48 years ago today, Ward Christensen and Randy Suess launched CBBS, the worldโs first dial-up bulletin board system โ and online communities were born.
From one phone line and a modem in 1978 to the BBS revival weโre still running today, the spirit of connection lives on.
This Months Dialed Back in Compute!โs Gazette celebrating these 48 years from Ward and Randy!
A Baudy anniversary to the pioneers who started it all.
prototype raymarching SDF engine experiment #003.
Using a smoothMin function let's us combine geometric primitives to form complex surfaces. Now to figure out how to upscale the final swapchain image with a single #vulkan #compute "shader" because my poor cheap minipc struggles to cast more than 900x600 rays@30fps. #screenshotsaturday
Fetching the video to learn
Thank you ๐ for sharing
https://www.youtube.com/watch?v=Ef70x0izkFU
๐ฆ๐โค๏ธ๐#Lobi ๐๐๐น๐๐๐ฆ
#Raspberry #Pi5 #SBC #Compute #model #technology #notebook #Linux #OpenSource #programming #POSIX
Avinav Sahoo(he/ him/ his) (@avinavsahoo)
์์ฑ์๋ '์ฐ๋ฆฌ๊ฐ ์ฌ์ค์ ๊ณ์ฐ ํจ์จ์ฑ ๊ฒฝ๊ณ(compute efficient frontier)๋ฅผ ๋์์์ง๋ ๋ชจ๋ฅธ๋ค'๊ณ ์ฃผ์ฅํ๋ฉฐ Sam Altman ๋ฑ ์ ๊ณ ์ฃผ์ ์ธ๋ฌผ๋ค์ ํ๊ทธํด ์ค์ํ ์ฐ๊ตฌยท๊ณตํ์ ์ด์ ํ ๊ฐ๋ฅ์ฑ์ ์ ๊ธฐํฉ๋๋ค. ์ด๋ ๋ชจ๋ธ ํ๋ จ ๋น์ฉยท์ค์ผ์ผ๋ง ํจ์จ ์ธก๋ฉด์์ ํฐ ๋ณ๊ณก์ ์ผ ์ ์์์ ์์ํ๋ฉฐ AI ์ฐ๊ตฌยท์ ์ฑ ๋ ผ์์ ํ์ฅ์ ์ค ๋งํ ๋ฐํ๋ก ํด์๋ฉ๋๋ค.
I don't think rallying against AI no-matter-what is the right way to oppose its deployment.
I'm super bullish on Chinese AI labs, which, due to export controls favour efficiency in their models, and it is foreseeable that pushing for efficient AI models will allow us to break the massive centralisation and inevitable enshittification of AI.
Having skilled professionals wielding AI, similar to the philosophy of Primeagen's 99 AI coding plugin for neovim, will probably be a net benefit and augmentor rather than a detractor
Also, efficient, deployable at home models will even further eat into the negative margins that Western AI labs have, similar to how pushing against AI-derived work being copyrightable will make the billions invested by Hollywood into AI worthless.
Here I borrowed from @pluralistic 's thought that traditional copyright litigation is not a effective tool for protecting creators from gen AI, as that will primarily benefit major corporations rather than individual creators.
NVIDIA Just Killed OpenAIโs $100 Billion Deal
The landscape of AI infrastructure has shifted overnight. NVIDIAโs latest move into direct cloud services has effectively undermined OpenAIโs massive $100 billion compute deal, signaling a new era of competition between the hardware giants and the software pioneers.
#NVIDIA #OpenAI #Technews #Compute #AI #SiliconValley #Business #technology #News #Science
https://www.technology-news-channel.com/nvidia-just-killed-openais-100-billion-deal/
โ #Canada must establish a framework to govern #Canadian-owned #AI #compute and #data. This includes regulating #stablecoins and #digitalpaymentsystemsโthe #infrastructure of tomorrow's #commerce.โ
RE: https://bsky.app/profile/did:plc:5zca2ola2zxpkw37w4f3wxtu/post/3mduxhq435c2c
Q*Satoshi (@AiXsatoshi)
์๋ จํ ๊ธฐ์ ์๊ฐ ๊ณ์ฐ ์์์ ๋จ์ํ ์๋์ง ๋ณํ๊ธฐ๊ฐ ์๋๋ผ ์ง๋ฅ์ ์ฆํญํ๋ ์ฅ์น๋ก ์ ํํ๋ค๊ณ ์ฃผ์ฅ. ๊ธฐ์ ํ์ ์ด ์ฐ์ฐ ํจ์จ์ ๋์ด์ฌ๋ฆฌ๋ฉด ์์ฑ๋๋ ๊ฐ์น๊ฐ ๊ธฐํ๊ธ์์ ์ผ๋ก ์ฆ๊ฐํ๋ค๋ ์ ์ ๊ฐ์กฐํ๋ ๋ด์ฉ์ด๋ค.
