Tired of slow thinking with OpenAI o1? Try our demo using the fastest Llama 405B on SambaNova Cloud built by @KyleLiang5
. Try it here:
COMPUTING
-
SambaNova Llama 405B faster alternative to OpenAI o1
By
–
-
PyTorch Async Tensor Parallelism Now Publicly Available
By
–
While async Tensor Parallelism is common among the elite private codebases, the PyTorch team put together a public, accessible and readable one. pretty cool work from @foofoobuggy @cHHillee @lessw2020 @lc_whr tianyu and @wanchao_ Read more here: https://
discuss.pytorch.org/t/distributed-
w-torchtitan-introducing-async-tensor-parallelism-in-pytorch/209487
… -
Tech Giants Face Scrutiny Over Data Center Emissions
By
–
excellent report on squishy emissions math at large-cap tech from @xisabelobrienx https://
theguardian.com/technology/202
4/sep/15/data-center-gas-emissions-tech
… -
Volvo EX90 Debuts NVIDIA Drive Orin-Powered AI Platform
By
–
.
@volvocars rolls out the EX90 SUV, first vehicle in its EV lineup built on @NVIDIADRIVE Orin, running NVIDIA DriveOS; Announces plans to integrate NVIDIA DRIVE Thor in future models and taps NVIDIA DGX for advanced #AI training. -
Study on the continuation of AI scaling through 2030
By
–
Study:
https://
epochai.org/blog/can-ai-sc
aling-continue-through-2030
… -
AI Compute Power: 10,000x GPT-4 by 2030, Is It a Bottleneck?
By
–
Analysis: By 2030, we will see a 10,000-fold increase in computing power compared to GPT-4. Epoch AI's analysis forecasts that AI training runs could reach up to 2e29 FLOP (floating point operations) in about 5 years. Will we hit an electricity/chips/data bottleneck? Link to
-
Oracle promises accelerating growth through 2029 amid stock surge
By
–
Oracle: Enough in the trunk to keep the shares rising? Oracle dazzled analysts on Thursday with a promise of accelerating revenue growth through 2029. Will it be enough to keep the stock gaining after a fifty-four percent rise this year? https://
thetechnologyletter.com/the-posts/orac
le-enough-in-the-trunk-to-keep-the-shares-rising
… $ORCL -
Cloud Strategy Success Depends on Enabling People Leadership
By
–
Talking with a Fortune 500 company about their tech choices, almost everything comes down to two words: "It depends." But, mostly, it depends on people. I wrote down some thoughts on how to drive change by enabling people to lead that change https://
infoworld.com/article/352212
4/when-your-cloud-strategy-is-it-depends.html
… for @infoworld -
MathWorks Automates Code Generation for Qualcomm Hexagon NPU
By
–
Exciting news! MathWorks has launched a hardware support package to automate code generation from MATLAB & Simulink models to the @Qualcomm Hexagon NPU, eliminating manual code optimization and enabling Processor-in-the-Loop testing. Read more https://
spr.ly/6017WtL25 -
India’s AI Infrastructure Readiness: Scalable Compute Strategies
By
–
The panel discussion on 'IndiaAI: India’s Infrastructure Readiness for AI' held at the Global IndiaAI Summit in June 2024, centered on key issues related to AI infrastructure readiness in India, focusing on strategies for developing scalable and future-ready AI compute