[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.] #  @sudoingX Sudo su Sudo su posts on X about hardware, open ai, gpu, stack the most. They currently have XXX followers and XX posts still getting attention that total XXXXX engagements in the last XX hours. ### Engagements: XXXXX [#](/creator/twitter::1555661341914198016/interactions)  - X Week XXXXXX +292% - X Month XXXXXXX +671% ### Mentions: XX [#](/creator/twitter::1555661341914198016/posts_active)  - X Week XX +140% - X Month XXX +33% ### Followers: XXX [#](/creator/twitter::1555661341914198016/followers)  - X Week XXX +10% - X Month XXX +42% ### CreatorRank: XXXXXXX [#](/creator/twitter::1555661341914198016/influencer_rank)  ### Social Influence [#](/creator/twitter::1555661341914198016/influence) --- **Social category influence** [technology brands](/list/technology-brands) XXXXX% [stocks](/list/stocks) XXXX% [finance](/list/finance) XXXX% [countries](/list/countries) XXXX% [travel destinations](/list/travel-destinations) XXXX% [celebrities](/list/celebrities) XXXX% **Social topic influence** [hardware](/topic/hardware) 5.77%, [open ai](/topic/open-ai) 3.85%, [gpu](/topic/gpu) #929, [stack](/topic/stack) 1.92%, [epoch](/topic/epoch) 1.92%, [$googl](/topic/$googl) 1.92%, [pip](/topic/pip) 1.92%, [taiwan](/topic/taiwan) 1.92%, [$270m](/topic/$270m) 1.92%, [announces](/topic/announces) XXXX% **Top accounts mentioned or mentioned by** [@svpino](/creator/undefined) [@dolartrooper](/creator/undefined) [@domie_apps](/creator/undefined) [@yairdev](/creator/undefined) [@theaiinvestor](/creator/undefined) [@theahmadosman](/creator/undefined) [@loloelwolf97](/creator/undefined) [@cmiondotdev](/creator/undefined) [@hotaisle](/creator/undefined) [@amperlycom](/creator/undefined) [@hypertechinvest](/creator/undefined) [@outreachocta](/creator/undefined) [@amdgpu](/creator/undefined) [@cbdoge](/creator/undefined) [@aibreakfast](/creator/undefined) [@theinformation](/creator/undefined) [@stocksavvyshay](/creator/undefined) [@nastiazik](/creator/undefined) [@basedbeffjezos](/creator/undefined) [@domieapps](/creator/undefined) **Top assets mentioned** [Alphabet Inc Class A (GOOGL)](/topic/$googl) ### Top Social Posts [#](/creator/twitter::1555661341914198016/posts) --- Top posts by engagements in the last XX hours "@AMDGPU_ XXX vs XXX TOPS = negligible. $2300 vs $4000 = significant. x86 vs ARM = depends on your stack. If ROCm supports your inference pipeline AMD wins on economics. That's the real gating factor" [X Link](https://x.com/sudoingX/status/1978698271401902407) [@sudoingX](/creator/x/sudoingX) 2025-10-16T05:43Z XXX followers, 1400 engagements "Running GPUs 24/7 for 16+ months taught me: Power costs hardware costs over time. PCIe bandwidth matters more than VRAM. Cooling is not optional. Infrastructure problems look exactly like software problems until they don't" [X Link](https://x.com/sudoingX/status/1980160481966104908) [@sudoingX](/creator/x/sudoingX) 2025-10-20T06:33Z XXX followers, XXX engagements "AWS: down My GPU cluster: up This is the compute convenience trade-off. When cloud infrastructure fails local infrastructure just. keeps working. Expensive upfront. Priceless when everyone else is locked out" [X Link](https://x.com/sudoingX/status/1980541801661641114) [@sudoingX](/creator/x/sudoingX) 2025-10-21T07:48Z XXX followers, XXX engagements "Unless. the browser IS the path to AGI. Not technically. Economically. AGI requires: massive compute + massive data + massive distribution Browser = infinite training data from every user action Browser = platform lock-in that funds the compute