"Mแบกng computeprices.com ฤฦฐแปฃc ra mแบฏt ฤแป phรขn tรญch trung tรขm dแปฏ liแปu, huแบฅn luyแปn AI vร thแป trฦฐแปng thuรช cloud. Tuแบงn trฦฐแปc, trang bแป sung giรก mรด hรฌnh ngรดn ngแปฏ lแปn (LLM) song song giรก GPU. ฤang nรขng cแบฅp theo dรตi giรก LLM & mแป rแปng nguแปn. Bแบกn muแปn tรญnh nฤng gรฌ thรชm? #Compute #AI #CloudComputing #TrungTรขmDแปฏLiแปu #MรดHรฌnhNgรดnNgแปฏLแปn"
Startup Project (@startupprojectt)
Cerebras์ ์ต๊ณ ๊ฒฝ์์ง์ด ์ ์ํ ํต์ฐฐ: 70B(700์ต) ํ๋ผ๋ฏธํฐ๊ธ ๋ชจ๋ธ์์ ํ ๋จ์ด๋ฅผ ์์ฑํ๋ ค๋ฉด ์ ์ฒด ๊ธธ์ด ์ํ 100ํธ ๋ถ๋์ ํด๋นํ๋ ๋ฐ์ดํฐ ์ด๋์ด ํ์ํ๋ค๋ ์ฃผ์ฅ. ์ด๋ ๋ํ ์ธ์ด๋ชจ๋ธ(LLM) ์ถ๋ก ์ ๋์ ๋์ง ์๋ ๋ง๋ํ ๋ฐ์ดํฐ ์ด๋ ๋ฐ ๊ณ์ฐ ๋น์ฉ์ ๊ฐ์กฐํ๋ ๋ด์ฉ์ ๋๋ค.
investment data (@value_charts)
๊ธ๋ก๋ฒ AI ์ปดํจํ (์ฐ์ฐ) ์ฉ๋์ด ์ฝ 7๊ฐ์๋ง๋ค ๋ ๋ฐฐ๋ก ์ฆ๊ฐํ๊ณ ์์ผ๋ฉฐ, ์ ๋ฐ์ ์ธ AI ์ฐ์ ์ด ๊ธฐํ๊ธ์์ ์ฑ์ฅ์ธ๋ฅผ ๋ณด์ธ๋ค๋ ๊ด์ธก์ ๋๋ค. ์ด๋ฌํ ๊ธ์ํ ์ปดํจํ ํ์ฅ์ ๋ชจ๋ธ ํ๋ จยท์ถ๋ก ๋ฅ๋ ฅ ๋ฐ ์๋น์ค ๋ฐฐํฌ ์ญ๋์ ๋น ๋ฅด๊ฒ ํ๋ํด ์ฐ๊ตฌยท์ ํํ ์๋์ ํฐ ์ํฅ์ ์ค ๊ฐ๋ฅ์ฑ์ด ์์ต๋๋ค.
Just give me back affordable RAM, SSDs and GPUs already.
Q*Satoshi (@AiXsatoshi)
์ด ๊ธ์ ์์ง: ์๋์ฌ ์๋์๋ ํ ์ง๊ฐ ํต์ฌ ์์ฐ์ด์๋ฏ์ด, ๋์งํธ ์๋์๋ ๊ณ์ฐ(์ปดํจํ ) ์์์ด ๋์งํธ ๋์ง์ ๊ธฐ๋ฐ์ด๋ผ๋ ๊ด์ ์ด๋ค. ์๋์ง๋ง์ด ์๋๋ผ ๊ณ์ฐ ์์์ ํ๋ณด์๋ ๊ตญ๊ฐ ์ ๋ต ์์ค์ ์๊ธ ํฌ์๊ฐ ๊ณ์๋ ๊ฒ์ด๋ผ๋ ์ ๋ง์ ์ ์ํ๋ค.
