link](https://x.com/SebAaltonen/status/1979876513232408910) [@SebAaltonen](/creator/x/SebAaltonen) 2025-10-19T11:45Z 45.7K followers, 44.8K engagements


"@JoaoBapt Link is just a 32-bit aligned offset if your data is 4-byte aligned then a 32-bit link it can address up to 16GB of voxel data. 16GB address space is fine for gaming purposes"  
[X Link](https://x.com/SebAaltonen/status/1979879500071702901) [@SebAaltonen](/creator/x/SebAaltonen) 2025-10-19T11:57Z 45.7K followers, XXX engagements


"Lots of hydrogen car hype recently. Let's get some facts straight. Hydrogen car = EV with hydrogen fuel cell to produce electricity. Hydrogen can be created with water and electricity but this process needs 3x as much electricity as charging EV directly from the grid. Thread"  
[X Link](https://x.com/SebAaltonen/status/1979895309808595005) [@SebAaltonen](/creator/x/SebAaltonen) 2025-10-19T13:00Z 45.7K followers, 11.7K engagements


"I think USA made a massive mistake. A short sighted move. They controlled the chip market and the AI hardware market. But they forced China to become their biggest competitor. Let's see how many years it takes for China to catch up in process tech. Less than a decade for sure"  
[X Link](https://x.com/SebAaltonen/status/1979920262045859968) [@SebAaltonen](/creator/x/SebAaltonen) 2025-10-19T14:39Z 45.7K followers, 36.1K engagements


"The only sane explanation is that this is a stalling technique. They need couple of years to conquer the AI market worldwide. And it doesn't matter for them if they lose everything else since the AI market will dominate. It's an extremely risky move with lots of failure modes"  
[X Link](https://x.com/SebAaltonen/status/1979922529440465337) [@SebAaltonen](/creator/x/SebAaltonen) 2025-10-19T14:48Z 45.7K followers, 4513 engagements


"@Sebasti66855537 Nvidia's geometry pipelines are super awesome. Always been and Turing also bumped those up drastically (both the classic pipes and added mesh shaders). Nvidia can achieve awesome occupancy in vertex-only workloads. Nobody can touch them when it comes to pure triangle throughput"  
[X Link](https://x.com/SebAaltonen/status/1979931195740754074) [@SebAaltonen](/creator/x/SebAaltonen) 2025-10-19T15:22Z 45.7K followers, 1268 engagements


"@JBrooksBSI Yes if it is axis aligned (best case). The maximum amount of touched voxels is x+y+z: X for a 2x2x2 brick and XX for a 4x4x4 brick. 2x2x2 unused bits = X. 4x4x4 unused bits = XX. However both have XX bits of metadata which is a waste"  
[X Link](https://x.com/SebAaltonen/status/1979970987392176544) [@SebAaltonen](/creator/x/SebAaltonen) 2025-10-19T18:00Z 45.7K followers, XXX engagements


"@JBrooksBSI Also the worst case 4x4x4 traversal = XX voxels would touch six 2x2x2 bricks in the same region thus we load 6x 32-bit metadata instead of just a single 32-bit metadata. This is a big win. It's not just about voxels the metadata is also a waste"  
[X Link](https://x.com/SebAaltonen/status/1979971636242641035) [@SebAaltonen](/creator/x/SebAaltonen) 2025-10-19T18:03Z 45.7K followers, XXX engagements


"@DrEliDavid Big European and American companies moved their physical goods production to China because it was cheaper. Even iPhones are made in China. Physical good production requires factories and factories cause pollution. Now attributed to China. We are the top consumers of Temu + Shein"  
[X Link](https://x.com/SebAaltonen/status/1980163464942240239) [@SebAaltonen](/creator/x/SebAaltonen) 2025-10-20T06:45Z 45.7K followers, 4312 engagements


"The latest Pixel XX has GPU performance equivalent to iPhone XX Pro. Google barely increased the GPU performance in X generations. Maybe I am the problem We shouldn't aim to improve mobile game graphics anymore"  
[X Link](https://x.com/SebAaltonen/status/1980165535695311309) [@SebAaltonen](/creator/x/SebAaltonen) 2025-10-20T06:53Z 45.7K followers, 29.6K 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](https://x.com/SebAaltonen/status/1980186233553998224) [@SebAaltonen](/creator/x/SebAaltonen) 2025-10-20T08:16Z 45.7K followers, 7336 engagements


"@never_released IP per lane is a bit similar to texture heap index per lane. That doesn't allow us to sample divergent textures either. We must do a loop that takes lowest exec mask lane broadcasts the texture index of it sets exec mask on to all lanes with identical index sample goto 0"  
[X Link](https://x.com/SebAaltonen/status/1980213326631325912) [@SebAaltonen](/creator/x/SebAaltonen) 2025-10-20T10:03Z 45.7K followers, XXX engagements


"WTF is this Some windows in MacOS XX have different curve radius on the corners. Steve would have never approved this"  
[X Link](https://x.com/SebAaltonen/status/1980240352289415185) [@SebAaltonen](/creator/x/SebAaltonen) 2025-10-20T11:51Z 45.7K followers, 11.5K engagements