Chrome made Google invincible. Atlas could do the same for OpenAI" [X Link](https://x.com/sudoingX/status/1980797638699282536) [@sudoingX](/creator/x/sudoingX) 2025-10-22T00:45Z XXX followers, 3848 engagements "Main blocker for switching from NVIDIA: Issues: RDNA X installation: device detection fails driver signatures missing dependency hell Newest AMD chips unsupported (Ryzen AI 300) Support dropped for recent cards unpredictably Setup: CUDA = XX min ROCm = 4-40 hours Would switch if: X. All current-gen RDNA supported X. pip install simplicity X. 5+ year support windows Hardware advantage is real (price/VRAM). Software experience is the gap" [X Link](https://x.com/sudoingX/status/1981197548531622310) [@sudoingX](/creator/x/sudoingX) 2025-10-23T03:14Z XXX followers, XXX engagements "@amperlycom @redtachyon Fair correction. Meta's not fighting for profit. They're fighting for relevance in an AI race they're losing on mindshare despite winning on margins" [X Link](https://x.com/sudoingX/status/1981260883897831537) [@sudoingX](/creator/x/sudoingX) 2025-10-23T07:26Z XXX followers, XX engagements "@HyperTechInvest AMD investing $270M in Taiwan infrastructure while NVIDIA announces space datacenters. One company building cooling for 1500W chips. Other launching H100s to orbit. Different strategies. One's about margins. One's about headlines" [X Link](https://x.com/sudoingX/status/1981671254232039604) [@sudoingX](/creator/x/sudoingX) 2025-10-24T10:36Z XXX followers, XXX engagements "@andre_banandre This is the AMD hardware advantage we were discussing earlier. Beats NVIDIA on raw performance. Question remains: ecosystem maturity. If Strix Halo ships with stable ROCm support this changes the inference economics" [X Link](https://x.com/sudoingX/status/1981646989642391746) [@sudoingX](/creator/x/sudoingX) 2025-10-24T09:00Z XXX followers, XXX engagements "@ns123abc Non-profit for-profit transitions are messy. The real question: can you build AGI-scale compute without billions in capital Structure follows funding reality even if optics are bad" [X Link](https://x.com/sudoingX/status/1978323878100296012) [@sudoingX](/creator/x/sudoingX) 2025-10-15T04:55Z XXX followers, 4833 engagements "@JonhernandezIA ATLAS is OpenAI's models in a Chromium based browser wrapper not "your" ChatGPT. "Your ChatGPT" = local model on hardware you own. No API. No company between you and inference" [X Link](https://x.com/sudoingX/status/1981702618658545869) [@sudoingX](/creator/x/sudoingX) 2025-10-24T12:41Z XXX followers, XXX engagements "@Yuchenj_UW Measuring engineering impact by lines of code is like measuring chef quality by number of ingredients used. Gordon Ramsay would be unemployed. Guy Fieri thriving" [X Link](https://x.com/sudoingX/status/1982004433958633836) [@sudoingX](/creator/x/sudoingX) 2025-10-25T08:40Z XXX followers, 2562 engagements "1. Power: Even at solo builder scale this is real. My 6x3090s = 2.1kW sustained. Apartment wiring was not designed for this. X. HBM premium: Why 3090s (GDDR6X) still make sense vs newer cards. Price/VRAM ratio matters more than peak FLOPS. X. Storage for inference: Underrated. NVMe speed matters when loading 70B models" [X Link](https://x.com/sudoingX/status/1982009966195097646) [@sudoingX](/creator/x/sudoingX) 2025-10-25T09:02Z XXX followers, XXX engagements "Running 6x3090s in Bangkok taught me things AWS docs won't tell you: PCIe riser quality matters: Cheap risers = mysterious stability issues. Good risers cost 3x more but save weeks of debugging "random" crashes. Ambient temp: 30C room = thermal throttling starts way earlier