Communications of the ACM
The blog archive provides access to past blog postings from Communications of the ACM and other sources by date.
October 2024
From Computational Complexity
Favorite Theorems: Gradient Descent
Who thought the algorithm behind machine learning would have cool complexity implications?
John Fearnley, Paul Goldberg, Alexandros Hollender and Rahul Savani…From Schneier on Security
California AI Safety Bill Vetoed
Governor Newsom has vetoed the state’s AI safety bill.
I have mixed feelings about the bill. There’s a lot to like about it, and I want governments to regulate in this space. But, for now, it’s all EU.
(Related, the Council of…From Putting People First
Dutch Media Authority: “Government should force Big Tech to do news”
Research by the Dutch Media Authority shows that 78% of Dutch youth get their news mainly from social media, in particular Instagram, TikTok and YouTube, reports Wilfred Takken in the Dutch NRC newspaper today. The research…
From insideHPC
VAST Data Unveils InsightEngine with Nvidia
AI data platform company VAST Data today announced VAST InsightEngine with NVIDIA, which the company said is the first solution to securely ingest, process, and retrieve all types of enterprise data (files, objects, tables, and…
From insideHPC
Hyperion Research Announces AI in the Cloud Findings
ST. PAUL, MN, Oct. 1, 2024—HPC-AI industry analyst firm Hyperion Research announced the results of their recently completed AI in the Cloud study, entitled, Cloud-based AI Activity for HPC: Widespread but Primarily Exploratory…
From insideHPC
Optical I/O: The Key to Unlocking AI Infrastructure Profitability and Performance
[SPONSORED GUEST ARTICLE] Scaling GenAI inference performance requires increasing the number of GPUs or accelerators working in parallel within the scale-up domain. In-package optical I/O offers a path forward by breaking the…
From insideHPC
Nvidia H100s GPUs Available on IBM Cloud for AI and HPC
Oct. 1, 2024 — IBM announced the availability of NVIDIA H100 Tensor Core GPU instances on IBM Cloud. It joins IBM Cloud’s existing lineup of accelerated computing offerings for enterprise’s AI implementation. IBM said that despite…
From Schneier on Security
Hacking ChatGPT by Planting False Memories into Its Data
This vulnerability hacks a feature that allows ChatGPT to have long-term memory, where it uses information from past conversations to inform future conversations with that same user. A researcher found that he could use that …