L' #ANSSI publie un guide sur les recommandations de sécurité pour un système d’ #IA générative.

🎯Il vise à sensibiliser les administrations et entreprises aux risques et à promouvoir les bonnes pratiques dans la mise en œuvre de ce type de système.

🔗cyber.gouv.fr/publications/r…

We're excited to announce the launch of Phi-3, a groundbreaking family of small language models that outperform larger models on a range of benchmarks. Learn how these small language models trained on high-quality data are doing more with less: msft.it/6010YHP32

dell.com/en-us/blog/bui…

🎉 Now announcing our Web3 portal—the home for all things Web3 at Google Cloud → goo.gle/4b8VBmj

🚰 Experiment on testnet
💡 Discover blockchain datasets & developer tools from Google Cloud and our partners
⚡ Learn how the Web3 Startup Program can help you build faster

HPE Aruba Networking is unveiling its first line of Wi-Fi 7-capable access points. The new products are differentiated in the wireless market for several reasons, none of which are related to the Wi-Fi 7 standard, according to HPE Aruba Networking.
bit.ly/49UMkgC

youtu.be/ooL2asjrHNQ?si…

🥁 Llama3 is out 🥁
8B and 70B models available today.
8k context length.
Trained with 15 trillion tokens on a custom-built 24k GPU cluster.
Great performance on various benchmarks, with Llam3-8B doing better than Llama2-70B in some cases.
More versions are coming over the next

We are proud to release the first open-source multi-modal, multi-task and multi-domain model! Called JAT. A crucial step for generalist agents. What started out as an open reproduction of GATO with Quentin Gallouédec, Clément ROMAC and myself, has evolved into a far greater project.

Meta's new AI—Llama3—beats an old version of GPT-4, Claude 2.1, and GPT-3.5!

You can run it:
- 100% free
- with 100% privacy (no data leaves your machine)

See the comments for an easy way to install 👇

My mind is blown.

Groq Inc is serving LLaMA 3 at over 800 tokens per second!

800. Tokens. Per. Second.

This unlocks so many incredible use-cases.

It's one thing to see my demo — it's another thing entirely to experience it for yourself.

Do yourself a favor and try it asap.

Already almost 1,000 llama3 model variations have been shared publicly on HF (many more in private use at companies): huggingface.co/models?sort=cr….

Everyone should fine-tune their own models for their use-cases, languages, industry, infra constraints,...

10,000 llama3 variants by

5Y 🎂

Marseille se prépare pour le coup d'envoi du salon Vivatech avec la Région Sud varactu.fr/marseille-se-p… #var #var actu #infos #actus #Vivatech2024 #RégionSud #Innovation #IA #TechMarseille #Startups #EconomieCirculaire #MarseilleTech #IntelligenceArtificielle #EcosystèmeTech

Introduced today, Cisco Hypershield is a brand-new approach to securing data centers and clouds to protect AI workloads, according to Cisco's top security executives, including Jeetu Patel and Tom Gillis.

CiscoSecure CiscoPartners
#CiscoHypershield
bit.ly/3U1Ek7S

VMWare : Pourquoi ces entreprises européennes sont-elles vent debout contre la société américaine Broadcom ? 01net.com/actualites/vmw…

🔴 live stream: real-time stream processing and dwell time calculation

- challenges of real-time stream processing
- hardware acceleration - Intel CPU, Nvidia GPU, Apple M1 chips
- processing multiple streams simultaneously

drop by to say hello: youtube.com/watch?v=u7XUC-…

. Intel said its upcoming Gaudi 3 AI accelerator chip can best Nvidia’s powerful H100 GPU for training large language models and offer similar or better performance than its rival’s memory-rich H200 for large language model inferencing.

#IntelVision
bit.ly/3PUGOUj

❤️ developers.googleblog.com/2024/04/gemini…

Granite Rapids will FUNDAMENTALLY change the way AI will be implemented in the enterprise. ⭐

On stage at #IntelVision , Pat Gelsinger ran a 70 billion parameter 4-bit model entirely on Xeon in just 86 milliseconds. intel.ly/43PdVP5