than specs suggest. AC isn't optional infrastructure. Power delivery: Your breakers and wiring matter more than GPU specs. My apartment wasn't designed for 2.1kW sustained load. The gap between "works in the cloud" and "works on hardware you own" is where real infrastructure knowledge lives. Cloud abstracts this away." [X Link](https://x.com/sudoingX/status/1982102725929406926) [@sudoingX](/creator/x/sudoingX) 2025-10-25T15:11Z XXX followers, XX engagements "@HyperTechInvest FPGAs for quantum error correction makes sense. Need: real time processing reconfigurability deterministic latency Don't need: massive parallel compute (GPU territory) Right tool for the job. AMD's Xilinx acquisition looking smarter every quarter" [X Link](https://x.com/sudoingX/status/1982014353395134791) [@sudoingX](/creator/x/sudoingX) 2025-10-25T09:20Z XXX followers, XX engagements "@elonmusk Grok Imagine looks good. When does Grok CLI drop 2M context in terminal = massive upgrade over everything else" [X Link](https://x.com/sudoingX/status/1982179555222966649) [@sudoingX](/creator/x/sudoingX) 2025-10-25T20:16Z XXX followers, XXX engagements
[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.]
@sudoingX Sudo su
Sudo su posts on X about hardware, open ai, gpu, stack the most. They currently have XXX followers and XX posts still getting attention that total XXXXX engagements in the last XX hours.
Engagements: XXXXX #
- X Week XXXXXX +292%
- X Month XXXXXXX +671%
Mentions: XX #
- X Week XX +140%
- X Month XXX +33%
Followers: XXX #
- X Week XXX +10%
- X Month XXX +42%
CreatorRank: XXXXXXX #
Social Influence #
Social category influence technology brands XXXXX% stocks XXXX% finance XXXX% countries XXXX% travel destinations XXXX% celebrities XXXX%
Social topic influence hardware 5.77%, open ai 3.85%, gpu #929, stack 1.92%, epoch 1.92%, $googl 1.92%, pip 1.92%, taiwan 1.92%, $270m 1.92%, announces XXXX%
Top accounts mentioned or mentioned by @svpino @dolartrooper @domie_apps @yairdev @theaiinvestor @theahmadosman @loloelwolf97 @cmiondotdev @hotaisle @amperlycom @hypertechinvest @outreachocta @amdgpu @cbdoge @aibreakfast @theinformation @stocksavvyshay @nastiazik @basedbeffjezos @domieapps
Top assets mentioned Alphabet Inc Class A (GOOGL)
Top Social Posts #
Top posts by engagements in the last XX hours
"@AMDGPU_ XXX vs XXX TOPS = negligible. $2300 vs $4000 = significant. x86 vs ARM = depends on your stack. If ROCm supports your inference pipeline AMD wins on economics. That's the real gating factor"
X Link @sudoingX 2025-10-16T05:43Z XXX followers, 1400 engagements
"Running GPUs 24/7 for 16+ months taught me: Power costs hardware costs over time. PCIe bandwidth matters more than VRAM. Cooling is not optional. Infrastructure problems look exactly like software problems until they don't"
X Link @sudoingX 2025-10-20T06:33Z XXX followers, XXX engagements
"AWS: down My GPU cluster: up This is the compute convenience trade-off. When cloud infrastructure fails local infrastructure just. keeps working. Expensive upfront. Priceless when everyone else is locked out"
X Link @sudoingX 2025-10-21T07:48Z XXX followers, XXX engagements
"Unless. the browser IS the path to AGI. Not technically. Economically. AGI requires: massive compute + massive data + massive distribution Browser = infinite training data from every user action Browser = platform lock-in that funds the compute Chrome made Google invincible. Atlas could do the same for OpenAI"
X Link @sudoingX 2025-10-22T00:45Z XXX followers, 3848 engagements