"@LAFadhel Didn't they just do a big deal with Qualcomm Snapdragon Elite for Galaxy"  
[X Link](https://x.com/SebAaltonen/status/1980262788409839925) [@SebAaltonen](/creator/x/SebAaltonen) 2025-10-20T13:20Z 45.7K followers, 1446 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](https://x.com/SebAaltonen/status/1980349179957490011) [@SebAaltonen](/creator/x/SebAaltonen) 2025-10-20T19:03Z 45.7K followers, 12.5K engagements


"@Meito_1 That's a good point. More polished Air X would be great. A slightly bigger 64Mpix sensor with 16Mpix crop for 50mm equivalent image would be perfect. And a full speed USB-C port too"  
[X Link](https://x.com/SebAaltonen/status/1980365222750957883) [@SebAaltonen](/creator/x/SebAaltonen) 2025-10-20T20:07Z 45.7K followers, XXX 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](https://x.com/SebAaltonen/status/1980575428659257695) [@SebAaltonen](/creator/x/SebAaltonen) 2025-10-21T10:02Z 45.7K followers, 22.6K engagements


"@simplex_fx Safe to overflow to beginning of the data structure That corrupts data silently"  
[X Link](https://x.com/SebAaltonen/status/1980583986150937085) [@SebAaltonen](/creator/x/SebAaltonen) 2025-10-21T10:36Z 45.7K followers, XXX engagements


"@chrisoffner3d @AnxiousHolly Say that to Steve. He wanted people to polish everything before he was happy"  
[X Link](https://x.com/SebAaltonen/status/1980584301017330120) [@SebAaltonen](/creator/x/SebAaltonen) 2025-10-21T10:37Z 45.7K followers, XXX engagements


"@RouaniJihad Fortnite was initially running at 30fps on consoles. There was a big optimization effort to lift consoles to 60fps around the UE XXXX time frame. After that Epic was able to ship Fortnite to mobile phones. These big optimizations were back-ported to mainstream UE4"  
[X Link](https://x.com/SebAaltonen/status/1980619400077046073) [@SebAaltonen](/creator/x/SebAaltonen) 2025-10-21T12:57Z 45.7K followers, XXX engagements


"@RouaniJihad Sure it's a long time since UE XXXX. Fortnite has bloated and the engine has bloated. Lumen Nanite ray-tracing etc have become common"  
[X Link](https://x.com/SebAaltonen/status/1980619561146749438) [@SebAaltonen](/creator/x/SebAaltonen) 2025-10-21T12:57Z 45.7K followers, XXX engagements


"OK let's do a quick comparison: Tesla Model X LR vs Mercedes CLA 250+ Frunk: 88-100l (+12l) Range: 750-792km (+42km) Charging: 250kW(400V)-320kW(800V) (+70kW) Price: 45-50k (+5k = downside) Not all German cars are sub-optimal. CLA also feels great to drive (D+ mode)"  
[X Link](https://x.com/SebAaltonen/status/1980660993928884309) [@SebAaltonen](/creator/x/SebAaltonen) 2025-10-21T15:42Z 45.7K followers, 4247 engagements


"People always shit on phone companies saying that there's no progress. Just look at how much space old phones wasted for a button and microphone. Air has 3x display area (+6x nits and 2x refresh rate) and it's only XX% heavier That's progress"  
[X Link](https://x.com/SebAaltonen/status/1980674414233387211) [@SebAaltonen](/creator/x/SebAaltonen) 2025-10-21T16:35Z 45.7K followers, 4781 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](https://x.com/SebAaltonen/status/1980710090031120865) [@SebAaltonen](/creator/x/SebAaltonen) 2025-10-21T18:57Z 45.7K followers, XXX engagements


"How about iPhone X X years difference. iPhone Air advantages: - Almost 2x screen area - Over 4x brightness - 2x refresh rate - XX% bigger battery - Only XX% heavier X years old iPhone has measly XXXX% screen to body ratio. iPhone Air uses the area much better at 89%"  
[X Link](https://x.com/SebAaltonen/status/1980711326101250452) [@SebAaltonen](/creator/x/SebAaltonen) 2025-10-21T19:02Z 45.7K followers, 1580 engagements


"The GPU in iPhone Air is 4x faster. The camera much better too. 4x megapixels and captures 2x light (bigger sensor faster lens)"  
[X Link](https://x.com/SebAaltonen/status/1980714913921392671) [@SebAaltonen](/creator/x/SebAaltonen) 2025-10-21T19:16Z 45.7K followers, 1682 engagements


"@fedyac @atomlib Yes. It's cherry picking. I also waited for the iPhone Air release. If I would have done this comparison last year the iPhone XX Pro Max would have fared much worse"  
[X Link](https://x.com/SebAaltonen/status/1980719883240501554) [@SebAaltonen](/creator/x/SebAaltonen) 2025-10-21T19:36Z 45.7K followers, XXX engagements