I’m here at Microsoft’s big #MicrosoftEvent that’ll showcase its next updates for Surface and AI. Kicking things off here in NYC is @satyanadella
.
AI HARDWARE
-
Microsoft Event Showcases Surface Updates and AI Innovations in NYC
By
–
-
Amazon Launches Alexa LLM-Powered Smart Devices in AI Wars
By
–
The generative AI wars are heading home, starting with Amazon’s big debut of smart devices powered by Alexa LLM. My story about yesterday’s announcements:
-
Google Plans to Design Own AI Chips to Cut Broadcom Costs
By
–
To Reduce AI Costs, Google Wants to Ditch Broadcom as Its TPU Server Chip Supplier via @theinformation What's going on here:
Google is planning to end its relationship with Broadcom and design its own AI chips in-house by 2027 to save billions in costs. What does this mean? -
Google Reduces AI Costs by Ditching Broadcom TPU Supplier
By
–
“Google is also using a significant number of TPUs to train Gemini, its collection of advanced LLMs, which will soon power all of its key AI products for consumers and enterprises” https://
theinformation.com/articles/to-re
duce-ai-costs-google-wants-to-ditch-broadcom-as-its-tpu-server-chip-supplier
… -
Cerebras Roofline: Peak Performance Through Ultra-Fast SRAM
By
–
(1/2) This roofline plot shows theoretical performance of systems in Flops as a function of Flops per memory access Cerebras has a flat roofline: we achieve peak compute-bound perf. at lower operational intensity, thanks to our 40GB ultra-fast SRAM – accessible in a single cycle
-
Groq Booth and Security Stage Demo at Applied Live Austin
By
–
Don't miss the fun at booth #620 today and tomorrow at @AppliedEvents in Austin, TX. And you can see our public demo on the Security Stage at 2:50 CT. #AppliedLive #LPU #LanguageProcessingUnit #GenAI
-
Cerebras CS-2 Achieves Record 92.58 Petabytes Memory Bandwidth
By
–
(2/2) Key to this achievement is the development of a Tile Low-Rank Matrix-Vector Multiplication kernel, which exploits the Cerebras CS-2 architecture. This resulted in a record sustained memory bandwidth of 92.58 petabytes per second Read the paper: https://
repository.kaust.edu.sa/handle/10754/6
94388
… -
Groq CEO Says GPU Is AI Economy’s Weakest Link
By
–
"The GPU is becoming the weakest link in the AI economy. We intend to set a new standard for what the AI experience should be, while making it accessible to everyone." – @JonathanRoss321
, CEO & Founder of Groq Read more about his upcoming SCSP talk at http://
groq.link/scsppr. -
SambaNova CEO Unveils Revolutionary SN40L RDU Hardware
By
–
WATCH: @SambaNovaAI CEO and Co-Founder @RodrigoLiang live on @CNBC discussing the launch of the revolutionary SN40L RDU at #TCDisrupt2023