"Main blocker for switching from NVIDIA: Issues: RDNA X installation: device detection fails driver signatures missing dependency hell Newest AMD chips unsupported (Ryzen AI 300) Support dropped for recent cards unpredictably Setup: CUDA = XX min ROCm = 4-40 hours Would switch if: X. All current-gen RDNA supported X. pip install simplicity X. 5+ year support windows Hardware advantage is real (price/VRAM). Software experience is the gap"
X Link @sudoingX 2025-10-23T03:14Z XXX followers, XXX engagements
"@amperlycom @redtachyon Fair correction. Meta's not fighting for profit. They're fighting for relevance in an AI race they're losing on mindshare despite winning on margins"
X Link @sudoingX 2025-10-23T07:26Z XXX followers, XX engagements
"@HyperTechInvest AMD investing $270M in Taiwan infrastructure while NVIDIA announces space datacenters. One company building cooling for 1500W chips. Other launching H100s to orbit. Different strategies. One's about margins. One's about headlines"
X Link @sudoingX 2025-10-24T10:36Z XXX followers, XXX engagements
"@andre_banandre This is the AMD hardware advantage we were discussing earlier. Beats NVIDIA on raw performance. Question remains: ecosystem maturity. If Strix Halo ships with stable ROCm support this changes the inference economics"
X Link @sudoingX 2025-10-24T09:00Z XXX followers, XXX engagements
"@ns123abc Non-profit for-profit transitions are messy. The real question: can you build AGI-scale compute without billions in capital Structure follows funding reality even if optics are bad"
X Link @sudoingX 2025-10-15T04:55Z XXX followers, 4833 engagements
"@JonhernandezIA ATLAS is OpenAI's models in a Chromium based browser wrapper not "your" ChatGPT. "Your ChatGPT" = local model on hardware you own. No API. No company between you and inference"
X Link @sudoingX 2025-10-24T12:41Z XXX followers, XXX engagements
"@Yuchenj_UW Measuring engineering impact by lines of code is like measuring chef quality by number of ingredients used. Gordon Ramsay would be unemployed. Guy Fieri thriving"
X Link @sudoingX 2025-10-25T08:40Z XXX followers, 2562 engagements
"1. Power: Even at solo builder scale this is real. My 6x3090s = 2.1kW sustained. Apartment wiring was not designed for this. X. HBM premium: Why 3090s (GDDR6X) still make sense vs newer cards. Price/VRAM ratio matters more than peak FLOPS. X. Storage for inference: Underrated. NVMe speed matters when loading 70B models"
X Link @sudoingX 2025-10-25T09:02Z XXX followers, XXX engagements
"Running 6x3090s in Bangkok taught me things AWS docs won't tell you: PCIe riser quality matters: Cheap risers = mysterious stability issues. Good risers cost 3x more but save weeks of debugging "random" crashes. Ambient temp: 30C room = thermal throttling starts way earlier than specs suggest. AC isn't optional infrastructure. Power delivery: Your breakers and wiring matter more than GPU specs. My apartment wasn't designed for 2.1kW sustained load. The gap between "works in the cloud" and "works on hardware you own" is where real infrastructure knowledge lives. Cloud abstracts this away."
X Link @sudoingX 2025-10-25T15:11Z XXX followers, XX engagements
"@HyperTechInvest FPGAs for quantum error correction makes sense. Need: real time processing reconfigurability deterministic latency Don't need: massive parallel compute (GPU territory) Right tool for the job. AMD's Xilinx acquisition looking smarter every quarter"
X Link @sudoingX 2025-10-25T09:20Z XXX followers, XX engagements
"@elonmusk Grok Imagine looks good. When does Grok CLI drop 2M context in terminal = massive upgrade over everything else"
X Link @sudoingX 2025-10-25T20:16Z XXX followers, XXX engagements