@MetaAI championing #openAI
- next-generation open source LLM #Llama2 released; weights and model publicly available
- Model available from 7B to 70B parameters trained on 2T tokens
- free for commercial use (<700M MAU)
- available at @Azure@awscloud@huggingface
🧵
Scale of models available by #Meta allow the #AI to break out of the traditional DC infra and actually be deployed on USER devices
@Qualcomm announced the #Llama2 based AI implementations on #Smartphones and #PC starting from 2024
> #Smartphone METAmorphosis underway
>> #Datacenter 🔼42% at $1.7B & #Embedded 🔼 1868% at $1.4B as the silver lining of the revenues for Q4;
>> #PC🔻51% at $0.9B #Gaming 🔻7% at $1.6B bear the brunt of macro #headwinds
>>Inventory normalization still to happen in PC market and surprisingly #datacenter
>> So the supply issues have been cleared out for AMD both for #Cloud and #networking
>> #Genoa seems to have an adoption roadblock; as #cloud and #enterprise demand not as strong
>> As per @LisaSu the cloud demand should be strong for H2; Bergamo also to be launched in H2
For #Client#Computing Group:
>> Intel seems bullish on the total PC Market
• Revenue🔻36% YoY to $6.6B
• Operating Income🔻82% to $700M
• Huge #Inventory Corrections expected in Q1 2023
• 2023 Sell in units (TAM) to be ~270M for PC market
#Intel$INTC
For DCAI #datacenter Group:
• Revenue🔻33% YoY to $4.3B
• Operating Income🔻84% to $400M
• #Genoa#AMD giving LOT of competition and taking away market share
• #Intel betting on Installed base in #Cloud (95%) with SPR for $$$
• SPR to ship 1M units by mid-2023
#Intel Sapphire Rapids is out after a long pause!
▶️The #Xeon launched with 60 cores w/focus on in-chip acceleration to compete with #AMD
▶️ #XeonMax for #HPC workload
▶️ 64GB of HBM2e for #XeonMax
▶️ Uses 4 Intel 7 node tiles using EMIB
▶️ DDR5, PCIe5 and CXL support
@intel announces one of its important CPU lineup
🔥4th Gen Xeon (Sapphire Rapids) w/ built-in tons of accelerators and security engines.
🔥Intel Xeon CPU Max Series (Sapphire Rapids HBM)
🔥Intel Data Center GPU Max Series (Ponte Vecchio)
#IntelXeon#SapphireRapids
Race to become the smartest just got interesting!
With #google#nvidia all focusing on #AI with their S/W stacks
#Microsoft aims to mainstream #AI usage by intention to acquire #ChatGPT
The panel talking about how viewpoint is changing the #NBA experience at home with different views (the BEST ones!) , also opening up new ventures of #organicadvertising integrating seamlessly in #VR systems