OpenAI (@OpenAI)
AI์์ ์ฐ์ฐ(compute)์ด ๊ฐ์ฅ ํฌ์ํ ์์์ด๋ฉฐ ์์๊ฐ ๊ณ์ ์ฆ๊ฐํ๋ค๋ ๋ ผ์์ ๋๋ค. OpenAI ํ์บ์คํธ์์ CFO ์ฌ๋ผ ํ๋ผ์ด์ด์ Khosla Ventures ์ฐฝ์ ์ Vinod Khosla๊ฐ ํธ์คํธ Andrew Mayne์ ํจ๊ป ์ฐ์ฐ ์์, ์ธํ๋ผ ํ๊ณ, AI ํํ์ ๋ ๋ง์ ์ฌ๋์๊ฒ ์ ๋ฌํ๋ ๋ฐฉ๋ฒ์ ๋ ผ์ํฉ๋๋ค. ๊ฐ๋ฐยท์ธํ๋ผ ๊ด์ ์์ ์ฐธ๊ณ ํ ๋งํ ๋ํ์ ๋๋ค.
#OpenAl says its #compute power grew from 0.2 GW in 2023 to 1.9 GW in 2025 and annualized #revenue grew from $2B in 2023 to $20B+ in 2025. OpenAIโs #business model scales with the value delivered by intelligence, from #consumer subscriptions to #enterprise solutions and #commerce. https://openai.com/index/a-business-that-scales-with-the-value-of-intelligence/?AIagents.at #AIagent #AI #ML #NLP #LLM #GenAI
This is my #Atari8Bit #NOS setup for writing my #Compute Articles with AtariWriter:
NOS from: TNFS://irata.online/Atari_8-bit/DOS/NOS.atr
AtariWriter in XEX form from atariwiki.org
AUTORUN.BAT in pic.
The NTRANS 2 tells #FujiNet to make LINUX ASCII text files. Windows users would use 1.
Save documents on N2: and they are readable on my PC.
TechRadar (@techradar)
Meta๊ฐ AI ์ธํ๋ผ ํ์ฅ์ ๊ด์ฅํ ์กฐ์ง 'Meta Compute'๋ฅผ ๋ฐํํ์ต๋๋ค. ๋ฆฌ๋์ญ์ ํตํฉํด AI ์ธํ๋ผ๋ฅผ ๊ด๋ฆฌยทํ๋ํ๊ณ ํฅํ 10๋ ๊ฐ ์์ญ์์ ์๋ฐฑ ๊ธฐ๊ฐ์ํธ ๊ท๋ชจ์ ์ ๋ ฅ ๊ณํ์ ์ธ์ฐ๋ฉฐ ๋๊ท๋ชจ ๋ฐ์ดํฐ์ผํฐยท์ปดํจํธ ํ์ถฉ์ ์ค๋นํ๋ค๋ ๋ด์ฉ์ ๋๋ค. ํ์ฌ ์ฐจ์์ AI ์ธํ๋ผ ์ ๋ต๊ณผ ์๋์ง ์์ ์ ๋ง์ด ํต์ฌ์ ๋๋ค.
Latent.Space (@latentspacepod)
์ต๊ทผ ๋ช ๋ ๋์ '์ง๋ฅ์ ๋น์ฉ'์ด ๊ทน์ ์ผ๋ก ํ๋ฝํ๊ณ ์๋ค๋ ๊ด์ฐฐ๋ก, ์ง๋ฅ์ ํ ๋ผ ๋จ์๋ก ์๊ฐํ ๋ ๋จ๊ฐ๊ฐ ํฌ๊ฒ ๋จ์ด์ก์ผ๋ฉฐ ๊ทธ ๊ฒฐ๊ณผ GPT-4 ์์ค์ ์ง๋ฅ์ 100๋ฐฐ ์ด์ ์ ๋ ดํ๊ฒ ์ป์ ์ ์๋ค๋ ์ฃผ์ฅ์ ๋๋ค. AI ๋ชจ๋ธ ์ด์ฉยท๊ฐ๋ฐ์ ๊ฒฝ์ ์ฑ์ด ํฌ๊ฒ ๊ฐ์ ๋๊ณ ์์์ ์์ฌํ๋ ์ฐ์ ํธ๋ ๋์ ๋๋ค.
